Virtual Library

Abstract:
The current IASLC/CAP/AMP guidelines indicate that testing for EGFR mutations and ALK fusions is recommended before instituting therapy in patients with advanced lung cancers with adenocarcinoma features. The guidelines indicate that the guidelines would be updated as other therapies become available for specific drivers. Since these guidelines were published, considerable information has become available on molecular changes that cause resistance to the first generation EGFR and ALK tyrosine kinase inhibitors (TKIs) and new 2nd and 3rd generation drugs to treat molecular resistance are in the clinic are likely to be approved in 2015. New drugs are also in development for other molecular drivers that have been reported in lung cancer including fusions that activate ROS1, RET, NTRK, and FGFR; mutations that activate KRAS, BRAF,HER2, and FGFR; and amplifications that activate MET, FGFR and HER2. It is also likely that some of these agents will approved in the near future. These developments bring into focus the most appropriate methods to test for molecular drivers in lung cancer. The techniques include direct sequencing, PCR testing for mutation hot spots, Next generation whole exome sequencing for mutations, translocations and amplifications, immunohistochemical testing for activated proteins or specific mutations, FISH testing for specific translocations, fusions, and amplifications. The sequential testing of one gene at a time is very inefficient especially with respect to the time it takes to complete testing, with the total cost and with respect to the amount of tissue necessary to complete the testing. Thus, multiplex testing is becoming far more common. However, in the US, the FDA’s companion diagnostic tests are one a time for specific tests for specific drugs. This one off policy is threatening the ability to perform one test for multiple analytes simultaneously. Hopefully, these issues can be resolved in the near future. What are the molecular changes that appear to be drivers for which there exist therapeutic TKIs? We can first discuss the development of resistance to 1st generation EGFR and ALK TKIs. About half of all patients who progress on a 1st generation EGFR TKI will have a secondary gatekeeper mutation, T790M that limits binding of 1st generation EGFR TKIs to the EGF receptor. 3rd generation EGFR TKIs such as AZD9291 and CO1686 (rocelitinib) have been shown to produce objective clinical responses in about 60% of patients with T790M who progress on a first generation EGFR TKI. These responses are associated with a median PFS of about 10 months. These drugs bind irreversibly to the T790M receptor and to activating mutations such as del19 and L858R but do not bind to the wildtype receptor. Therefore they have far less skin rash and diarrhea compared to the 1st generation inhibitors. Rocelitinb may cause hyperglycemia in up to a third of patients that is believed to be caused by a metabolite that inhibits IGFR. The use of oral anti-hyperglycemic agents such as metformin may be required for glucose control in these patients. Rocelitinib may also cause prolongation of the QTc on EKGs and thus monitoring is required. This side effect is rare but requires monitoring. AZD9291 may be associated with slightly more skin rash and diarrhea and uncommon reports of pneumonitis have been reported. The mechanism of resistance to these agents is under investigation but most often appears to be activation of alternative signaling pathways or phenotypic switching to a mesenchymal state. There are multiple tests available to detect the T790M mutation and many studies are evaluating its presence in circulating free DNA. Such analyses seem to be quite specific but less sensitive to analyses of tumor tissue. Testing for mutations in circulating free DNA is quite appealing because it does not require another tumor biopsy. Crizotinib which was the first agent approved for ALK positive cases, is also a potent ROS1 and MET inhibitor. ROS1 may be activated by fusion to other genes on the same chromosome and detected by FISH or Next generation sequencing. Crizotinib has been reported to produce objective response is about 60% of cases with ROS1 fusions with median progression free survival of about 16 months. Crizotinib has also been reported to produce objective responses in about 2/3 of patients with high MET amplification although the number of patients studied is very low. Additional studies are ongoing. Because MET amplification is frequent in patients who progress on 1st generation EGFR TKIs that do not have T790M, MET inhibitors are also being studied on this setting. Both ceritinb and alectinib have been approved for patients with ALK fusions who have progressed on crizotinib. It is clear that the two drugs have different activity among various resistance mutations for ALK. Thus, rebiopsy of tumor or testing of circulation free DNA may become standard in patients progressing on ALK TKIs as it is in patients progressing on EGFR TKIs. BRAF mutations occur in about 2% of advanced NSCLC patients and the V600E mutation is the most common. BRAF TKIs such as vemurafinib have been reported to produce frequent responses in NSCLC patients with V600E BRAF mutations. In melanoma the combination of BRAF inhibitors with MEK inhibitors has been more effective than BRAF inhibitors alone and this combination is being studied in NSCLC patient with BRAF mutations. HER 2 may be activated by either mutation or amplifications and the response to various HER inhibitors may vary by the method of activation. About 2% of patients have activation by amplification and about 2% by mutation (usually exon 20 insertions). The irreversible pan HER TKIs such as neratinib, afatinib and dacomitinib have not produced high response rates in these patients. However a phase 1 trial of the combination of niratinib with temsirolimus produced higher responses in both breast and lung cancers and a phase 2 study of the combination in lung cancers with HER2 mutations/amplifications is in progress. Other HER2 inhibitors have been reported to produce occasional responses but we await formal study of such agents. RET and NTRK fusions have been reported I about 1% of patients and there are sporadic reports of responses to specific TKIS. Formal studies are in progress but may not be reported for some time due to the rarity of the abnormalities and the fact that multiple TKIs are available. FGFR may be activated by amplification, fusion and mutations and there are both quite specific FGFR TKIs and multi-TKIs that have been studied in small numbers of patients. It is fair to say that response rates to specific FGFR TKIs are low in patients with squamous cell carcinoma patients with FGFR amplification. Additional studies with other biomarkers and other agents are ongoing. The most frequent oncogenic driver in lung cancer is KRAS. No specific TKI has been developed. Downstream inhibitors such as MEK inhibitors and FAK inhibitors have the most study to date. Response rates as single agents are relatively low and combination studies are in progress. In summary, the increased numbers of TKIs specific for various molecular drivers in lung cancer is becoming far more important not only a diagnosis but also at the time of progression. Future studies will focus on multiplex testing and testing of circulating free DNA.

Abstract:
Technologic and scientific development in medicine and the biomedical sciences has led to almost immediate transfer of knowledge and methodology applied initially in the research lab to the clinic, producing a profound impact in screening, diagnosis and treatment. These enormous amount of information and specialized skills in the medical practice have made it necessary to integrate multidisciplinary teams to improve quality of cancer care.Lung cancer is the most common neoplasia worldwide as well as the leading cause of cancer-related death, with more than 50% of patients presenting with stage IV disease at diagnosis (1). Therefore, lung cancer management usually requires the collaboration of surgeons, medical oncologists, radiation oncologists, pathologists, nurses, and other health care professionals (2). Given the fact that many patients will present with locally advanced or metastatic disease, only small amounts of tissue or cell preparations will be available for morphologic analysis, immunohistochemistry and molecular testing. Recommendations from academic centers and agencies are in favor of limiting the amount of immunohistochemical stains in order to save tissue for molecular assays, stressing the need to integrate pathologist to multidisciplinary teams, where clinical information is exchanged, and specific differential diagnosis and objectives established in a case-to-case basis (3). The traditional role of pathologist in lung cancer has been to establish histological diagnosis of malignancy, as well as proper taxonomic allocation according to widely accepted classification schemes. (4) This approach sets pathologist within single moment interventions early in the course of patient management, aside from opportunities to collaborate during the rest of care. However, the advent of personalized medicine, characterized by the identification of biological features in cells that predict benefit from specific targeted drugs, opened up the way for pathologist to actively participate with the team in the selection of treatment and patient follow-up. Overall, from 60-70% of NSCLC have specific mutations of well characterized oncogenic drivers, some of them with 1st line treatment drugs and many with targeted therapies under development. Current guidelines of advance stage disease recommend initial characterization of EGFR mutation status and ALK rearrangement, although there is building evidence to support testing of a number of actionable genes such as HER2, BRAF, MET, RET, ROS1, or other biomarkers with predictive capacity such as microsatellite instability (5). Multiple methods and technological platforms have been developed to identify gene mutations, and although most of them have very high specificity values, the sensitivity to detect a mutant clone from a background of wild type DNA is wide. Conventional Sanger sequencing will identify a mutation if it is present in 10-20% of the sampled cells, pyrosequencing increases the level to 1%, mutant-enriched polymerase chain reaction (PNA-LNA PCR, ARMS, etc.) can detect a mutant gene among as many as 10[3] wild-type alleles (0.1-1%) with comparative performance to next generation sequencing(6). These differences in analytical sensitivity do not only affect the number of EGFR mutated cases identified, but may also impact the clinical results obtained when using TKI therapy. For example, studies suggest that high EGFR mutation allele burden at diagnosis may be associated with increased progression-free survival and overall survival in patients treated with tyrosine kinase inhibitors, based in sensitivity differences between conventional sequencing and allele-specific PCR(7,8). In the light of these challenges, pathologist face the need to secure tissue availability and adequacy for testing in order to increase the diagnostic yield of molecular characterization. This demands the establishment of changes in sample management and processing, depending on the biological material to be tested. For example, rapid on-site evaluation may be performed in cytological specimens from fine needle aspirates. In the case of CT-guided transthoracic biopsies, one initial core may be submitted for frozen section or studied with cytological imprints to assess tumor viability. If proper cellular material is identified, this core may be entirely used for molecular testing and subsequent cores destined for histological processing. Once the tissue is paraffin-blocked, the tissue cuttings product of facing the block may be saved in a sterile, DNAase/RNAase-free tube for later use if necessary. It is established that patients will ultimately develop resistance to targeted therapies through different mechanisms, either the emergence of mutations in the target gene or the acquisition of mutations or over-expression of oncogenes that overcome this inhibition. Studies have proved that at tumor progression, a number of cases may have a morphological switch from adenocarcinoma to sarcomatoid carcinoma or small-cell carcinoma (9), requiring therapy adjustments. Re-biopsy allows molecular evaluation as well as morphologic analysis, however; it is an invasive procedure that not all patients may undergo. Alternative highly sensitive molecular methods may be used for patient follow-up without the need of invasive interventions. Blood sample-based PCR or NGS can detect circulating free DNA from the tumour (liquid biopsy), the concordance rate between tissue and plasma for EGFR mutation going from 58.3% to 93.1%, stressing the need of analytical improvement. Of especial interest is the fact that when examining the appearance of the T790M mutation in serial blood samples, this mutation could be detected in the plasma DNA before clinically evident disease progression. (10) NSCLC diagnostics is rapidly changing to combine a dual morphologic-molecular approach, where initial HE-slide examination is used to confirm malignancy and to allocate the tissue sample to a group of molecular test relevant to the cellular composition of the tumour. Continuous increase in the number of genes responsible of oncogenesis in lung carcinoma ensures the development of new targeted drugs as well as active communication from all the members of the multidisciplinary team. References 1) Aisner DL, Marshall CB. Molecular pathology of non-small cell lung cancer: a practical guide.Am J Clin Pathol. 2012 Sep;138(3):332-46. 2) Pan CC, Kung PT, et al. Effects of Multidisciplinary team care on the survival of patients with different stages of Non-Small Cell Lung Cancer: A National Cohort study. PLoS One. 2015 May 12;10(5):e0126547. 3) Lindeman NI, Cagle PT, Beasley MB, et al: Molecular testing guideline for selection of lung cancer patients for EGFR and ALK tyrosine kinase inhibitors: Guideline from the College of American Pathologists, International Association for the Study of Lung Cancer, and Association for Molecular Pathology. Arch Pathol Lab Med 137:828-860, 2013 4) Cagle PT, Myers J.Precision medicine for lung cancer: role of the surgical pathologist. Arch Pathol Lab Med. 2012 Oct;136(10):1186-9. 5) Dacic S, Nikiforova MN.Present and future molecular testing of lung carcinoma. Adv Anat Pathol. 2014 Mar;21(2):94-9. 6) Young EC, Owens MM, Adebiyi I, et al. A comparison of methods for EGFR mutation testing in non-small cell lung cancer. Diagn Mol Pathol. 2013 Dec;22(4):190-5. 7) Kim HS, Sung JS, Yang SJ. Predictive efficacy of low burden EGFR mutation detected by next-generation sequencing on response to EGFR tyrosine kinase inhibitors in non-small-cell lung carcinoma. PLoS One. 2013 Dec 20;8(12):e81975. 8) Zhou Q, Zhang XC, Chen ZH. Relative abundance of EGFR mutations predicts benefit from gefitinib treatment for advanced non-small-cell lung cancer. J Clin Oncol. 2011 Aug 20;29(24):3316-21. 9) Sequist LV, Waltman BA, Dias-Santagata D. Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors. Sci Transl Med. 2011 Mar 23;3(75):75ra26. 10) Sorensen BS, Wu L, Wei W, et al. Monitoring of epidermal growth factor receptor tyrosine kinase inhibitor-sensitizing and resistance mutations in the plasma DNA of patients with advanced non-small cell lung cancer during treatment with erlotinib. Cancer. 2014 Dec 15;120(24):3896-901.

Abstract:
Remarkable progress has been made in the treatment of ALK-rearranged NSCLC with the initial approval of crizotinib. Multiple next-generation agents are impacting care as approved therapies in some regions of the world or in first- and second+-line clinical trials where early data are promising. Additional development of crizotinib and newer agents in ROS-1-positive NSCLC is quickly changing how patients are evaluated at diagnosis. Moreover, maturing data in BRAF-mutated NSCLC with BRAF and MEK inhibitors, along with early data in RET- and FGFR-positive NSCLC support broader and earlier testing in the care of patients with advanced NSCLC. The data for, and challenges of, selecting the best first-line (and beyond) options for patients with ALK, ROS-1, BRAF, RET, FGFR, or other altered NSCLC will be reviewed.

Abstract:
When more than one drug exists for the same target (e.g. erlotinib, icotinib, gefitinib and afatinib for EGFR mutant NSCLC), decisions in terms of which drug to commence targeted therapy with may be based on personal experience, side effects and cost. Until head-to-head trials in the same sensitive target population read out, any potential efficacy differences remain speculative. Other than to switch out drugs based on tolerability issues, the anti-cancer benefit from sequential use of different drugs in the same class/generation as a strategy separate from any benefit of ongoing target suppression post-progression with the same drug also remains largely unproven. However, as new generation drugs are developed, drugs directed against the same dominant driver now exist which have activity against both the initial form of the target and common acquired resistance forms (e.g. ceritinib and alectinib for ALK+ NSCLC, or rociletinib and AZD9291 for EGFR mutant NSCLC). While initial licensing strategies have focused on sequential use of such drugs after first generation drugs, studies are also underway looking to see if first line use of these next generation drugs could be more beneficial in the relevant molecularly selected population. As mechanisms of acquired resistance become better understood, specific actionable second drivers, co-existing with the initial sensitive form of the oncogenic driver, are now being identified (e.g. MET amplification with EGFR mutations). Preclinically, inhibition of both drivers is required to achieve cancer control in this setting. Clinically, trials of combination therapy are showing promise with determination of the exact predictive cutpoint in continuous variables such as MET emerging as a key issue. With more extensive molecular testing being deployed upfront, rather than in tiered testing strategies and separate from in the acquired resistance setting, multiple potential molecular drivers on each patient are now being reported to treating physicians. When some of these are known acquired mechanisms of resistance (e.g. MET or T790M in EGFR mutant), concern may arise re whether initial therapy will be effective. However, most diagnostic assays do not give information on the proportion of each molecular aberration. Consequently, a highly sensitive test can detect a small clonal fraction of a resistance mechanism that will later be selected out by use of the initial targeted therapy, but does not preclude an initial response. In contrast, germline abnormalities present in all cells, such as can occur with T790M, would preclude an initial response to the relevant targeted therapy. As such germline events are very rare, initial use of a targeted agent still makes most sense. While defined oncogenic drivers are often perceived to be mutually exclusive (e.g. EGFR and ALK), exceptions do occur. While there appears to be no specific growth advantage to having two oncogenic drivers in the same cell in the absence of a specific selection pressure, such examples of double drivers could reflect false positives, true positives where one of them is somehow non-functional (e.g. an ALK rearrangement detected by FISH, which is not transcribed), or true positives where each is present in a different cell population. Again, as proportional positivity is not a feature of most diagnostic assays, starting with monotherapy for the abnormality that is either easiest to target or has the lowest chance of being non-functional (e.g. a point mutation over a chromosomal abnormality), makes the most clinical sense. At acquired resistance, rebiopsy and reanalysis for changes in biology including the dominance of the other initial driver should be strongly considered. Perhaps the biggest challenge has been the proliferation of multiplex reports detailing a range of abnormalities in the same cancer, where the exact driver status and biological significance of the abnormalities remains unclear. Caution should be exercised in assuming that all changes are true drivers and extreme caution should be exercised if attempting to combine available targeted drugs that have not been combined before in the absence of a formal phase I study.

Background:
Early detection of lung cancer using low-dose CT led to a 20% reduction in mortality. However, this strategy has several limitations including high false-positive rates, potential over-diagnosis, and the potential harm associated with radiation exposure. The aim of this study was to identify differentially-expressed miRNAs in the serum of non-small cell lung cancer (NSCLC) patients that might be a clinically-useful tool for lung cancer early detection.

Methods:
We performed miRNA expression profile analysis using TaqMan OpenArray Human panel in a discovery set of 70 serum samples obtained at lung tumor resection including lung adenocarcinoma (AD) and lung squamous carcinoma (SCC) and 22 non-cancer subjects (NC). To construct the diagnostic signature, the miRNA candidates were selected based upon the following criteria: miRNAs significantly up-regulated (adjusted t-test p < 0.001) in the NSCLC tissue and serum as compared to normal lung tissue and NC serum respectively, not overexpressed in circulating blood cells and with Area Under the Curve (AUC) > 0.840 for discriminating stage I LC from NC in the receiver-operating characteristic (ROC) plots. Selected serum miRNAs were then validated by quantitative PCR using an independent validation set of serum samples from LC patients (n=84) and NC (n=23).

Results:
Sixty miRNAs were significantly up-regulated and 31 were down-regulated in the serum from NSCLC patients versus NC (adjusted p<0.001). Four miRNAs (miR-193b, miR-301, miR-141 and miR-200b) were selected for validating their diagnostic value in an independent cohort. A diagnostic signature was obtained by logistic regression based upon the expression values of these 4 serum miRNAs in the discovery set. This miRNA signature generated an AUC of 0.985 (95% CI 0.961 – 1.000, p < 0.001) for detecting NSCLC (all stages) and of 0.989 (95% CI 0.967 – 1.000, p < 0.001) for detecting stage I NSCLC in the discovery set. In the test set, the diagnostic utility of this miRNA signature was validated and exhibited an AUC of 0.993 (95% CI 0.979 – 1.000, p < 0.001).

Conclusion:
We identified a serum 4-miRNA signature that discriminated with high accuracy lung cancer patients from NC. Further prospective validation of this miRNA signature is warranted using an independent cohort of serum samples from patients who participated in a lung cancer screening program.

Background:
In the context of CT screening in current and former smokers at high risk for lung cancer, the false positive rate is high (26% at first NLST screening; 13% with Lung-RADS criteria applied to NLST) and indeterminate nodules are frequently discovered. Noninvasive biomarkers are urgently needed to reduce false positives with screening CT and to improve risk stratification in those with indeterminate nodules. The Colorado (CO) Lung SPORE program performed a retrospective longitudinal evaluation (Pepe Phase 3 validation) to assess the potential of chromosomal aneusomy detected in sputum via fluorescence in situ hybridization (CA-FISH) as a biomarker for early detection in four nested case-control studies. Two of the cohorts (ACRIN/NLST and PLuSS) enrolled current and former smokers to investigate use of low dose CT to diagnose lung cancer. The other two were Colorado cohorts in which pulmonary clinic patients (mostly current and former smokers) were enrolled to investigate biomarkers to predict lung cancer. One of these cohorts (CO High Risk) was a COPD population and the other, still in the accrual phase, comprises patients referred for care of indeterminate lung nodules (CO Nodule).

Methods:
The cohorts were grouped into a Screening cohort (ACRIN/NLST (49 cases, 96 controls) and PLuSS (48 cases, 89 controls)) and a High Risk cohort (CO High Risk (55 cases, 59 controls) and CO Nodule (13 cases, 10 controls)). The CA-FISH assay was a 4-target panel including genomic sequences encompassing the EGFR and MYC genes, and the 5p15 and centromere 6 regions or the FGFR1 and PIK3CA genes. At the subject level, the assay was scored on a 4-category scale representing normal, probably normal, probably abnormal and abnormal. Operating characteristics (with 95% CI) of the assay were estimated for each group of cohorts overall and separately for COPD patients: sensitivity, specificity, likelihood ratio+ (LR+) and likelihood ratio- (LR-).

Results:
Using the cutoff of abnormal vs. not abnormal for CA-FISH, sensitivity and specificity for Screening subjects are 0.20 (0.13, 0.30) and 0.84 (0.78, 0.89), respectively; and for High Risk subjects are 0.67 (0.55, 0.78) and 0.94 (0.85, 0.98), respectively. Likelihood ratios for Screening subjects are LR+: 1.36 (0.81, 2.28) and LR-: 0.93 (0.83, 1.05), and for High Risk subjects are LR+: 11.66 (4.44, 30.63), and LR-: 0.34 (0.24, 0.48). Similar results were observed when only COPD subjects were analyzed.

Conclusion:
The high LR+ of sputum CA-FISH indicates that this noninvasive biomarker could be a clinically useful adjunct to CT among patients in high risk settings. Whether this same high level of LR+ will be reproducible in patients at high risk because of their indeterminate nodules remains to be seen. If so, a hypothetical patient with indeterminate nodules and a pre-test (CA-FISH) lung cancer risk of 20% would have a post-test probability of lung cancer of 78% if the CA-FISH test were positive. In the screening setting, however, the low LR+ of CA-FISH limits its clinical utility. Prospective assessment of sputum CA-FISH is ongoing in the Nodule Cohort of the CO Lung SPORE.

Background:
Lung cancer is the leading cause of cancer death worldwide. Non-small cell lung cancer (NSCLC) accounts for about 80% of primary lung cancer cases and approximately two thirds of them are diagnosed at an advanced stage . The poor prognosis of this disease is partially due to the lack of an effective means of early diagnosis. Discovery of an effective and reliable tool for early diagnosis of lung cancer would play a pivotal role in improving the prognosis of patients with lung cancer. Pulmonary ground-glass nodules (GGNs) are increasingly detected in clinical practice. GGNs are related to lung cancer, especially lung adnocacinoma . The subject of how to manage the pulmonary GGNs remains controversial. It is necessary to identify biological markers that can be used to screen high-risk patients in order to allow better lung adenocacinoma presenting with GGNs detection, earlier intervention and increase the likelihood of successful treatment. MicroRNAs are small non-coding RNAs of 18–24 nucleotides, typically excised from 60–110 nucleotide foldback RNA precursor structures . MicroRNAs have drawn significant attention in cancer research after it was linked to oncogenesis and tumor metastasis. Abnormal expression of microRNAs has been found in both haematopoietic and solid tumours by various genome-wide techniques. There is no report about the relationship between microRNA and pulmonary GGNs. It is necessary to identify biological markers that can be used to screen high-risk patients presenting GGNs in order to allow early lung adenocacinoma detection. Our study investigated microRNA expression with the intention to identify a panel of microRNAs for the diagnosis of lung adenocarcinoma presenting with GGNs.

Methods:
73 pairs of samples (tumorous and non-tumorous) were surgically resected from lung adnocacinoma patients presenting with GGNs from Shanghai Pulmonary Hospital between May 2012 and June 2014. After obtaining the approval of the patient consent, fresh tissues samples were taken during surgical resection, snap-frozen on dry ice and stored at−80◦C. MicroRNA expression of tumor and non-tumorous tissues was investigated in 3 participants by the next generation sequencing. Then, we analyzed the difference expression microRNA profiles which were identified by second generation sequencing in 73 pairs of lung adenocacinoma presenting with GGNs and adjacent non-tumorous tissues using a quantitative reverse-transcriptase polymerase chain reaction assay (qRT-PCR).

Results:
When we compared microRNA expression among lung cancer tissues versus corresponding noncancerous lung tissues via next-generation sequencing, 23 microRNAs had statistical differences in expression between groups. Five microRNAs (hsa−miR−548ar−5p, chr10_7330_star, chr17_10932_star, hsa−miR−148a−3p, hsa−miR−210−3p) exhibited higher expression in the adnocacinoma samples than that in the non-tumorous samples, eighteen microRNAs (hsa−miR−548x−5p, hsa−miR−144−3p, hsa-miR-106a-5p, hsa−miR−548ay−5p, hsa−miR−199a−3p, hsa−miR−378d, hsa−miR−4732−3p, hsa−miR−486−3p, chr7_5517, hsa−miR−1307−5p, chr17_10880, hsa−miR−127−3p, hsa−miR−411−5p, chr1_1402, chr16_10269, hsa−miR−138−5p, hsa−miR−212−3p, hsa−miR−33b−5p) demonstrated lower expression in adnocacinoma samples than that in the non-tumorous samples (P＜0.05). Further validated by qRT-PCR, six microRNAs (chr17_10932_star, hsa−miR−148a−3p, hsa−miR−210−3p, chr1_1402, hsa−miR−378d, hsa−miR−138−5p) were statistically differentially expressed in tumorous compared with non-tumorous tissues.

Conclusion:
We found a microRNA panel that has considerable clinical value in diagnosing lung adenocacinoma presenting with GGNs. Thus, patients who would have otherwise missed the curative treatment window can benefit from optimal therapy.

Background:
A previous, non-blinded study presented at the American Society of Cytopathology demonstrated performance of the LuCED[®] test for early detection of lung cancer and showed a sensitivity to cancer of 93.6% with 100% specificity based on 94 patients. Sensitivity was consistent across tumor histology, stage, size and location. Here, LuCED performance is presented where the pathologist was blinded to the case diagnosis. Data for this evaluation was produced as part of the CLIA validation of LuCED for use in the VisionGate Biosignatures Laboratory.

Methods:
Sputum from 42 patients was processed by LuCED: 23 patients had biopsy-confirmed lung cancer and 19 patients were normal. Sputum was collected from three spontaneous morning coughs, fixed and stained with hematoxylin, and enriched for epithelial cells using fluorescence activated cell sorting. Each enriched specimen was analyzed using the Cell-CT® platform that computes 3D digital images of single cells through tomographic reconstruction with isometric, sub-micron resolution. 3D morphometric biosignatures were automatically measured to produce a probabilistic score that identified abnormal cell candidates while a second score identified normal bronchial epithelial cells to determine specimen adequacy. Specimen adequacy was achieved when either abnormal cells were detected or 800 normal bronchial epithelial cells were enumerated by the classifier, whichever came first. Data was randomized by case and cell images of abnormal candidates were viewed using a CellGazer® workstation for blinded, cytopathologist confirmation. Cases were run until one of the following conditions was met: an abnormal cell was discovered, the specimen was exhausted, the criterion for specimen adequacy was reached. Example images of positive cells are shown in Figure 1. Figure 1



Results:
For cancer cases, lung cancer histology included adenocarcinoma (10 cases), squamous cancer (7), small cell lung cancer (3) and undifferentiated cancer (3); representing TNM stages I (5), II (10), IV (5), and unknown (3). Abnormal cells were found in all 23 cancer cases for 100% case sensitivity (lower 95% CI bound: 85.1%). Non-cancer lung diseases may produce reparative changes whose morphology can mimic cancer cell features. To stress test LuCED, patients with COPD, bronchitis, etc., were included in the normal group. 100,645 cells were processed from the 19 normal cases with 0.47% identified by the classifier for review using CellGazer. No abnormal cells were found. Case specificity is 100% (lower 95% CI bound: 82.4%).

Conclusion:
This interim blinded study of LuCED performance demonstrates highly sensitive (100%) and specific (100%) early lung cancer detection suggesting utility as a non-invasive screening test.

Abstract not provided

Background:
Lung cancer (LC) is the leading cause of cancer mortality. Computed tomographic (CT) screening detects LC at earlier stages but also results in finding many more smaller, benign nodules. Positron-Emission-Tomography (PET) is commonly used to evaluate suspicious lesions prior to invasive biopsies. However, PET accuracy is confounded by various factors including size, inflammation and tumor metabolic activity. Another potential biomarker of cancerous tissue is a non-invasive measure of transcutaneous bioconductance. This study compares Electro Pulmonary Nodule (EPN), a tissue bioconductance scan, to 18FDG-PET in evaluating CT detected suspicious lung lesions.

Methods:
Cohort- 27 patients with suspicious nodules evaluated with both PET and EPN scanning (IRB approved protocol) prior to biopsy or long-term radiologic follow-up. An EPN Scan measuring bioconductance was performed on bilateral anatomic skin sites and results were scored as either positive or negative dependent on a defined cut-off point. The PET results were interpreted as positive, negative or indeterminate.

Results:
There were 18 LCs (16 non-small cell LC, 2 small cell LC) and 9 benign lesions. PET results yielded 7 indeterminate readings. Excluding these 7, PET had 100% sensitivity (14/14 true positives) and 67% specificity (4/6 true negatives). EPN Scan evaluation of these 20 determinate PET cases had 86% sensitivity (12/14 true positives) with 83% specificity (5/6 true negatives). When evaluating the entire cohort of 27, the EPN results improved sensitivity and specificity to 89% (16/18 true positives) and 89%(8/9 true negatives), respectively. Table 1 describes the 7 lesions 18FDG-PET indeterminate lesions compared to EPN Scanning. Figure 1 In these 7 cases, the EPN Scan correctly classified all cases for 100% accuracy. Of note, 2 of the 4 cancers classified by PET as “indeterminate” were < 1 cm and were correctly categorized by the EPN Scan.



Conclusion:
While 18FDG-PET is often used as a clinical adjunct in the evaluation of suspicious CT detected pulmonary lesions, it has recognized limitations. In this feasibility study of measuring transcutaneous bioconductance as a pre-biopsy assessment, EPN Scan performed favorably versus PET, especially in evaluating smaller or PET-indeterminate lesions.

Background:
Lung cancer (LC) is the leading cause of cancer death in the United States with more than 158,000 estimated deaths in 2015. Early detection of LC has been well established as a significant key point in patients' survival and prognosis, yet unfortunately, the vast majority of new LC patients are being diagnosed at advanced disease stages. Exhaled breath analysis can serve as a non-invasive method in early detection of LC. The tumor's micro-environment releases various compounds to blood, some of which are then exhaled at breath as Volatile Organic Compounds (VOCs). This study evaluates the potential of exhaled breath analysis in LC detection and to further diagnose histology, EGFR mutational status and to discriminate early from advanced disease in a multinational study.

Methods:
Breath samples were taken from untreated LC patients and matching controls. Patients were enrolled in a large tertiary referral hospital in Israel. Analysis was performed by gold nanoparticle-based Artificial Olfactory System (NaNose®) and Pattern recognition methods were used to analyze the results obtained from the NaNose®. Histology, EGFR mutation status and staging was taken from patient's files.

Results:
A total of 174 patients participated in this study, and Inter-group analysis of 80 LC patients (64 advanced stage) and 31 matched controls showed a significant discrimination between disease and control. Among all patients, 83 were adenocarcinoma and 11 were squamous. EGFR mutations were detected in 24 patients. The comparisons resulted in: early LC versus control: p < 0.0001; accuracy 85.11%, advanced LC versus control: p < 0.0001; accuracy 82.11%, early LC versus advanced LC: p < 0.0001; accuracy 78.75%. Histology (Adenocarcinoma vs. Squamous cell carcinoma) and EGFR status was also significantly determined by the volatile signature.

Conclusion:
Breath analysis may support early detection of cancer as well as histological diagnoses, staging and mutational testing in lung cancer. This innovative method may pose as an important non-invasive tool for lung cancer early detection, thus promoting better prognosis and therapeutic possibilities for patients.

Background:
Autoantibodies is an attractive diagnostic approach for early detection of malignant tumors. Our previous studies found a panel of 7 TAAs(p53, GAGE7, PGP9.5, CAGE, MAGE A1, SOX2, GBU4-5) was associated with lung cancer. We performed this large-scale, multi-center clinical trial to validate their ability to aid early diagnosis of lung cancer in Chinese population. Autoantibodies is an attractive diagnostic approach for early detection of malignant tumors. Our previous studies found a panel of 7 TAAs(p53, GAGE7, PGP9.5, CAGE, MAGE A1, SOX2, GBU4-5) was associated with lung cancer. We performed this large-scale, multi-center clinical trial to validate their ability to aid early diagnosis of lung cancer in Chinese population.

Methods:
The 7 TAAs were selected from 43 candidate TAAs from our previous studies, which were detected by ELISA in 1915 participants from 5 clinical centers in China. These samples including lung cancer (n = 818), benign lung diseases (n = 386), healthy volunteers (n = 415) and interference group (n = 296). The sensitivity and specificity from 7 TAAs and the traditional cancer biomarkers CEA, NSE and CYFRA21-1 were compared.

Results:
The sensitivity and specificity of autoantibody assay were 61% and 90% respectively, which were similar in different subgroups such as age, gender, smoker status and histological type. As for the enrolled patients with lung cancer, the sensitivities were 60% for patients with stage I/II, which were significantly higher than 27% ( p < 0.01)when using the combination of CEA, NSE and CYFRA21-1 to detect patients with lung cancer. While in patients with stage III/IV lung cancer, sensitivities were similar (63% vs. 56%, p > 0.05) and specificity was significantly improved (90% vs 71%, p < 0.01). The specificity was consistent in benign lung diseases and autoimmune diseases(interference group) and were 90% and 94% respectively and a concentration decrease of 7 TAAs were also observed after tumor resection.

Conclusion:
This study suggest that the 7 TAAs autoantibody panel can be used to aid diagnosis of lung cancer, and show a significantly improving sensitivity in patients with early stage lung cancer when comparing with the combination of CEA, NSE and CYFRA21-1.

Background:
Since the majority of lung cancer cases are detected at a late stage the prognosis remains poor at present. The National Lung Screening Trial (NLST) reported 20% reductions in lung cancer mortality in 2011, however as a primary screening modality CT is expensive and may lead to significant morbidity in individuals whose tests are false positives. The EarlyCDT-lung test detects autoantibodies to proteins in the earliest stages of the disease with a specificity of 93%. Research question Does using the EarlyCDT-Lung test reduce the incidence of patients with late-stage lung cancer (3 & 4) or unclassified presentation (U) at diagnosis, compared with standard practice?

Methods:
We are conducting an RCT of 12 000 participants in areas of Scotland within the most deprived quintile of the population whose mortality from lung cancer is high by international standards. Adults aged 50 to 75 who are at 1.2% risk over the next 2 years are eligible to participate. They should also be healthy enough to undergo curative interventions. We will undertake a comparison of the EarlyCDT-lung test and follow-up imaging at six monthly intervals for 2 years with standard clinical practice. The primary outcome is the difference, after 24 months, between the rates of patients with stage 3, 4 or unclassified lung cancer at diagnosis. Participants who develop lung cancer will be followed-up via electronic record-linkage to assess both time to diagnosis and stage of disease at diagnosis. The secondary outcomes are cost-effectiveness, and a range of psychological measurements. There is a nested qualitative study of the psychological effects test of results on participants.

Results:
In the first 14 months of recruitment 8 848 patients have been recruited and 9.0% of those tested have had a positive blood test with eight early cancers and 13 abnormalities undergoing further investigation detected to date in those who tested positive. Six of the eight cancers have been staged and four of these are early cancers. Provisional data reported to the trial team on those tested negative include three cancers. No data are currently available for the main trial comparison. From prior observational studies the test performance is expected to be: 40% sensitivity and 90% specificity these early data. Based on the study so far the current Positive Predictive Value of the test is 2.0%.

Conclusion:
The study will determine the EarlyCDT-Lung test’s clinical and cost effectiveness. It will also assess potential morbidity arising from the test and potential harms and benefits of a negative EarlyCDT-Lung test result. Early results in the test only arm of the trial are encouraging.

Abstract not provided

Background:
Early CDT-Lung, a serum based biomarker consisting of a panel of seven autoantibodies that develop in response to tumor associated antigens, has been shown to detect lung cancer in all stages of disease. We hypothesized that this biomarker when used in combination with a low-dose CT (LDCT) in screening of a high-risk population would increase the detection of early stage lung cancer.

Methods:
A prospective study of 1,600 subjects at high risk for lung cancer was designed. Eligibility criteria included persons 50-75 years of age, current or former smokers of ≥ 20 pack years and < 10 years since quit smoking. Those with a history of lung cancer in first degree relative(s) and any history of smoking were included. Exclusion criteria ware any history of cancer within 10 years (except skin cancer), any use of oxygen, and life expectancy of < 5 years. A direct mail campaign was conducted with study announcements sent to the homes of potentially high-risk individuals, who then contacted us for consideration of participating in this screening study. Those fitting inclusion criteria received the Early CDT-Lung blood test and a LDCT. A nodule of ≥ 3mm was considered as a positive scan. The Early CDT-Lung test was considered positive if any one of the seven autoantibodies was positive. All participants are to have yearly telephone follow-up for two years.

Results:
From May 2012 through November 2014, 815 individuals were enrolled and 814 completed the initial blood and LDCT screening tests. The cohort median age was 59 years with 55% female and 45% male gender distribution. The mean smoking history was 44 pack-years. Fifty-four per cent were current smokers while 46% were former smokers. Forty-six per cent of the LDCTs were negative for any lung nodule while 38% were positive. Incidental non-lung cancer findings were identified in 15% of the study group. The Early CDT-Lung biomarker was positive in 60 (7%) of participants, 23 males and 37 females. In those with a positive LDCT (n=313), the biomarker was positive in 25 (8%). As of January 30, 2015, there have been six confirmed lung cancers: two limited stage small cell, two Stage IB adenocarcinoma (ACA), and two Stage IA (one ACA and one squamous cell). The Early CDT-Lung blood test was positive in two of the four Stage IA/B lung cancers and negative in the two small cell cancers.There are 35 Early CDT-Lung biomarker positive individuals whose LDCT had no nodule. These individuals are being followed with yearly LDCT for two years. The study is continuing to accrue with a goal of 1,600. (NCT01700257)

Conclusion:
The Early CDT-Lung biomarker was positive in 7% of our high-risk population. The biomarker was positive in two of six lung cancers, specifically in two of four Stage I lung cancers. Accrual to the study and follow-up of 35 biomarker positive but LDCT negative participants continues.

Background:
The United States Preventive Services Task Force recommends annual CT-screening for lung cancer in high risk adults but also acknowledges that one disadvantage of CT-screening is the large number of false positive results. Circulating biomarkers may provide a noninvasive, cost-effective means of addressing this disadvantage by assisting with classification of patients with indeterminate pulmonary nodules. Here, we describe the development and testing of a blood-based 5-analyte panel to classify these patients.

Methods:
A 5-analyte panel was developed in a training study comprising stage I NSCLC patients (n=95) and healthy smoker controls (n=186). The ability of the biomarker to resolve patients with benign nodules from those with malignant lesions was investigated in two validation studies: (1) Prostate, Lung, Colorectal, Ovarian (PLCO), a CXR-based screening trial, cases n=56, controls n=56; (2) Conversant Bio (CB), cases n=22, controls n=22.

Results:
In the training study, the 5-marker classifier (TFPI, OPN, CEA, CYFRA, SCC) resolved malignant cases with 72% sensitivity and 90% specificity (AUC=0.90). In the PLCO validation study, the biomarker distinguished pre-diagnostic cases with an AUC=0.65. In the CB study, a clinical model developed integrating nodule size, nodule location and gender, classified subjects with an AUC=0.79. When added to the clinical model, the biomarker significantly improved overall accuracy (P=0.016; AUC=0.86).

Conclusion:
A blood-based biomarker has been developed that accurately classifies patients with indeterminate nodules. Adding this biomarker to currently employed clinical and imaging-based evaluations of pulmonary nodules, may prove valuable in assessing malignant risk.

Background:
Lung cancer is the worldwide leading cause of cancer-related mortality. Almost 85% of lung cancer cases are diagnosed at late stages with a five-year-survival probability at the time of diagnosis of 16.8%. The National Lung Screening Trial (NLST) showed a 20% reduction in lung cancer mortality using low-dose computed tomography (CT) screening, but there was also a 96.4% false positive rate. Lung cancer screening might be improved through cancer specific biomarkers detected in body fluids such as plasma or sputum. Previous studies using DNA methylation failed to achieve adequate sensitivity because of use of infrequently methylated genes and detection techniques unable to detect the small amounts of DNA yielded from blood and sputum. We sought to improve the diagnostic accuracy using gene promoter methylation in blood and sputum through the use of Methylation On Beads (MOB) and a highly lung-cancer specific panel of genes for detection of lung cancer.

Methods:
We conducted a prospective case-control study obtaining cases and controls from the Lung Cancer Spore. Cases had pathological confirmation of Non-Small Cell Lung Cancer (NSCLC) lesion stage IA or IB. Controls were defined as patients with pathological confirmation of non-cancerous lesion in the surgical specimens. Plasma, sputum and CT scans were obtained pre-operatively. We quantified methylation levels and the amplification cycle threshold from sputum and plasma samples by using MOB and quantitative methylation specific real-time PCR lung cancer-related genes previously identified from The Cancer Genome Atlas (TCGA). This panel of genes include: CDO1, TAC1, HOXA7, HOXA9, SOX17 and ZFP42.

Results:
A total of 210 subjects fulfilled inclusion criteria, including 150 patients with NSCLC and 60 patients with non-cancerous lesions. All six genes were methylated in significantly more people with cancer than without cancer in both plasma and sputum (p<0.001) with the exception of HOXA9 in sputum, which was methylated in more than 90% of people with cancer and more than 90% of people without cancer. After adjusting by age and pack·year, the methylated genes that were significantly associated with risk of lung cancer stage IA & IB from blood samples were: CDO1 (p=0.009), TAC1 (<0.001), HOXA9 (p=0.005), SOX17 (<0.001) & ZFP42 (p=0.003) and from sputum samples were: CDO1 (p=0.066), TAC1 (p=0.007), ZFP42 (p=0.009). Sensitivity and specificity for lung cancer diagnosis using the 3 best genes in plasma was 91% and 68% respectively and for sputum 91% and 88%. Area under the curve for 3 best genes in plasma was 0.78 95% confidence interval (CI) (0.69-0.87) (p<0.001) and for the best 3 genes in sputum 0.88 95% CI (0.77-0.99) (p<0.001).

Conclusion:
This study shows that its is possible to obtain high diagnostic accuracy for Lung Cancer in early stages using a panel of methylated promoter genes in Plasma and Sputum, by using Methylation-on-beads. These epigenetic biomarkers could potentially be used to identify patients with high risk of lung cancer development. reducing unnecessary tests and increasing the chance to diagnose lung cancer at earlier stages

Background:
There is a need for non-invasive airway-based biomarkers in lung carcinogenesis for both risk assessment of the ex-smoker, and earlier diagnosis. Exhaled breath condensate (EBC) contains airway molecules, presumably in part from bronchial and alveolar epithelial cellular origins. Our previous study showed microRNAs could qualitatively be detected in EBC. Here both qualitative and quantitative multivariate analysis were applied to look for microRNA candidates in EBC from a new sample of lung cancer patients and controls.

Methods:
MicroRNA expression profiling using RNA-specific RT-qPCR was performed in EBC from 41 patients and 41 contols with clinical and microRNA expression data. The panel of microRNAs was assembled based on literature-derived reports of blood or lung microRNAs which segregate with case-control status, combined with our own lung tissue-based discovery effort using microRNA-seq on lung tumor-non-tumor pairs. The assembled panel for this effort included n=19 tumor-non-tumor differentiating microRNAs (miR-9, 18a, 20a, 31, 130b, 142, 146, 182, 183, 196a, 200a, 200c, 205, 210, 212, 221, 224, 330 and 708) chosen from the literature and our own lung tissue-based discovery data. Small nuclear RNA U1 was a housekeeping gene in the study based on its universality. Qualitative and quantitative (miRNA qPCR data normalized to internal reference U1 small ncRNA) analyses were considered. Multivariate analyses considered clinical information, including age, smoking status, underlying lung disease (COPD or not).

Results:
By univariate analyses, between cases (all histologies) and controls, qualitative/binary data showed miR-221 (p=0.030; OR=3.11) and miR-708 (p=0.016; OR=3.04) were significantly different. The case-adenocarcinoma subgroup (n=13) also differed from the controls in miR 708 frequency (p=0.034, OR=4.71). Examples of multivariate analyses (qualitative/binary data, case – all histologies) are shown in the Table: ontrols.

miRNA	Odds Ratio	lower bound of CI	upper bound of CI	p-value
miR.221	3.339	0.994	12.482	0.059
age	1.084	1.026	1.158	0.008
smoking 1vs0	1.467	0.304	8.372	0.642
smoking 2vs0	2.211	0.411	14.436	0.371
Underlying lung dz (COPD vs no COPD)	3.400	1.184	10.349	0.026
miR.708	5.041	1.651	17.603	0.007
age	1.093	1.031	1.172	0.006
smoking former vs never	1.378	0.273	8.145	0.704
smoking current vs never	2.144	0.386	14.269	0.397
Underlying lung dz (COPD vs no COPD)	4.437	1.448	15.047	0.012

Similar multivariate models were obtained for miR 221 and miR708 in the cancer-adenocarcinoma subgroup. No clear case-control discriminant exhaled microRNAs were found in the analogous quantitative data (delta CT) analyses, by univariate or multivariate analyses.

Conclusion:
From the qualitative analysis, two possible miRNA biomarkers of case status (miR-221 and miR-708) were obtained. Previous work had suggested miR 221 as a discriminant microRNA in lung cancer case versus control setting. Quantitative data was not informative. We are working on expanding and refining the miR panel, and larger sample size to partition covariates such age, underlying lung disease, and other factors. Our goal is to test this non-invasive biomarker approach to lung cancer risk assessment.

Abstract not provided

Background:
The National Lung Screening Trial (NLST) established a 20% reduction in lung cancer-specific mortality with low-dose computed tomography (LDCT) in 30 pack year smokers. However, approximately 25% of all lung cancers occur in non-smokers, and screening guidelines are needed for this large cohort. Pre-test probability of lung cancer can be estimated in this group using a validated risk prediction model [Liverpool Lung Project, LLP]. The LLP compares risk in 579 lung cancer cases with 1157 age and sex matched controls.

Methods:
We used the LLP model to illustrate risk profiles for non-smoking females compared to 30 pack year smokers [the NLST target population]. This tool revealed the individual and cumulative effect of risk factors in non-smoking females. The LLP estimates the probability of developing lung cancer within 5 years based on age, sex, smoking history, family history of lung cancer, infectious and occupational exposures, and prior diagnosis of a malignant tumor other than lung cancer. This tool has been validated in a Caucasian population including never and ever smokers up to 79 years of age (cross validation of tool: AUC=0.70).

Results:
We generated risk profiles for female non-smokers between 65-79 years old and no other co-morbidity, and compared the risk against those for women in the same age bracket with 30-pack year smoking history or additional non-tobacco risk factors (i.e. previous pneumonia, asbestos exposure, having a relative with lung cancer < 60years, and the combination of all factors listed). Significant risk with increasing age was predicted by the LLP model for women with 30 pack year smoking history (peak risk at age 75 years 2.2% over next 5 years). This is less than the risk of 6.7% over the next 5 years (at age 75 years) for non-smoking women with the combination of all mentioned risk factors. Relative risk of lung cancer of non-smoking women with all noted risk factors was 3.5 compared to women with no other risk factors other than 30 pack-years smoking history. Relative risk of smoking women compared to non-smoking without other risk factors was 4, while relative risk of non-smoking women with cumulative risk factors was 14 compared to non-smoking women with no other risk factors. Figure 1



Conclusion:
Therefore, the development of lung cancer risk prediction models is a key advance in the assessment of patients at risk. Individual risk assessment can be judged using the LLP model and could encourage refinement of screening recommendations.

Background:
Epidemiological studies consistently show that chronic obstructive pulmonary disease (COPD) is associated with an increased risk of lung cancer among smokers. However, debate exists as to whether there is a linear relationship between the severity of COPD and risk of lung cancer. The National Lung Screening Study (NLST) and it’s sub-study by the American College of Radiology and Imaging Network (ACRIN), provides the means to re-examine these findings. We examined the effect of spirometry-defined COPD (according to GOLD status at baseline), on the risk of lung cancer in the NLST-ACRIN cohort (according to lung cancer incidence), in a large prospective lung cancer screening study of high risk smokers.

Methods:
In the NLST-ACRIN cohort of 18,475 screening participants eligible for the NLST, 6,436 screening participants had COPD (35%) according to baseline pre-bronchodilator spirometry and were followed for a mean of 6.4 years. From this group, 401 lung cancer cases were identified. The 6,436 screening participants with COPD were sub-grouped according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) stages 1 (N=1607), 2 (N=3528), 3 (N=1083) and 4 (211). Lung cancer incidence at the end of follow-up was compared between the GOLD subgroups and those with normal spirometry (N=12,039).

Results:
Compared to those with normal spirometry, where the lung cancer incident rate was 4.63/1000 person years, the lung cancer incident rate was 7.58/1000 person years for GOLD 1, 9.43/1000 person years for GOLD 2, 12.7/1000 person years for GOLD 3 and 15.55/1000 person years for GOLD 4 (all P<0.0001). The lung cancer histology was significantly different, with more squamous and non-small cell cancers in those with COPD but more adenocarcinoma and Bronchoalveolar carcinoma in those with normal lung function (P<0.004). Figure 1



Conclusion:
In a large prospective study of unselected high risk smokers with and without COPD, we report a strong linear association between increasing severity of COPD and increasing lung cancer risk (incidence). This suggests that the risk of lung cancer is greatest in those with the most severe COPD and 3-4 fold greater than those with normal lung function. We also report that lung cancers of more aggressive histology were more common in those with COPD. Funding This study was funded by a grant from Johnson and Johnson and grants U01-CA-80098 and U01-CA-79778 to the American College of Radiology Imaging Network

Background:
Chronic obstructive pulmonary disease (COPD) is a major intermediate phenotype for Lung cancer (LC); the presence of COPD conferring a three- to 10-fold increased risk of LC when compared with smokers without COPD. Variability in lung function and risk for COPD in people with similar cigarette smoking histories, together with studies of familial aggregation, support an important role for genetic factors in COPD. Therefore, we employed a targeted sequencing approach to identify variants associated with susceptibility to COPD. We focused on 107 common susceptibility loci identified in recent genome-wide association studies (GWAS) catalog in LC, COPD, lung function and smoking behavior.

Methods:
We employed an extreme phenotype approach in two carefully phenotyped extreme categories of smokers from the COPDGene study: 1) Long-term smokers with normal lung function defined as post-bronchodilator FEV1 ≥ 80% predicted, FEV1/FVC ≥ 0.7, with smoking histories of 15+ pack-years, considered as resistant to the effects of smoking, n = 318.; 2) Susceptible smokers with severe COPD defined as GOLD Stages 3-4 (post-bronchodilator FEV1 < 50% predicted and FEV1/FVC < 0.7), with smoking histories of 10+ pack-years, n = 309. We performed exome sequencing and analyzed rare (minor allele frequency [MAF] < 0.01 in reference exome databases) substitution and indel variants predicted to be functional in susceptibility loci previously identified by GWAS.

Results:
Our analysis revealed eight potentially causative non-synonymous substitution variants, occurring in 3+ susceptible smokers with COPD, and with none in resistant smokers. The two most intriguing associations were TGM5 Thr42Asn and ZBTB9 Leu43Val, that presented in six and four susceptible smokers with severe COPD, respectively. Moreover, we found an additional TGM5 compound heterozygous mutation, Val202Ile, carried by two severe COPD patients with none in the resistant smokers. TGM5 is located on 15q15.2, a susceptibility locus only reported in LC GWAS, and the p.Thr42Asn in exon 2 is only 1563bp away from the LC GWAS hit (rs504417) in intron 1. ZBTB9 is located on 6p21.31, a locus common to LC, lung function and smoking behavior. Table 1. List of top candidate deleterious mutations in susceptible smokers

Marker	Gene	Protein	# Mutated COPD	# Mutated Control	MAF in KG
rs148913728	TGM5	Thr42Asn	6	0	0.0012
rs144575810	TGM5	Val202Ile	2	0	0.0002
rs41267651	ZBTB9	Leu43Val	4	0	0.0008
rs147018937	NID2	Lys1296Arg	3	0	0.0004
rs147278493	SLC6A18	Gly496Arg	3	0	0.0002
rs116926108	IFIT3	Leu390Arg	3	0	0.0002
rs142934543	LAMA1	Lys2086Thr	3	0	0.0002
rs41316996	DBH	Gly482Arg	3	0	0.0004
rs41298243	MYOF	Phe1400Leu	3	0	0.0002

LC, lung cancer; COPD, chronic obstructive pulmonary disease; SM, smoking behavior; PF, pulmonary function; MAF, minor allele frequency; KG, thousand genome.

Conclusion:
Our targeted exome sequencing results demonstrate highly disruptive COPD risk-conferring TGM5 and ZBTB9 rare mutations that are associated with susceptibility to lung cancer in smokers, and strengthen the concept of a shared genetic link between COPD and LC.

Background:
The National Lung Cancer Screening Trial found, after 6.5 years, a 20% reduction in LC mortality in high-risk patients (pts) screened with low-dose computed tomography compared to chest x-ray. However, LC screening programs (SP) result controversial due to potential cost-effectiveness issues. Familial LC aggregation (fLCa) has been described previously. The estimated relative risk of LC is ∼1.8 for offspring of parents with LC. Linkage analysis has mapped a dominant locus to chromosome 6 in LC pedigrees. Therefore, in this high-risk subpopulation, SP may have clear advantages. This is the first study to investigate the incidence of fLCa conducted in a sEp.

Methods:
Overall, 509 cancer pts of Spanish (n = 473) or Portuguese (n = 36) origin were included in the analysis. A cohort of 236 consecutive pts (cases) diagnosed with LC was studied for family history (FH) of any type of cancer including LC. Another cohort of 273 pts (controls) with similar demographic characteristics diagnosed with cancer types other than LC was also studied for FH of cancer. We investigated whether LC pts show a higher incidence of fLCa than subjects with other solid tumors.

Results:
Among LC pts with a positive FH for LC, 36.7% showed one of their parents as the only LC relative, 26.5% showed one or more siblings, 18.4% one or more either uncle or aunt, 6.1% their grandfather/grandmother and 12.2% other combinations. Regarding the number of relatives affected, in our LC cohort one relative was the most frequent finding with 42/49 pts (85.7%), 2 in 3 cases (6.2%) and > 3 relatives in 4 subjects (8.1%). We studied the overall incidence of any type of family cancer among cases and controls. No differences were found between groups (72.9% vs 67.4%; p = 0.18). However, in our cohort of LC cases, 49/236 pts (20.8%) had a FH of LC in first or second degree whereas among cancer controls only 29/273 pts (10.6%) showed a LC FH (p = 0.002).

Conclusion:
This is the first estimation of LC FH in a non-selected sEp with LC. 20.8% of LC cases showed a positive FH for LC, being significantly higher (twofold) compared to other cancer pts. Therefore, the usefulness of directed SP for subjects with positive FH of LC should be prospectively evaluated and potential genomic drivers studied.

Table 1. Comparison of incidence of any type of familial cancer and fLCa between a cohort of LC patients and a cohort of subjects with other solid tumors
	LC patients	Other solid tumor patients		p value
	N= 236	N=273
Familial cancer (any type) (n (%))
Yes	172 (72.9)	184 (67.4)		0.18
No	64 (27.1)	89 (32.6)
Familial Lung Cancer (n (%))
Yes	49 (20.8)	29 (10.6)		0.002*
No	187 (79.2)	244 (89.4)
*Statistically significance at p < 0.05
fLCa: familial lung cancer; LC: lung cancer

Background:
Lung cancer with different molecular profile behaves in different ways in terms of etiology, clinical characteristics, and prognosis. It implies that fighting against lung cancer should be moving forward in the direction of modifying the screening strategies on genetically-defined high risk groups, initiating early chemoprevention trials on selected high-risk subjects, and developing personalized cancer treatments. Lung adenocarcinoma in never-smokers, more often harboring epidermal growth factor receptor (EGFR) mutations especially in Asians, has a remarkable therapeutic response to specific tyrosine kinase inhibitors. Most single nucleotide polymorphisms (SNPs) identified by candidate gene and genome-wide association studies are of genes involved in carcinogen metabolic pathway, DNA repair pathway, inflammatory pathway and tumor suppressor pathway. Conflicting results are likely due to heterogeneity of the study population and lack of focus on specific molecular subgroups (e.g. EGFR-mutants). We therefore embark on the current study to identify susceptibility genes in a molecularly-defined (EGFR-mutated) subgroup of never-smoking lung adenocarcinoma in Hong Kong population.

Methods:
Eligible patients with confirmed primary lung adenocarcinoma were recruited from Queen Mary Hospital, Hong Kong. Voluntary healthy controls were recruited from blood donors in Hong Kong Red Cross. 10mL venous blood samples were taken from both cases and controls for DNA extraction and SNP assays. 51 SNPs of 14 genes involved in four different pathways were tested using MassARRAY. A structured questionnaire including the information of environmental exposures at home and workplace, family history of lung cancer and other cancer for all subjects, and clinical characteristics (cell type, EGFR mutation status, staging and treatment etc.) for lung cancer patients were administered to cases and controls. Using SNPstats package, logistic regression analysis adjusted for age and gender was performed to evaluate the association between the studied SNPs and lung cancer development.

Results:
From September 2006 to February 2015, a total of 614 lung cancer patients regardless of histological type and smoking status were recruited. Out of which, 267 never-smoking lung adenocarcinoma patients (72% females, mean age 61.6+/-12.6 years) were regarded as cases in the study. From July 2013 to August 2014, a total of 453 healthy controls (40% females, mean age 53.8+/- 8.3 years) were recruited. Most cases (69%) were at advanced stage with chemotherapy treatment (67.8%). Higher proportion of cases (41.7% at home; 35.4% at workplace) than controls (24.9% at home; 26.9% at workplace) had been exposed to second-hand smoke. Genetic analysis was restricted to 184 pairs of age and gender well-matched cases and controls. Two of the 51 SNPs showed a significant association with lung adenocarcinoma, which were rs2069840 of IL-6 gene in inflammatory pathway (OR: 5.80; 96%CI: 1.60-20.99) and rs1106087 of XPC gene in DNA repair pathway (OR: 3.72; 95%CI: 1.40-9.83).

Conclusion:
Our results suggest that IL-6 rs2069840 and XPC rs1106087 are susceptibility genes for development of EGFR-driven lung adenocarcinoma in never-smoking Hong Kong population.

Abstract not provided

Background:
Lung SCC arises in the epithelial layer of the bronchial airways and is preceded by the development of premalignant lesions (PMLs). The molecular events involved in the progression of PMLs to lung SCC are not clearly understood as not all PMLs that develop go on to form carcinoma. Our group is using high-throughput genomic techniques to characterize the process of premalignant progression by examining PMLs and non-lesion areas of individuals with PMLs (“field of injury”) to identify events that lead to the development of SCC. Pathway analysis revealed enrichment oxidative phosphorylation (OXPHOS) /respiratory electron transport among genes up-regulated in the airways of subjects with PMLs. OXPHOS is the most efficient metabolic pathway that generates energy in the form of ATP by utilizing the structures and enzymes of the mitochondria. OXPHOS is often elevated during epithelial tissue repair and is superseded by glycolysis in the development of cancer.

Methods:
mRNA-Seq was conducted on cytologically normal airway epithelium collected from indviduals with (n=50) and without (n=25) PMLs. Linear modeling strategies were used to identify genes altered between subjects with and without PMLs (n=206 out of 13,900, genes at FDR<0.001). Pathway analysis by GSEA revealed enrichment (FDR<0.05) of oxidative phosphorylation (OXPHOS)/respiratory electron transport genes among genes up-regulated in subjects with PMLs. To validate these findings, we examined oxygen consumption rates (OCR) and extracellular acidification rates (ECAR) in primary airway epithelial cells cultures from PMLs and non-lesion areas and cancer cell lines that have high OXPHOS/ moderate glycolytic (H1299), moderate OXPHOS/ high glycolytic (HCC4006) or low OXPHOS/ low glycolytic (H2085) gene expression. In addition, protein expression of genes elevated in the field of injury including, translocase of the outer mitochondrial membrane (TOMM 22) and cytochrome C oxidase (COX-IV) were measured in FFPE sections of human PMLs and PMLs from the N-nitroso-tris-chloroethylurea (NTCU) mouse model of lung SCC.

Results:
OCR and ECAR values in the lung cancer cell lines were consistent with gene expression patterns. Perturbations of OXPHOS resulted in 3 fold (H1299) and 2 fold (HCC4006) higher OCR vales than those in H2085 cells (p<0.05) reflecting higher OXPHOS activity. Whereas the ECAR values were 2.5 fold (HCC4006) and 1.5 fold (H1299) higher than those in H2085 cells (p<0.05), reflecting higher glycolytic metabolism. The OCR and ECAR patterns in the primary premalignant cultures also supported the computational findings in the field of injury of PMLs. The baseline OCR/ECAR values were 1.5 fold higher in the cultures from PMLs compared to non-lesions controls (p<0.001). Additionally the OCR and ECAR values were elevated in response to perturbations in OXPHOS in the PMLs compared to controls. Protein levels of TOMM 22, and COX-IV were found to be elevated in dysplastic lesions compared to controls.

Conclusion:
Together these data suggest that metabolism-associated gene expression is correlated with cellular metabolism and there is an increase in OXPHOS associated with the development of PMLs. Furthermore, there is potential that therapeutically increasing or maintaining OXPHOS in premalignant lesions or the field of injury may be a mechanism of prevention for lung cancer.

Background:
Recently, SPNs have become more frequently encountered in bronchology. Therefore, an efficient and reliable method for detecting SPNs based on their morphological characteristics is needed The aim of this study was to compare the diagnostic value of near infrared (NIR) spectroscopy, in vivo Raman spectroscopy and radial endobronchial ultrasound (EBUS) for solitary pulmonary nodule (SPN). Fluoroscopic guidance with transbronchial biopsy and needle biopsy was performed in all patients.

Methods:
Between February 2014 and February 2015 we examined 22 male and 29 female patients having a median age of 68 years with positron emission tomography-computed tomography findings of metabolically active SPN between 1, 5 to 3 cm in diameter. We used tree types of point monitoring systems. Fluoroscopic guidance (with guide- sheath) was combined with a radial EBUS. In the case radial EBUS conclusively showed catheter position in the centre of SPN (41 cases) than in-vivo Raman Spectroscopy and NIR spectroscopy probes were placed into the guide sheath in order to gather tissue information. Mean measurement time was less than five minutes after establishing ideal position of guide -sheath. Results of in spectroscopy measurements from both Raman spectroscopy and NIR spectroscopy were obtained as differences between spectral characteristics of normal tissue (same side, different lobe) to SPN tissue.

Results:
The results are expressed as sensitivity of RAMAN spectroscopy and NIR spectroscopy towards EBUS navigated biopsies. Statistical analysis of the results showed comparable very high sensitivities for NIR spectroscopy and Raman spectroscopy in confirmation of SPN tissue. From 41 EBUS positive visualisations of SPN there were 38 conclusive histological findings. Both Raman spectroscopy (positive differences in 38 cases) and NIR spectroscopy (positive differences in 36 cases) showed good performance in tissue discrimination.

Conclusion:
Every confirmatory method brings different information about tissue. EBUS describes volume of the SPN and gives valuable information about the position of catheter in the SPN. Raman spectroscopy and NIR spectroscopy bring information about biochemical/ optical characteristics of the tissues.

Background:
Probe-based confocal laser endomicroscopy (pCLE) allows for real-time noninvasive histological imaging via bronchoscopy. Interpreting pCLE images and correlating with pulmonary disease remains challenging. We performed an in vivo study to evaluate the correlation between pathological diagnosis and pCLE imaging of pulmonary disease.

Methods:
We sequentially enrolled the patients with undiagnosed lung lesion, and randomly grouped into control group (TBLB and peripheral EBUS) and pCLE group (TBLB + pCLE and peripheral EBUS + pCLE). pCLE was performed with Cellvizio system (Mauna Kea Technologies, Paris, France). All patients were consent to the procedure. Pathologists and pulmonologists reviewed the images by the Columbus Classification (CC). Questionnaires were applied post the procedure to collect patients’ condition. We developed a computer assistant diagnostic (CAD) system to calculate alveolar diameter, vessel diameter and optical density percentage and compare the CAD diagnostic accuracy with CC standard. The CAD system involved image processing methods to calculate the diameters in pixel domain and then transformed them into the real value. Pseudo-color processing was used to show the density percentage of different tissues. And the histogram was also calculated to figure out the distribution alone gray scale.

Results:
258 patients enrolled in the study, 98 under pCLE examination, while 160 under control group. Among them 128 lesions were diagnosed as malignant tumor by pathological diagnosis, 87 cases were diagnosed as benign disease. Primary features were observed in the samples using pCLE in the lesion of cancer: The normal alveolar in malignant nodules is smaller than benign nodules. While, the vessel in the malignant nodules is thicker than the benign ones. The cellular structure and vessel domination in various subtypes of lung cancer is different. There was no significance on procedure time between control and pCLE group, as well as patients’ secretion, tolerance and willing for repeat examination.

Conclusion:
pCLE can identify lung carcinoma in in vivo procedure with well tolerance and with limit procedure time. As a non-invasive method, pCLE could improve accuracy and avoid unnecessary biopsy. The Computer Assist Diagnosis system could help pulmonologists to better acquire the right image and to differentiate diseases on the site.

Background:
Lung squamous carcinoma is believed to arise from premalignant bronchial epithelial dysplasia, which demonstrates progressive histologic changes leading up to invasive cancer. However, only a small subset of these lesions progress to carcinoma. Recent studies have shown that somatic chromosomal alterations (SCAs) status is a better biomarker than premalignant histology alone. Single-nucleotide polymorphism microarray (SNP array) has been frequently used to delineate these genomic alterations across the whole genome. However, the cellular heterogeneity, from clinical samples such as endobronchial specimens, is a basic obstacle to perform sensitive and accurate detection of SCAs.

Methods:
We used: 1) a lung cancer cell line (NCI-H1395) and its matched lymphoblastoid (NCI-BL1395) cell line; 2) frozen lung tissues containing different percentage of invasive cancer cells surgically resected from a patient; and 3) biopsies and brushings obtained at the visually concerning areas during bronchoscopy. The histology of the clinical samples were graded by the study pathologist. Genomic DNA was isolated from each sample, quantified, and labeled for Illumina SNP array (HumanOmni 2.5-Quad BeadChip). Data analysis and visualizations were performed using Partek Genomic Suite 6.6 software.

Results:
Our study focused on the detection of SCAs by the comparison of genomic DNAs from cancer/premalignant cells (subject) to blood/normal cells (reference) from the same individual. We tested a B allele frequency metric, the subtraction of allelic fractions (delta-θ), on a standardized mixture of genomic DNAs from a lung cancer cell line and its matched lymphoblastoid cell line. Delta-θ proved to be a sensitive parameter to clearly delineate SCAs present in the tumor cell line even with a large proportion of normal cells (up to 90%). To explore the utility of using delta-θ for heterogeneous samples, we used clinical lung cancer specimens with known cancer cell content. In comparison to the other publicly available analytical metrics/algorithms (conventional Log R Ratio plot, mirrored B Allele Frequency plot, and GAP algorithm), delta-θ performed as well or better (with lower computational power needed), and enabled the detection of SCAs even in highly heterogeneous clinical samples (<30% tumor cell content). In addition, we completed a study using a number of bronchial biopsies and brushings with histologic grade ranging from normal to squamous cell carcinoma. SCAs were rarely detected in those of low to mild dysplasia, while they were detected in approximately 25% of moderate or severe dysplasia, and in all carcinoma in situ (CIS) and squamous cell carcinoma specimens. Longitudinal, repeated samplings from a high risk patient who persistently showed high grade dysplasia across the bronchus, revealed that delta-θ could identify SCAs continuously across the whole genome. The fact this individual had highly overlapping SCAs between different bronchial locations indicates genomic field cancerization may occur, along with the histological field effect in premalignant epithelium.

Conclusion:
In SNP microarray studies, delta-θ is a highly sensitive metric for detecting SCAs even in heterogeneous dysplastic bronchial specimens. SNP array may be a powerful tool to understand premalignant genetic alterations and field cancerization.

Abstract not provided

Background:
Lung cancer is the result of a multi-step accumulation of genetic and/or epigenetic alterations; therefore, a better understanding of the molecular mechanism, by which these alterations affect lung cancer pathogenesis, would provide new diagnostic procedures and prognostic factors for early detection of recurrence. In this regard, many have studied molecular or other markers in pre-neoplastic and neoplastic lesions to discover what might relate to tumor recurrence and shortened survival.

Methods:
A series of 136 lung/bronchial and lung parenchyma tissue samples from 136 patients consisting of basal cell hyperplasia, squamous metaplasia, moderate dysplasia, adenomatous hyperplasia, severe dysplasia, squamous cell carcinoma and adenocarcinoma were analyzed for the distribution of hyaluronidase 1 (HYAL1) and 3 (HYAL3), and hyaluronan synthases 1 (HAS1), 2 (HAS2) and 3 (HAS3) by immunohistochemistry.

Results:
HYAL 1 was significantly more expressed in basal cell hyperplasia compared to moderate dysplasia (p=0.01), atypical adenomatous hyperplasia (p=0.0001) and severe dysplasia (p=0.03). A lower expression of HYAL 3 was found in atypical adenomatous hyperplasia compared to basal cell hyperplasia (p=0.01) and moderate dysplasia (p=0.02). HAS 2 was significantly higher in severe dysplasia compared to basal cell hyperplasia (p=0.002), and equally higher in squamous metaplasia compared to basal cell hyperplasia (p=0.04). HAS 3 was significantly expressed in basal cell hyperplasia compared to atypical adenomatous hyperplasia (p=0.05) and severe dysplasia (p=0.02). A lower expression of HAS 3 was found in severe dysplasia compared to squamous metaplasia (p=0.01) and moderate dysplasia (p=0.01). Epithelial HYAL 1 and 3 and HAS 1, 2 and 3 expressions were significantly increased in pre neoplastic lesions compared to neoplastic lesions. Comparative Cox multivariate analysis controlled by N stage and histologic types of tumors showed a significant association between poor survival and high pre neoplastic cell associated to HAS3 (HR=1.19; p=0.04).

Conclusion:
We concluded that localization of HYALs and HASs in lung/bronchial pre-neoplastic and neoplastic lesions was inversely related to malignancy, these factors emerging as potentially important diagnostic markers in patients with suspicion of lung cancer.

Background:
The extracellular sulfatases (SULF1 and SULF2) are overexpressed in a wide assortment of human cancers. SULF2, in particular, has been shown to drive carcinogenesis in non-small cell lung cancer (NSCLC), malignant astrocytoma, and hepatocellular carcinoma. As extracellular enzymes that are both tethered to the cell membrane and secreted, the SULFs and their heparan sulfate proteoglycan (HSPG) substrates are present in the extracellular environment. We hypothesize that the blood levels of SULF2 can serve as biomarkers for the early detection of NSCLC and malignant astrocytoma. The primary goal of this study is to evaluate the patient tumor and blood samples for the presence of the SULF2 in order to develop novel biomarkers for the early detection of NSCLC.

Methods:
We identified patients who underwent lung resection for adenocarcinoma (ADC) (41 patients) or squamous cell carcinoma (SCC) (51 patients) at our institution from 2000 to 2006. We excluded patients with recurrent lung cancer, or less than 3 mm of invasive tumor on H&E slide. A section from each paraffin-embedded tissue specimen was stained with a monoclonal antibody to SULF2. A pathologist determined the percentage (0-100%) and intensity (0-3) of tumor cells staining. Survival analysis was performed using a multivariate Cox proportional hazards model. We developed an ELISA to detect SULF2 in human blood. After testing a number of different strategies including using different combinations of our anti-SULF2 mAbs, we determined that a sandwich ELISA with capture mAb 5C12 followed by detection with biotinylated mAb 8G1 was best for the most sensitive detection of SULF2.

Results:
SULF2 staining (either tumor or stroma) was positive for 82% of the samples The SCC samples had a higher mean percentage of tumor staining compared to the ADC samples (100% vs. 60%; p<0.0005). However, after adjusting for age, sex, race, histologic type, stage, and neoadjuvant therapy, there was no significant association between percentage of SULF2 tumor staining and overall survival. Nonetheless, these initial findings are very encouraging, because the vast majority of ADC samples, including early stage disease, and all of the SCC tumor samples have some degree of staining for SULF2 protein. Using our SULF2 ELISA assay, we analyzed plasma samples from 54 healthy donors and 85 patients with newly diagnosed early stage NSCLC before surgical resection. The level of SULF2 protein is significantly higher in patients with NSCLC compared with healthy controls (738.4 ± 55.17, vs. 439.4 ± 40.88 pg/mL; p<0.0001).

Conclusion:
SULF2 protein was detected in the vast majority of tumor and blood samples of patients with lung cancer. Although additional studies are required, these data provide the first indication that SULF2 blood level may be a useful biomarker for the early detection of lung cancer.

Background:
Long non-coding RNAs (lncRNAs) are new-founding RNAs which could regulate many biological processes. Our previous study shown that lncRNA-GAS5 was decreased in lung cancer tissue, which contributed to the proliferation and apoptosis of non-small lung cancer (NSCLC). GAS5 was also associated with the prognosis of lung cancer patients. However, the plasma samples were more easily available than the tissue sample in the clinic. And the expression of GAS5 in the plasma of NSCLC patients was unknown.

Methods:
90 patients with NSCLC and 33 health controls were included in our study. Blood samples were collected before surgery and therapy. We extracted the free RNA in the plasma and analyzed the expression of GAS5 with quantitative reverse transcription polymerase chain reaction (qRT-PCR). Suitable statistics methods were used to compare the plasma GAS5 levels between the NSCLC patients and health controls, preoperative and postoperative plasma samples. Receiver operating characteristic (ROC) analysis was used to evaluate the diagnostic sensitivity and specificity of plasma GAS5 in NSCLC.

Results:
The 2[-][△][CT ]of GAS5 in the plasma of NSCLC patients and health controls are 1.053774 and 3.019817, respectively. GAS5 in NSCLC plasma was down-regulated compared with health controls (P=0.001), which was significantly correlated with TMN stage (P=0.024). Furthermore, plasma GAS5 increased markedly on day 7 after surgery compared with preoperative levels in NSCLC patients (P=0.003). The CT values of preoperative and postoperative are 2.225909 and 1.050455, respectively. The area under the ROC curve of GAS5 was up to 0.832. The combination of the GAS5 and CEA could produce 0.909 area under the ROC curve in distinguishing NSCLC patients from control subjects (95% CI 0.857–0.962,p=0.000).These results indicated that lncRNA GAS5 may be a more precise biomarker in NSCLC.

Conclusion:
We have demonstrated that GAS5 was decreased in NSCLC plasma expression and the plasma samples were more easily available than the tissue sample in the clinic. So GAS5 could be ideal biomarkers for the early diagnosis of NSCLC.

Abstract not provided

Background:
Several 2-dimensional computed tomography (CT)-based evaluation methods of small-sized lung adenocarcinomas have been reported as predictors of the disease invasiveness. They include the ratio of the maximum diameter of consolidation to the maximum entire tumor diameter (C/T ratio), tumor shadow disappearance rate on mediastinal window images (TDR), and visual estimation of the ratio of ground-glass opacity area (GGO-R). However, these measurements can be poorly reproducible due to possible inter-observer discrepancy, and can be unrepresentative because measuring is done only on one section of a lesion. We have developed a 3-dimensional quantitative entire-nodule evaluation method using novel image-analysis software. The aim of this study is to compare the new method to these 2-dimensional evaluation methods as a predictor of small-sized invasive lung adenocarcinomas.

Methods:
There were 101 consecutive patients with clinical stage IA adenocarcinoma of the lung who underwent complete resection between 2002 and 2005 at our institution, excluding patients undergoing preoperative treatment and those with multiple lung nodules or with a past history of other cancers. Of them, 75 had a nodule separated from the chest wall and mediastinum depicted on preoperative thin section CT scan without contrast enhancement, and they were the subject of this study. The reconstruction interval of the CT scans was 0.2mm and the reconstructed slice thickness was 0.5mm. The image analysis software recognizes a nodule as a collection of cubic voxels. Ground glass opacity (GGO) was defined as the area of increased attenuation in the lung with preservation of the bronchial and vascular margins. As the average CT value of pulmonary arteries on non-contrast-enhanced CT was 50 Hounsfield Unit (HU), we measured the percentage of the voxels over 50 HU in a nodule to identify voxels representing solid component, and the percentage was defined as R-50. Invasive cancer was defined as a nodule with pathological lymphatic permeation, vascular invasion or node involvement. The correlation between invasive lung cancer and clinicopathological factors, including the image findings (C/T ratio, TDR, GGO-R and R-50) was evaluated using multivariate analysis. The areas under the curve (AUC) of receiver operating characteristics curves were compared among the image evaluation methods.

Results:
There were 17 invasive cancers. C/T Ratio, TDR, GGO-R and R-50 were independent predictors of invasive lung cancers (p<0.01). R-50 was equivalent in AUC to the other evaluation methods (AUC: R-50, 0.807; C/T Ratio, 0.800; TDR, 0.809; GGO-R, 0.792, respectively).

Conclusion:
Our new 3-dimensional quantitative evaluation method using image-analysis software had invasive cancer predictability similar to the other 2-dimensional evaluation methods. As this method enables entire-tumor evaluation quantitatively and objectively, it should be more reproducible and reliable than the conventional methods.

Background:
Lung cancer screening by low-dose computed tomography (LDCT) is now recommended for high-risk individuals by US guidelines. New nodules detected after initial baseline screening may complicate management. So far, reported results of new nodules have been inconsistent as different definitions were used. The purpose of this study was to determine the occurrence of new solid nodules and their respective lung cancer rate at incidence screening rounds of the Dutch-Belgian Randomized Lung Cancer Screening Trial (NELSON).

Methods:
The NELSON trial was approved by the Dutch Ministry of Health. All participants gave written informed consent. In total, 7,557 individuals underwent baseline LDCT screening. Incidence-screening rounds took place after 1, 3 and 5.5 years. This study included participants with solid non-calcified nodules, newly detected after baseline and also in retrospect not present on any previous screen. Lung cancer diagnosis was based on histology, and benignity was based on either histology or a stable size for at least two years. Nodule volume was generated semi-automatically by Lungcare software (Siemens, Erlangen, Germany), and the nodule detection limit was 15mm[3].

Results:
During the incidence screening rounds, 1,484 new solid nodules were detected in 949 participants (77% male), with a median age of 59 years (interquartile-range 55-63 years). At the second screening round (1 year after baseline), at least one new solid nodule was present in 4.7% (344/7,295) of participants, and at the third screening round (2 years after the second screening round) additional new nodules were found in 7.1% (491/6,922) of participants. Eventually, a new solid nodule was proven to be lung cancer in 7.9% (75/949) of participants with new solid nodules (77 cancers). A higher number of pack-years smoked increased the risk of a new nodule being cancer significantly (P=0.004). Age and gender distribution were comparable between participants with and without lung cancer detected in a new solid nodule (P=0.236 and P=0.157 respectively). The majority of cancers was diagnosed at stage I (48/77 [62.3%]). Most of the lung cancers were adenocarcinoma (30/77 [39.0%]), squamous cell carcinoma (20/77 [26.0%]) or small cell lung cancer (9/77 [11.7%]).

Conclusion:
New solid nodules are common findings in LDCT lung cancer screening and possess a comparably high risk of malignancy. Guidelines may need to consider a more stringent follow-up for new nodules. More research concerning new nodules is necessary to determine a sufficient follow-up strategy and evaluate distinguishing nodule features of benign and malignant new nodules.

Background:
Currently, there is little known about prevalence of multi-nodularity in a high risk screening population. Radiologists often find more than one nodule per screenee. Whether the number of lung nodules plays a role in the probability of lung cancer, remains still largely unknown.

Methods:
In the Dutch-Belgian randomized lung cancer screening trial (NELSON), launched in 2003, participants were selected with at least one non-calcified nodule at baseline. The NELSON trial was approved by the Ministry of Health and the ethics board of each participating center. All participants gave written informed consent. The per-participant number of baseline nodules was determined. The probability of lung cancer was compared for categories based on the number of baseline nodules, using chi-square testing. Lung cancer diagnosis was confirmed by histology. Nodules were classified as benign if they did not show growth for up to six years after baseline.

Results:
3,392 participants (84,4% male, median age 58 years, median pack years 37,9) with 7,258 nodules at baseline CT screening were included. Of these 3,392 screenees, 1,746 (51,5%) had one nodule, 800 (23,6%) had two nodules, 354 (10,4%) had three nodules, 191 (5,6%) had four nodules, and 301 (8,9%) had five or more nodules. The probability of lung cancer was 61/354 (3.5%) in subjects with one nodule, 37/800 (4.6%) in those with two nodules, 17/354 (4.8%) for three nodules, 12/191 (6.3%) for four nodules and 10/301 (3.3%) when a participant had over four nodules (p=NS). In the baseline screening round, 62 subjects had a malignant nodule. Lung cancer diagnosis was made in the nodule with the largest volume in 60/62 (96.8%) cases. Overall, lung cancer was diagnosed in 137/3,392 subjects (4.0%) in whom nodules were found at baseline. Mean nodule count in screened subjects with only benign nodules was 2.1±1.8, compared to 2.3±2.2 in those with a malignant nodule.

Conclusion:
At baseline CT lung cancer screening, nearly half of screened participants with lung nodules have more than one lung nodule. Nodule count did not have predictive value in the determination of lung cancer probability in lung cancer screening participants. In the first screening round, of all detected nodules per screenee, lung cancer was detected most frequently in the nodule with the largest volume.

Background:
We have previously reported a practical predictive tool that accurately estimates the probability of malignancy for lung nodules detected at baseline screening LDCT (New Engl J Med. 2013;369:908-17). Manual measurement of nodule dimensions and generation of malignancy risk scores is time consuming and subjected to intra- and inter-observer variability. The goal of this study is to prepare a nodule malignancy risk prediction model based on automated computer generated nodule data and compare it to an established model based on radiologists’ generated data.

Methods:
Using the same published PanCan dataset (New Engl J Med. 2013;369:908-17) with the number of lung cancers updated, we prepared a logistic regression model predicting lung cancer using computer-generated imaging data from the CIRRUS Lung Screening software (Diagnostic Imaging Analysis Group, Nijmegen, The Netherlands). Ninety-one of the 2,537 baseline (first) scans were not available or could not be processed by CIRRUS. The remaining 2,446 scans were first annotated by the CIRRUS software. A human non-radiologist reader then accepted/rejected the annotated marks and manually searched the LDCT for nodules missed by CIRRUS or the study radiologist. New nodules found that were not recorded by the study radiologist were reviewed by a subspecialty trained chest radiologist with 14 years experience in lung cancer screening (JM). Nodule morphometric measurements (maximum and mean diameter, volume, mass, density) and total nodule count per scan irrespective of size were automatically generated by the CIRRUS software. The nodule type (nonsolid, part-solid, or solid), nodule description (lobulated, spiculated or well defined) and nodule location (upper versus middle or lower lobe) were manually entered. The variables were evaluated in models as untransformed and natural log transformed variables. Nonlinear relationships with lung cancer were also evaluated. Socio-demographic and clinical history predictors were not included in the model.

Results:
Radiologists evaluation identified 8,570 pulmonary nodules of any size in 2063 individuals - 124 nodules in 119 individuals were diagnosed as cancer in follow-up. Based on CIRRUS software annotated marks that were accepted by a human reader, computer analysis identified 11,520 pulmonary nodules in 2174 individuals - 121 nodules in 115 individuals were diagnosed as cancer in follow-up. Thirty-six percent of new nodules found by CIRRUS and/or second human reader were ≥4 mm (mean±SD, 5.9± 3.5 mm). Both the computer generated imaging data model (Model-CIRRUS) and the radiologist generated data model (Model-RAD) demonstrated excellent discrimination and calibration. Their predictive performances were also similar. Comparing Model-CAD to Model-RAD, the AUCs were 0.9537 versus 0.9541, the 90[th] percentile absolute errors were 0.0008 versus 0.0007, and the Brier scores were 0.0093 versus 0.0137. Mean nodule diameter is a better risk predictor than maximum nodule diameter, nodule density or mass.

Conclusion:
The predictive performances of computer and radiologist generated data models were similar. The model can be integrated to the CIRRUS Lung Screening software to automatically generate a nodule malignancy risk score to facilitate nodule management recommendation. Supported by the Terry Fox Research Institute, The Canadian Partnership Against Cancer and the BC Cancer Foundation on behalf of the Pan-Canadian Early Detection of Lung Cancer Study Group.

Background:
The recommendation by the US Preventive Services Task Force to screen high-risk smokers with low-dose computed tomography (LDCT) and the recent decision by the Centers for Medicare and Medicaid Services to fund LDCT screening under the Medicare program mean that LDCT screening will be implemented at the population level in the US and likely in other countries. With the large volume of scans that will be generated, accurate and efficient interpretation of LDCT images is key to providing a cost-effective implementation of LDCT screening to the large at risk population. Objective To evaluate an alternative workflow to identify and triage abnormal LDCT scans in which a technician assisted by Computer Vision (CV) software acts as first reader with the aim to reduce workload, improve speed, consistency and quality of interpretation of screening LDCT scans.

Methods:
A test dataset of baseline Pan-Canadian Early Detection of Lung Cancer Study LDCT scans (New Engl J Med. 2013;369:908-17) was used. This included: 136 scans with lung cancers, 556 scans with benign nodules and 136 scans without nodules. The scans were randomly assigned for analysis by the CV software (CIRRUS Lung Screening, Diagnostic Imaging Analysis Group, Nijmegen, The Netherlands). The annotated scans were then reviewed by a technician without knowledge of the diagnosis. The scans were classified by the technician as either normal (no nodules or benign nodules only, potentially not requiring radiologist review) or abnormal (suspicious of malignancy or other abnormality requiring radiologist review). The results were compared with the Pan-Can Study radiologists. Nodules found by CIRRUS but not by the radiologist were reviewed by a subspecialty trained chest radiologist with 14 years experience in lung cancer screening (JM).

Results:
The overall sensitivity and specificity of the technician to identify an abnormal scan were: 97.7% (95% CI: 96.3 - 98.7) and 98.0% (95% CI: 89.5 - 99.7) respectively. The technician correctly identified all the scans with malignant nodules. The time taken by the technician to read a scan was 208±120 sec.

Conclusion:
A technician assisted by CV software can categorize accurately abnormal scans for review by a radiologist. Pre-screening by a technician and CV software is a promising strategy for reducing workload, improving the speed, consistency and quality of scan interpretation of screening chest CTs. Supported by the Terry Fox Research Institute, The Canadian Partnership Against Cancer and the BC Cancer Foundation on behalf of the Pan-Canadian Early Detection of Lung Cancer Study Group.

Abstract not provided

Background:
Navigational bronchoscope including conventional electromagnetic navigation bronchoscope(ENB) and endobronchial ultrasound(EBUS) with a guide sheath(GS) for transbronchial lung biopsy (TBLB) has improved the diagnostic outcome for peripheral pulmonary lesions (PPLs). However, ENB required the bronchoscope for large diameter of the working channel(>2.6mm) which could limit the deep of the insertion and EBUS-GS could be regarded as the confirmation tool other than navigation system. A new, realtime electromagnetic guidance system for bronchoscopy using a thin bronchoscope(4.0mm) with a GS(1.95mm) for TBLB is a novel method to increase diagnostic yield of PPLs.

Methods:
A prospective, open label, two centers, randomized controlled pilot study involves two diagnostic arms: ENB-GS-TBLB and traditional GS-TBLB which was conducted to determine the ability and safety.ENB-GS-TBLB is performed using an electromagnetic navigation system with a GS and an internal locatable guide with diameter of 1.45 mm. Primary outcome was diagnostic yield. Secondary outcomes were yields by total procedure time, the time for finding lesions and the X-ray time using during operation. Complications were also documented.

Results:
Of the 86 patients recruited, 81 had a definitive histological diagnosis and were included in the final analysis. The diagnostic yield of the ENB-GS-TBLB (87%) was greater than GS-TBLB (64%; p<0.05). The time for finding lesions of the ENB-GS-TBLB(3min 43s) was significantly less than GS-TBLB(4min 44s; p<0.05). ENB-GS-TBLB was independent of lesion size or lobar distribution. No complications were found in both two groups.

Conclusion:
ENB-GS-TBLB seems to be an accurate and safe procedure. It allowed us to improve the diagnostic yield of flexible bronchoscopy in PPLs.

Background:
The National Lung Screening Trial (NLST) showed that chest CT screening for patients at risk for lung cancer reduces lung cancer mortality compared to Chest X-Ray (CXR) screening, but with considerable costs due to a high rate of false positive findings. The use of FDG PET has been advocated as a diagnostic tool to aid clinicians in evaluating nodules that may or may not be cancer, but no investigations to date have ascertained current practice patterns in a large group of patients across the U.S.

Methods:
Using data from the NLST, we determined the appropriateness and characteristics of diagnostic FDG PET use in patients with an abnormal finding (defined as a ≥ 4 mm nodule) during lung cancer screening via CT or CXR. Diagnostic FDG PET consisted of either a PET alone or combined PET/CT, which was done prior to a lung cancer diagnosis but after an abnormal finding. Appropriateness was defined as diagnostic FDG PET use for nodules ≥ 8 mm. We used multivariable logistic regression techniques to assess factors associated with diagnostic FDG PET use.

Results:
Of 9,964 patients with an abnormal finding during any of the three rounds of screening, 1,206 (12%) had a diagnostic FDG PET scan at 33 different medical centers across the U.S. (Table 1). Forty percent (n = 484) of these scans were recommended by a radiologist as a follow-up for an abnormal finding. Twenty-seven percent (n = 331) were performed for nodules less than 8 mm, and of these 24% (n = 81) were recommended by radiologists. There were no regional differences in PET use across U.S. areas with endemic fungal disease but patients from the Northeast and Southeast were twice as likely as the West to have a PET scan after a positive screen. Older age, nodule size ≥ 0.8 –2.0 cm, upper lobe location and a spiculated nodule border were associated with increased diagnostic FDG PET use.

Conclusion:
This is the first study to describe differential FDG PET use across the U.S and by medical specialty. Importantly, PET imaging was used inappropriately for small nodule evaluation in one out of four cases. Future studies should characterize associated costs and whether better adherence to current national guidelines can reduce such costs. Figure 1

Background:
CT-guided cutting needle lung biopsy is important for the diagnosis of lung cancer. The co-axial method is now widely used. However, co-axial method failed to decrease the incidence of pneumothorax. This study is to investigate whether our new-developed “Liquid withdraw” technique (to inject small amount of lidocaine during withdraw of the needle) can reduce incidence of pneumothorax when combined with co-axial technique. Figure 1 Fig.1.What is liquid withdraw



Methods:
From Jan 2013 to Dec 2014, We retrospectively studied 38 CT-guided percutaneous lung biopsy using co-axial and liquid withdraw techniques. The pathologies and complications secondary to biopsy procedure (pneumothorax, bleeding and hemoptysis) were noted. Pneumothorax was graded as mild, moderate, and severe.

Results:
37 cases was diagnosed out of 38 biopsies, of which 23 cases were adenocarcinoma (21 patients consented EGFR mutation test, and 15 cases had EGFR mutaions)，2 squamous cell carcinoma, 1 non-small cell lung cancer (cannot be further classified after IHC), 1 small-cell lung cancer, 2 primary lung cancer of other types, 5 metastatic lung cancer and 3 benign diseases. 4 cases (10.5%) happened pneumothorax (all were mild pneumothorax)，bleeding during biopsy happened in 1 (2.6%) case, 6 cases with a small amount of hemoptysis（15.8%）. No infection, tumor implantation or aeroembolism happened. Figure 1 Fig.2. A case of CT-guided cutting needle lung biopsy using “Liquid Withdarw” technique A: The co-axial inducer needle is located at the margin of the lesion B&C:after biopsy, the lidocaine can be seen in the needle passage and no pnemothorax is found.



Conclusion:
CT-guided percutaneous lung biopsy using co-axial and liquid withdraw is an accurate, safe，reliable technique. Compared to co-axial technique without liquid withdraw, the incidence of pneumothorax was reduced from approximately 35% to 10.5%. More studies according to liquid withdraw technique will be conducted in our future work.

Abstract not provided

Background:
Rates of resection of non-malignant lung nodules suspected pre-operatively to be lung cancer vary widely and are reported to be as high as 40%. Commonly used modalities in the pre-operative workup of new lung nodules suspicious for lung cancer include positron emission tomography (PET), bronchoscopy, and computed tomography (CT)-guided fine needle aspiration (FNA). We evaluated the non-malignant resection rate (NMRR) and the frequency of benign resections among patients with pre-operative FNA in our lung cancer center.

Methods:
The study population was identified using databases of the Mount Sinai Departments of Thoracic Surgery and Radiology. Eligible patients included those with a CT-guided FNA and/or surgical resection performed during the 12-month period between July 2013 – July 2014 for known or suspected first primary early stage lung cancer presenting with a lung nodule or mass. Cases were included if patients were >18 years of age with no history of cancer treated within 5 years. Patient data were abstracted from the electronic medical records.

Results:
A total of 283 nodules from 264 patients met inclusion criteria. Of these, FNA was performed in 217 (77%) of the 264 patients, with 131 results (60%) categorized as malignant. Similarly, 228 nodules (81%) were PET imaged, and 141 (62%) of these were positive (Standard Uptake Value >2). Sensitivity and specificity of FNA and PET for diagnosis are reported in Table 1. Post-FNA pneumothorax requiring a chest tube occurred in 11/193 FNAs performed at Mount Sinai (6%). Of 208 surgically resected nodules, 27 cases (13.0%) had a non-malignant diagnosis on pathologic examination. The non-malignant resection rate (NMRR) ranged from 0% to 39% by different surgeons and did not correlate with surgical case volume. Among the 142 resections preceded by FNA, 11 (7.7%) were found to have non-malignant pathology. In contrast, among the remaining 66 resections without a pre-operative FNA, 16 (24.2%) were benign (OR 3.81, 95%CI 1.52-9.69; p = 0.001). Figure 1



Conclusion:
In this single center retrospective analysis, the overall NMRR was lower than in previously published reports. Furthermore, the NMRR was significantly lower in thoracic operations preceded by a CT-guided FNA compared with those without a pre-operative FNA. Diagnostic accuracy of FNA in this cohort of patients at moderate to high risk for lung cancer is higher than that of PET, with an acceptably low complication rate. These findings suggest that pre-operative diagnostic confirmation by FNA results in a low rate of non-malignant resection.

Background:
This analysis compares the cost–effectiveness of radial endobronchial ultrasound (EBUS) and electromagnetic navigation (ENB) to transthoracic needle biopsy (TTNB) to achieve diagnosis of suspicious lung lesions. As more centers develop lung screening and lung nodule clinics, there will be a need for diagnosis of small pulmonary nodules. The National Lung Screening Trial (NLST) showed a probable positive Low Dose Computer Tomography (LDCT) incidence of 25%, indicating the number of patients requiring a diagnostic evaluation will increase. The expectation of increased lung cancer screening due to the Centers of Medicare and Medicaid Services approving coverage for LDCT for eligeable patients, and the American College of Chest Physicians (ACCP) recommendation for improved techniques to diagnose peripheral lung lesions necessitate utilizing clinical and economic assessment tools that support clinical decisions. The study seeks to identify the most cost-effective biopsy protocol to reduce costs, and deliver improved diagnostic accuracy.

Methods:
The study reviewed the NLST which enrolled over 50,000 people aged 55–74 years with at least a 30-pack/year smoking history, in fairly good health and non-symptomatic of lung disease. The study found low-dose computed tomography of the chest resulted in a 20% lower mortality from lung cancer compared with those who had chest x-rays. Approximately 25% of the LDCT-screened patients had a positive screen requiring confirmation of the lung lesion. Using an estimate of 5.2 million annual chests CT scans in the USA as a basis for the number of patients seeking a confirmative result before being recommended for surgery for possible benign lung lesions. The cost–effectiveness models employed estimate the direct costs to the hospital and to the patient. Direct costs were calculated using the Medicare Median cost files by Current Procedural Terminology (CPT) code for 2012. Additional costs were added to account for the fee of the procedure room, nursing and clinical support staff, and observation room time. Reimbursement reflects the 2013 Medicare allowable payment exclusive of the geographic adjustment factor. Reimbursement from commercial insurers is constructed upon a conservative multiple of the Medicare allowable. CPT codes subject to the multiple procedure discounts were properly reduced by 50% as they would be for reimbursement purposes for both Medicare and commercial insurance reimbursement.

Results:
Modeling 200 representative patients delegated to; TTNB, bronchoscopy, or R-EBUS-/ENB-enabled endobronchial percutaneous (Endo-Perc) when comparing procedure fees, insurance payment and clinical outcomes, the Endo-Perc technique lead to the most cost-effective option to biopsy the lung lesion. The lower adverse event profile of pneumothorax and reduced cost exhibited by the Endo-Perc procedure; resulted in a benefit of $130,464 compared with a loss of $562,863 for TTNB, or a loss of $103,487 for routine bronchoscopy.

Conclusion:
The results suggest combining R-EBUS with ENB provides a high-diagnostic yield at a lower cost due to the lower risk of a pneumothorax when compared with transthoracic lung biopsy.

Background:
In lung cancer resection, COPD is a risk factor for postoperative complications. There are few reports about postoperative complications that assume a pictorial emphysematous change an index. We examine a relationship of an emphysematous regional ratio in preoperative CT in patients with COPD who underwent lung cancer resection and cardiopulmonary complication.

Methods:
One hundred fifty-nine patients with COPD who underwent lobectomy for lung cancer in our hospital from 2002 to 2011 were retrospectively evaluated in this study. Preoperative factors, including proportion of emphysematous area measured by CT (percentage of low attenuation area: LAA%), and operative factors were analyzed. Cardiopulmonary complications include pyothorax, pneumonia, atelectasis, acute pulmonary injury, chest tube indwelling, O~2~ long supply and arrythmia.

Results:
Cardiopulmonary complications were observed among 61 patients (38%). Ages, FEV1.0%, LAA% and amounts of blood lost were significantly relevant to cardiopulmonary complications by univariate analysis. Multivariate analysis indicated that patient’s age and LAA% could be significant independent predictors. Table1.Complications incidence by LAA%

LAA%	N	complications：n=61	no complications：n=98	p value
～1% 1～10% 10%～	77 67 15	15(19.5%) 37(55.2%) 9(60.0%)	62(80.5%) 30(44.8%) 6(40.0%)	<0.001

Table2. Operative factors in relation to cardiopulmonary complications

variables	Odds ratio	95%Confidence Interval	p value
Age(>70 years) FEV1.0% GOLD PaO2 LAA%（1%～） blood lost（>150ml）	4.612 1.042 2.044 0.973 5.570 2.073	2.028-10.489 0.973-1.117 0.857-4.876 0.940-1.008 2.302-13.480 0.878-4.894	＜0.001 0.242 0.107 0.128 ＜0.001 0.096

Conclusion:
LAA% is useful for predicting cardiopulmonary complications in patients with COPD undergoing lobectomy for lung cancer.In patients with COPD undergoing lobectomy for lung cancer, 70 years of age or older, the LAA% 1% or more of the cases, more careful intraoperative, and postoperative management are required.

Background:
Lung cancer is one of the most difficult cancers to diagnose in primary care settings. Lung symptomatology alone is a poor indicator of likelihood of diagnosis and in Australia, most primary care practitioners (general practitioners, GPs) will see only one or two lung cancer cases annually. Early diagnosis leads to improved survival with 5-year lung cancer survival much higher in localised disease (30% versus 16% overall). Australia has a mixed public-private model of health services and referral pathways from primary to secondary and tertiary services are based on traditional or informal networks resulting in wide variations in practices. This presentation will report results from diagnostic pathway mapping in lung cancer across three multidisciplinary teams (MDTs) to inform future intervention strategies to reduce variation and improve patient outcomes.

Methods:
We conducted process-mapping workshops with each team to identify the barriers to delivering diagnostic and treatment services. We also developed qualitative interview schedules for GPs and patients. We recruited participants through multiple strategies (mail out, personal invitation from the clinical champion) with local ethics approval.

Results:
Forty-six lung cancer clinicians and four consumers participated in process mapping workshops across three sites. The resulting process maps highlight health system delays and complexities for patients navigating health services, particularly for those living in regional and rural areas. The provision of specialist services for lung cancer diagnosis varies significantly geographically with potential for patients to be lost to follow up. Twelve GPs completed in-depth qualitative interviews or participated in a focus group to identify barriers and enablers in diagnostic pathways. Qualitative analysis reveals that GPs need tailored information about appropriate referrals to specialist pulmonologists or oncologists at the time of a suspicious lung cancer. For GPs without established referral networks, there can be significant uncertainty about the most appropriate referral pathways. Analysis from qualitative interviews with 20 lung cancer patients and their carers indicates that they perceive their GP as having an advocacy role in coordinating their care across specialists’ appointments and diagnostic investigations. Patients reported that personal contact and networking across clinicians in primary, secondary and tertiary settings was a significant factor in the timeliness of investigations or being referred for treatment. In particular, the urgency or severity of symptoms significantly impacted timeliness in securing appointments for investigative diagnostic tests. Patients reported that GPs willing to coordinate their care played an enabling or facilitating role in their care pathway.

Conclusion:
This collaborative project between clinicians and researchers has identified significant barriers and enablers in diagnostic pathways in lung cancer. Primary care practitioners play a significant role in managing patient care and require timely and tailored information about how to refer to a specialist who actively participates in a MDT. We have subsequently developed a protocol to implement a referral decision prompt at the time of CT investigation for those people with a suspicious lung lesion. This prompt will be directed at primary care practitioners and we are currently undertaking a pilot study to examine its feasibility and acceptability.

Abstract not provided

Abstract not provided

Abstract:
In EGFR-mutant lung cancer, acquired resistance to EGFR tyrosine kinase inhibitors (TKIs) develops after a median of 9-14 months. The T790M gatekeeper mutation is the most common mechanism of TKI resistance, detected in >50% of tissue biopsies done after the advent of resistance. The recent clinical development of third-generation, irreversible EGFR TKIs that have preliminarily demonstrated durable tumor responses in patients who have developed the EGFR T790M mutation has generated a need for novel methods of T790M detection. Repeating tumor biopsies at the time of acquired resistance to help select second-line therapies is recommended in the NCCN guidelines. However, tissue biopsies do not always supply sufficient material for current sequencing strategies and thus may require multiple invasive procedures for adequate genotyping. Blood-based methods are more readily repeated when necessary and avoid the risks and discomfort of invasive tissue biopsies. As there may be heterogenous mechanisms of acquired resistance, a tissue biopsy of a single site of disease also may not capture the full spectrum of resistance. Blood-based methods theoretically have the potential of more comprehensively illustrating the principal mechanisms of resistance within a patient. Although there are multiple non-invasive sources of tumor-derived genetic material, circulating tumor cells (CTCs) and plasma circulating tumor DNA (ctDNA) are two that have received particular attention for blood-based genotyping. CTCs are cells shed into the bloodstream from primary and metastatic tumors that can be captured through multiple microfluidic platforms. Despite their rarity in the blood there is ongoing development of increasingly sensitive methods of CTC isolation. ctDNA is also shed into the bloodstream from tumor deposits. While more abundant than CTCs, ctDNA analysis is complicated by a high background of plasma DNA shed from normal cells. Techniques for genotyping from these blood-based sources of tumor-derived genetic material have proliferated rapidly, but there have been few studies directly comparing them. In this presentation, I will describe an exploratory study comparing T790M genotyping, using either CTCs or ctDNA versus concurrent tumor biopsies in patients with non-small cell lung cancer progressing on first line EGFR inhibitors.

Abstract:
Blood-based proteomics strategies for the early detection of lung cancer. Since the advent of the new proteomics era, large-scale studies of protein profiling have been exploited to identify the distinctive molecular signatures in a wide array of biological systems spanning areas of basic biological research, various disease states, and biomarker discovery directed toward diagnostic and therapeutic applications. Recent advances in protein separation and identification techniques have significantly improved proteomics approaches, leading to enhancement of the depth and breadth of proteome coverage. Proteomic signatures specific for invasive lung cancer and preinvasive lesions have begun to emerge. In this presentation, we will provide a critical assessment of the state of recent advances in proteomic approaches to the discovery and validation of blood based biomarker signatures for the early detection of lung cancer. Mass spectrometry and immuno-based detection methods will be reviewed including commercially available blood tests to aid the early detection of lung cancer. Much of this progress was driven by increasing knowledge of tumor-related aberrations that affect nucleic acids at genomic, transcriptional, and posttranscriptional levels. Proteins are the functional end product of genes that ultimately control vital biological processes via their expression level and posttranslational modifications. Moreover, the number of proteins produced by cells far exceeds the number of genes because proteins vary in their stability compared with mRNA and are subjected to many levels of posttranscriptional and posttranslational regulations, such as splicing variants, fusions, and posttranslational modifications. Therefore, to advance our understanding of the biology of lung cancer and to obtain a more integrated view of the disease biology, it is critical to capture the full spectrum of the variations in protein expression patterns, their posttranslational modifications, and their functions in cancer cells. Thus, we hope to take advantage of the molecular complexity of the proteome to improve early detection strategies for lung cancer. Proteomic analysis of blood represents an appealing choice to researchers addressing the discovery of biomarkers because it can be quickly and easily obtained noninvasively in large quantities over time. Several recent studies have investigated the extent to which proteomic technologies can unravel the complexity of the plasma proteome. In this regard, the Human Proteome Organization completed a comprehensive collaborative study to characterize the human serum and plasma proteomes. The rapid proteomic profiling of blood in particular has generated great enthusiasm but has been minimally successful at providing robust signatures to translate to the clinic. The major preanalytical challenges are related to the lack of standardized sample collection and preparation techniques, leading to the introduction of analytical bias and the lack of reproducibility. The extreme complexity of biofluids, such as blood, serum, or plasma, and the low abundance of most of the specific protein markers are among other factors that reduce the sensitivity of detection by proteomic technologies. After the discovery of new biomarkers, the next critical steps are to validate and evaluate their performance in clinically relevant patient populations. Multiple levels of validation have to take place before confirming the clinical utility of the biomarker. This includes confirmation of detected changes in protein level by different techniques and correlation with biological outcomes of lung cancer such as early detection, chemosensitivity, or survival. These phases of clinical validation will evaluate a biomarker's performance in relevant clinical context and how it may affect clinical management of risk or disease. Selected readings: 1. Zeng GQ, Zhang PF, Deng X, Yu FL, Li C, Xu Y, Yi H, Li MY, Hu R, Zuo JH, et al. Identification of candidate biomarkers for early detection of human lung squamous cell cancer by quantitative proteomics. Molecular & cellular proteomics : MCP. 2012;11(6):M111 013946. 2. Massion PP, and Walker RC. Indeterminate pulmonary nodules: risk for having or for developing lung cancer? Cancer Prev Res (Phila). 2014;7(12):1173-8. 3. Hassanein M, Callison JC, Callaway-Lane C, Aldrich MC, Grogan EL, and Massion PP. The state of molecular biomarkers for the early detection of lung cancer. Cancer Prev Res (Phila). 2012;5(8):992-1006. 4. Kikuchi T, Hassanein M, Amann JM, Liu Q, Slebos RJ, Rahman SM, Kaufman JM, Zhang X, Hoeksema MD, Harris BK, et al. In-depth proteomic analysis of nonsmall cell lung cancer to discover molecular targets and candidate biomarkers. Molecular & cellular proteomics : MCP. 2012;11(10):916-32. 5. Skates SJ, Gillette MA, LaBaer J, Carr SA, Anderson L, Liebler DC, Ransohoff D, Rifai N, Kondratovich M, Tezak Z, et al. Statistical design for biospecimen cohort size in proteomics-based biomarker discovery and verification studies. Journal of proteome research. 2013;12(12):5383-94. 6. Zhang B, Wang J, Wang X, Zhu J, Liu Q, Shi Z, Chambers MC, Zimmerman LJ, Shaddox KF, Kim S, et al. Proteogenomic characterization of human colon and rectal cancer. Nature. 2014;513(7518):382-7. 7. Neal JW, Gainor JF, and Shaw AT. Developing biomarker-specific end points in lung cancer clinical trials. Nature reviews Clinical oncology. 2015;12(3):135-46.

Abstract not provided

Abstract:
A commonly employed approach to understanding the pathogenesis of lung neoplasia is to use experimental animal models. The testing of potential chemopreventive (and chemotherapeutic) agents involves pre-clinical testing, and numerous animal models have been developed. In primary mouse lung tumor models, lung cancer develops through a predictable series of airway lesions that progress from normal epithelium to invasive cancer. Permanent genomic DNA alterations occur through either spontaneous, chemically, or environmentally-induced initiation events. The lung cancer chemoprevention field is increasingly reliant on animal studies as the results of negative, early, large scale human studies (for example, β-carotene) may have been predicted with extensive pre-clinical testing. Agents progressing to human trials now undergo extensive pre-clinical studies, and this review will focus on the commonly utilized models of adenocarcinoma (ADC) and squamous cell carcinoma(SCC). ADC: Multiple, well-characterized models of murine adenocarcinoma are available in which pulmonary adenomas progress to adenocarcinomas. These progression models also allow for the study of pre-malignant airway lesions. The most commonly studied models include initiator-promoter carcinogenesis(1), mutant KRAS(2) or EGFR(3) and the use of complete carcinogens. Urethane, a component of cigarette smoke, is a complete carcinogen because it leads to tumor development without the need for other carcinogens or promoters(4). 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) , another tobacco smoke carcinogen, is a chemically-induced model where the NNK is typically administered in drinking water or injected intraperitoneally. Tobacco smoke models also exist and they can reproducibly induce pulmonary adenocarcinomas, but they are fairly labor-intensive and do not result in robust tumor multiplicity(5). There are also multiple two-stage murine models of lung tumor promotion. Several of them use 3-methlycholanthrene (3-MCA), a polycyclic aromatic hydrocarbon, as the initiator at a low dose, followed by multiple doses of butylated hydroxytoluene (BHT). BHT is a well known antioxidant found in processed foods and packaging, however it is also an oxidant and well characterized promoter. In addition, genetically modified animal models, in which viral oncogenes or transforming ras mutants are selectively expressed in lung tissue, have been developed and extensively investigated. While there are differences between the human and murine respiratory tracts, the murine tumors derived from these models have many similarities to human adenocarcinoma, ranging from specific markers to gene expression patterns(6). SCC: Models of squamous cell lung cancer have also been developed, but they are much fewer in number. The most commonly utilized SCC model involves NTCU (N-nitroso-tris-chloroethylurea), which is applied topically and histopathological analysis of serial lung sections in this model revealed a range of lung pathology, including squamous-cell carcinoma, carcinoma in situ, and varying levels of bronchial dysplasia(7). Immunohistochemical studies on the premalignant lesions show staining that corresponds to analogous human lesions(8), and the NTCU model also induces dysplastic lesions that are similar to those found during bronchoscopy and can therefore be used to evaluate one proposed surrogate endpoint in pre-clinical studies (endobronchial dysplasia). Several positive murine chemoprevention studies examining ginseng, pomegranate fruit extract, and aerosolized budesonide +/- pioglitazone have also used NTCU (reviewed in (9)). Additional SCC models have been developed, the first consisting of a kinase dead IKKα knockin mouse (Ikkα[K44A/K44A], Ikkα[KA/KA]) that develop spontaneous SCC and marked pulmonary inflammation(10). A recently described model involved biallelic inactivation of LKB1 and PTEN in mouse lung leads to SCC that expresses the squamous markers keratin 5, p63, and SOX2(11). Chemopreventive interventions have been assessed in many of the murine preclinical models. This includes (but is not limited to): inhaled and systemic glucocorticoids; myoinositol; overexpression of prostacyclin synthase; dietary administration of the prostacyclin agonist iloprost; PPARγ overexpression; dietary administration of pioglitazone; COX inhibitors; the VEGF inhibitor vandetanib; and the anti-estrogen fulvestrant. The effect of COX inhibitors on lung cancer prevention has also been tested in murine models. It is likely the chemoprevention world will take a cue from lung cancer therapeutics by determining the altered pathways in specific premalignant lesions and employing targeted (or ‘precision’) chemoprevention in the next generation of trials. Reference List (1) Malkinson AM, Koski KM, Evans WA, Festing MF. Butylated hydroxytoluene exposure is necessary to induce lung tumors in BALB mice treated with 3-methylcholanthrene. Cancer Res 1997;57:2832-4. (2) Johnson L, Mercer K, Greenbaum D, Bronson RT, Crowley D, Tuveson DA, et al. Somatic activation of the K-ras oncogene causes early onset lung cancer in mice. Nature 2001;410:1111-6. (3) Regales L, Balak MN, Gong Y, Politi K, Sawai A, Le C, et al. Development of new mouse lung tumor models expressing EGFR T790M mutants associated with clinical resistance to kinase inhibitors. PLoS One 2007;2:e810. (4) Malkinson AM. Primary lung tumors in mice: an experimentally manipulable model of human adenocarcinoma. Cancer Res 1992;52:2670s-6s. (5) Witschi H. Tobacco smoke as a mouse lung carcinogen. Exp Lung Res 1998;24:385-94. (6) Stearman RS, Dwyer-Nield L, Zerbe L, Blaine SA, Chan Z, Bunn PA, Jr., et al. Analysis of orthologous gene expression between human pulmonary adenocarcinoma and a carcinogen-induced murine model. Am J Pathol 2005;167:1763-75. (7) Wang Y, Zhang Z, Yan Y, Lemon WJ, LaRegina M, Morrison C, et al. A chemically induced model for squamous cell carcinoma of the lung in mice: histopathology and strain susceptibility. Cancer Res 2004;64:1647-54. (8) Hudish TM, Opincariu LI, Mozer AB, Johnson MS, Cleaver TG, Malkoski SP, et al. N-nitroso-tris-chloroethylurea induces premalignant squamous dysplasia in mice. Cancer Prev Res (Phila) 2012;5:283-9. (9) Keith RL, Miller YE. Lung cancer chemoprevention: current status and future prospects. Nat Rev Clin Oncol 2013;10:334-43. (10) Xiao Z, Jiang Q, Willette-Brown J, Xi S, Zhu F, Burkett S, et al. The pivotal role of IKKalpha in the development of spontaneous lung squamous cell carcinomas. Cancer Cell 2013;23:527-40. (11) Xu C, Fillmore CM, Koyama S, Wu H, Zhao Y, Chen Z, et al. Loss of Lkb1 and Pten leads to lung squamous cell carcinoma with elevated PD-L1 expression. Cancer Cell 2014;25:590-604.

Abstract:
Although chemoprevention as a means of reducing cancer incidence has been successful for basal cell carcinoma of skin, breast, and prostate cancer, after three decades of research, none of the phase III trials with agents such as Beta-carotene, retinol, 13-cis-retinoic acid, alpha-tocopherol, N-acetylcysteine, acetylsalicylic acid, or selenium have demonstrated beneficial and reproducible results in preventing lung cancer, likely due to the complexity of genomic alterations in lung cancer (1). Intermediate endpoint biomarkers such as bronchial metaplasia or dysplasia have been used in Phase II trials (2). Studies on the natural history of pre-neoplastic lesions in the central airways showed that patients with high-grade dysplasia or carcinoma in-situ are more likely to develop invasive cancer at the same or another site in the lungs than those with low grade lesions. However, pre-invasive bronchial lesions may be more of a marker of lung cancer risk because more cancers developed from a separate site in the same individual than progression from an initially biopsied dysplastic site and a significant proportion of the cancers are found by CT rather than by bronchoscopy (3,4). The advantage of using bronchial metaplasia/dysplasia for phase II chemoprevention trials is that these lesions can be localized and biopsied using white light and autofluorescence bronchoscopy for histopathology confirmation. However, with a steady decline in the prevalence of centrally located squamous cell carcinomas and a shift to adenocarcinomas which are usually located in the peripheral lung beyond the range of sampling by standard flexible bronchoscopes, it has become increasingly difficult to enrol participants with bronchial dysplasia for clinical trial. With the implementation of low dose computed tomography for screening of lung cancer, alternative intermediate endpoint biomarkers, such as CT detected non-calcified lung nodules are being investigated for phase II lung cancer chemoprevention trials (5). The limitations of using CT-detected lung nodules as an intermediate endpoint are the lack of confirmation of the underlying pathology and variable growth behaviour of sub-solid nodules especially non-solid nodules (6,7). Without confirmation of the pathology (atypical adenomatous hyperplasia versus adenocarcinoma in-situ versus minimally invasive or invasive adenocarcinoma), when a nodule is first seen, volume doubling time measurement is meaningless to determine malignant behaviour. Endoscopic optical imaging tools such as combined auto-fluorescence-optical coherence tomography (8) are promising methods to localize and characterize small peripheral lung lesions for tissue or liquid biopsy for pathological diagnosis and molecular characterization (Example shown in Figure 1). The availability of accurate lung cancer risk prediction models such as the one developed by Tammemagi and co-workers (9 ) opens the possibility of using lung cancer as the endpoint for chemoprevention trials in high risk ever smokers instead of intermediate endpoint biomarkers to test chemopreventive agents that have sound biological basis. Prospective evaluation of a prototype PLCOm2012 lung cancer risk prediction model in the Pan-Canadian early Detection of Lung Cancer Study showed that a 3 years lung cancer risk of ≥2% can reliably identify a sufficient number of ever smokers who will develop lung cancer within 3 years to make it feasible to use lung cancer incidence as the endpoint for chemoprevention trials with a sample size of approximately 2500 participants. The sample size can be significantly reduced by using a nodule malignancy risk prediction model (10). There is great potential to partner with lung cancer screening programs for developing and testing biologically rationalized agents for chemoprevention clinical trials within this framework, which may lead to their eventual implementation in screening programs to improve patient outcomes. Figure 1. Autofluorescence –Optical Coherence Tomography images of a CT-detected lung nodule showing the invasive and lepidic components of the adenocarcinoma confirmed by transbronchial lung biopsy. Figure 1 Supported by the Terry Fox Research Institute, Canadian Partnership Against Cancer, the Canadian Institute of Health Research and Lung Cancer Canada. References 1. Szabo E, Mao JT, Lam S, Reid ME, & Keith RL (2013) Chemoprevention of lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest 143(5 Suppl):e40S-60S. 2. Keith RL, Blatchford PJ, Kittelson J, Minna JD, Kelly K, Massion PP, et al. Oral iloprost improves endobronchial dysplasia in former smokers. Cancer Prev Res (Phila). 2011;4:793-802. 3. Ishizumi T, McWilliams A, MacAulay C, Gazdar A, Lam S. Natural history of bronchial preinvasive lesions. Cancer Metastasis Rev. 2010;29:5-14. 4. Jeremy George P, Banerjee AK, Read CA, O'Sullivan C, Falzon M, Pezzella F, et al. Surveillance for the detection of early lung cancer in patients with bronchial dysplasia. Thorax. 2007;62:43-50 5. Veronesi G, Szabo E, Decensi A, Guerrieri-Gonzaga A, et al. Randomized Phase II trial of inhaled budesonide versus placebo in high-risk individuals with CT screen-detected lung nodules. Cancer Prev Res 2011; 1:34-42. 6. Massion PP. Walker RC. Indeterminate pulmonary nodules: Risk for having or for developing lung cancer? Cancer Prev Res 2014;7:1173-1178. 7. Pinsky PF, Nath PH, Gierada DS, Sonavane S, Szabo E. Short- and long-term lung cancer risk associated with noncalcified nodules observed on low-dose CT. Cancer Prev Res (Phila). 2014;7:1179-85. 8. Pahlevaninezhad H, Lee AM, Shaipanich T, Raizada R, Cahill L, Hohert G, Yang VX, Lam S, MacAulay C, Lane P. A high-efficiency fiber-based imaging system for co-registered autofluorescence and optical coherence tomography. Biomed Opt Express. 2014 Aug 6;5(9):2978-87. doi: 10.1364/BOE.5.002978. eCollection 2014 Sep 1. 9. Tammemagi MC, et al. (2013) Selection criteria for lung-cancer screening. The New England journal of medicine 368(8):728-736. 10. Tammemagi MC, Lam S. Screening for lung cancer using low dose computed tomography. BMJ. 2014 May 27;348:g2253. doi: 10.1136/bmj.g2253. Review.

Abstract:
Prevention of lung cancer could lead to a significant reduction in the mortality associated with this disease. Identification of individuals at high risk for the development of invasive lung cancer is critical to establishing efficient and effective screening and prevention programs. The presence of premalignant lesions including atypical adenomatous hyperplasia/adenocarcinoma-in-situ (AAH/AIS) and bronchial dysplasia (BD), which represent precursors of adenocarcinoma and squamous cell carcinoma (SCC) respectively, provide targets that can be studied by histologic, radiographic and molecular techniques to define biologic characteristics that are indicative of risk and potential cellular activities that can be targeted for prevention. The histologic features of premalignant lesions have been well described in published WHO defined classification systems (1). Accurate histologic assessment of precursor lesions of lung adenocarcinoma prior to development of invasive cancer is limited by sampling considerations. To establish a diagnosis of premalignant AAH or AIS, the whole lesion must be examined and size criteria and histologic confirmation of lack of invasion must be documented. Because resection is generally restricted to cases of invasive cancer, tissue from AAH or AIS prior to development of invasive adenocarcinoma are rare, and most analyses of these lesions are performed on lesions that are associated with or occur as synchronous independent lesions of invasive cancer in resection specimens. However, a number of recent publications have begun to describe non-lesion associated biomarkers that can be correlated with radiographic features that appear to faithfully distinguish premalignant from invasive peripheral lung lesions. While AAH and AIS are considered to be precursors of adenocarcinomas derived from the terminal respiratory unit (TRU), a recently described premalignant lesion, mucous columnar cell change (MCCC), appears to be the precursor of a less common subset of adenocarcinomas derived from a region of the distal airways that is proximal to the TRU. This lesion has been reported to be present in up to 70% of the mucinous variant adenocarcinomas that are derived from these more central sites (2). This suggests that MCCC may be amenable to sampling at a pre-invasive stage by bronchoscopic means. In contrast, BD is detectable prior to development of invasive cancer by bronchoscopy but cannot be identified by radiographic examination. Higher rates of progression to invasive SCC and/or carcinoma-in-situ for lesions with higher grades of atypia have been suggested in a number of studies and meta-analyses (3). We have assessed the relationship between persistence of BD and risk for development of invasive lung cancer employing a numeric scoring system (1=normal; 2-7=increasing levels of precursor atypia; 8=invasive cancer). These analyses have shown that higher histologic scores on follow-up biopsies at specific sites within the airway of individuals sampled over time are associated with higher baseline histologic score, the presence of papillary angiogenic change, and current smoking status (4). Multivariable analyses including these parameters show that sites in subjects that develop SCC have mean histologic scores on follow-up biopsy that are 1.55 higher than those in patient’s without development of invasive lung cancer. On a per subject basis, the frequency of SCC was significantly increased in subjects that showed multiple sites of BD at baseline that persisted as or progressed to high grade dysplasia (moderate dysplasia or worse, histology score > 5). A 33% increase in risk for development of SCC is associated with every 10% increase in percent of sites that persist/progress to HGD corresponding to an overall hazard ratio of 17.14 (CI 2.4, 123.3) for multifocal persistent BD. These findings lend support to the importance of a field effect in lung carcinogenesis and suggest a potential role for histologic demonstration of persistent field change as an indicator of risk for the development of lung SCC. A number of biomarkers have been studied to determine their relationship with outcomes in premalignant lung lesions. Direct analyses of AAH and AIS have shown that a significant number of these lesions harbor the activating EGFR driver mutations seen in invasive adenocarcinoma of the lung. The potential of these mutational events to act as predictors of progression is under investigation, and a case report has demonstrated response to EGFR inhibitors of radiographically established multifocal premalignant disease in which an EGFR mutation was demonstrated in biopsy tissue of one of the lesions (5). Interestingly, the tumors associated with MCCC show a higher proportion of KRAS mutations. Non-lesional biomarkers of risk such as a recently reported assay measuring germline DNA repair activity that correlates decreased repair capability with increased risk for malignancy show promise for risk prediction (6). BDs, like SCC, demonstrate frequent genetic alterations in tumor suppressor genes and show characteristic associated alterations in gene methylation, loss of heterozygosity and gene copy number gains that have been associated with increased risk (7-10). In an analysis of a small series of cases in which sites with BD were observed to directly progress to invasive SCC, we have demonstrated frequent TP53 and some other mutations in precursor lesions. Furthermore, via pathway analysis of genes that we have found to be differentially expressed between persistent and regressive bronchial dysplasia, we have identified altered control of cell cycle, adhesion and immune activity (see abstract #3026) to be associated with persistence of BD. Overexpression of polo-like kinase 1 (PLK1) is the most prominent cell cycle control alteration associated with persistence and its role as a mediator of progression through the G2-M DNA damage checkpoint suggests a potential mechanism by which genomic instability can be promoted in high risk premalignant BD. PLK1 inhibitor treatment of primary cultures derived from sites of persistent BD causes an arrest of growth in S/G2 phase and induces apoptosis, neither of which occurs when PLK1 inhibitor is applied to primary cultures of normal bronchial epithelium. Histologic features and molecular biomarkers of premalignant lung lesions provide means by which risk can be assessed, appropriate targets for prevention can be identified and efficacy of preventive therapies can be measured. References 1. Travis WD, Brambilla E, Muller-Hermelink HK and Harris CC. Pathology and genetics: tumors of the lung, pleura, thymus and heart. World Health Organization Classification of tumours. Lyon: IARC; 2004. p. 9-124. 2. Weichart W and Warth A. Early lung cancer with lepidic pattern: adenocarcinoma in situ, minimally invasive adenocarcinoma, and lepidic predominant adenocarcinoma. Curr Opin Pulm Med 2014, 20:309–316 3. Ishizumi T, McWilliams A, Macaulay C, Gazdar A and Lam S. Natural history of bronchial preinvasive lesions. Cancer Metastasis Rev 2010;29:5-14. 4. Merrick DT, Haney J, Petrunich S, Sugita M, Miller YE, Keith RLet. al. Overexpression of vascular endothelial growth factor and its receptors in bronchial dysplasia demonstrated by quantitative RT-PCR analysis. Lung Cancer 2005;48(1):31-45. 5. Pastorino U, Calabro E, TamboriniE, MarchianoA, Orsenigo M, Fabbri A, Sozzi G, Novello S, and De Marinis F. Prolonged Remission of Disseminated Atypical Adenomatous Hyperplasia Under Gefitinib. J Thorac Oncol 2009;4: 266–267. 6. Sevilya Z, Leitner-Dagan Y, Pinchev M, Kremer R, Elinger D, Rennert HS, Schechtman E, Freedman LS, Rennert G, Paz-Elizur T, and Livneh Z. Low Integrated DNA Repair Score and Lung Cancer Risk. Cancer Prev Res; 7(4); 398–406. 7. Nakachi I, Rice JL, Coldren CD, Edwards MG, Stearman RS, Glidewell SC, Varella-Garcia M, Franklin WA, Keith RL, Lewis MT, Gao B, Merrick DT, Miller YE, and Geraci MW. Application ofSNPMicroarrays to theGenome-Wide Analysis of Chromosomal Instability in Premalignant Airway Lesions. Cancer Prev Res; 7(2); 255–65. 8. Massion, P., Zou, Y., Uner, H., Kiatsimkul, P.,Wolf, H. J., Baron, A. E., et al. Recurrent genomic gains in preinvasive lesions as a biomarker of risk for lung cancer. PLoS ONE 2009;4(6):e5611. 9. Wistuba, I. I., Behrens, C., Virmani, Ak, Mele, G., Milchgrub, S., Girard, L., et al. High resolution chromosome 3p allelotyping of human lung cancer and bronchial epithelium reveals multiple, discontinuous sites of 3pallele loss and three regions of frequent breakpoints. Cancer Res 2000;60:1949–1960. 10. Belinsky SA[1], Liechty KC, Gentry FD, Wolf HJ, Rogers J, Vu K, Haney J, Kennedy TC, Hirsch FR, Miller Y, Franklin WA, Herman JG, Baylin SB, Bunn PA, Byers T. Promoter hypermethylation of multiple genes in sputum precedes lung cancer incidence in a high-risk cohort. Cancer Res 2006; 66(6):3338-44.

Abstract:
The ability to intervene in the process of carcinogenesis is predicated on an understanding of the pathways leading to invasive cancer and availability of targeted tools to abrogate the resulting processes. Thus, effective chemoprevention has been hampered by the evolving understanding of lung cancer as a heterogeneous set of malignancies arising from a multitude of diverse molecular deregulations. The simplistic view that early intervention (before the evolution of multiple complex mutational events that are characteristic of tobacco-related malignancies) is more likely to be effective than late intervention has been replaced by the realization that many complex abnormalities actually do occur early and we simply do not understand which individual abnormalities or combinations of abnormalities would derail the inevitable progression to invasive and metastatic cancer. To assess efficacy, well designed clinical trials need to have end points that are informative. For phase III trials, the ideal end point would be cancer-related mortality so that cancer overdiagnosis does not cloud the issue. Realistically, cancer incidence is a more achievable and sufficiently informative end point. Phase II trials, however, depend on intermediate end points that are surrogates for cancer incidence, in a manner analogous to tumor shrinkage or progression-free survival being a surrogate for survival in phase II cancer treatment trials. Examples of end points that have been used in a variety of phase II chemoprevention trials are premalignant lesions, proliferative indices, and a variety of biomarkers of risk or malignant potential. It must be emphasized that to be useful, intermediate end points should be integrally involved in the process of carcinogenesis, differentially expressed in at-risk vs. normal epithelium, and modulated by effective interventions well above the level of spontaneous fluctuation (1). To date, no intermediate end point has been validated to replace lung cancer incidence, but such biomarkers can significantly inform drug development and decision-making for subsequent phase III trials. Nevertheless, the histologic evolution of squamous carcinogenesis, with progression from bronchial metaplasia through varying grades of dysplasia to carcinoma in situ is well described (2). This knowledge has allowed for clinical trials based on pre- and post-treatment assessment of effect of interventions on bronchial histology. However, the rate of progression of dysplasia to invasive cancer is variable even though high grade histologies are associated with higher rates of progression. Therefore, studies assessing dysplasia need to have placebo controls to correct for spontaneous and biopsy-induced regression. A recent trial of a prostacyclin analogue, iloprost, showed improvement in bronchial histology in former smokers after 6 months of treatment (3). These results will be extended in a soon-to-open trial of inhaled iloprost in a similar population and will include analyses of potential molecular predictors of histologic progression. Ongoing efforts are focusing on understanding the process of carcinogenesis by profiling premalignant lesions, both in a cross-sectional manner with regard to lesions identified at time of lung cancer resection (4) and with longitudinal follow-up. Understanding the natural history of premalignant lesions will help determine which ones progress, why they progress, and, therefore, which end points are likely to be most informative. An alternative way to approach the issue is to examine the at-risk epithelial field to identify biomarkers associated with progressive carcinogenesis. Gustafson et al. showed that the PI3K pathway is upregulated early during lung carcinogenesis and that an intervention with a drug, myo-inositol, that resulted in regression of bronchial dysplasia also inhibited PI3K activation in the histologically normal bronchial epithelium obtained by bronchial brushings (5). These data suggest that upregulated PI3K could potentially identify smokers at increased lung cancer risk and that pathway inhibition could serve as an end point for assessing treatment effect. This hypothesis is undergoing further testing in a recently finished phase II trial of myo-inositol in current and former smokers with dysplasia, where normal bronchial epithelium was collected and will be tested for PI3K activation pre- and post-treatment. Similarly, Spira et al. showed that gene expression classifiers from bronchial brushings of histologically normal epithelium obtained from individuals undergoing bronchoscopy for suspect lung cancer can aid in the diagnosis of lung cancer and can serve as lung cancer biomarkers (6,7). These classifiers have potential to be adapted to surrogate tissues further up in the aerodigestive tract, such as the nasal epithelium, and are being tested in chemoprevention ongoing clinical trials (8; NCT02123849). How do we move forward? A better understanding of the early carcinogenic processes and which processes are operative in individual persons is key to designing clinical trials that bring the prospect of precision medicine to lung cancer chemoprevention. The focus on a molecular understanding of premalignant lesions and the at-risk field is at the center of current efforts to identify informative end points for chemoprevention clinical trials. References Szabo E. Phase II cancer prevention clinical trials. Semin Oncol 2010;37:359-66. Saccomanno G et al. Development of carcinoma of the lung as reflected in exfoliated cells. Cancer 1974;33:256-70. Keith R et al. Oral iloprost improves endobronchial dysplasia in former smokers. Cancer Prev Res 2011;4:793-802. Ooi AT et al. Molecular profiling of premalignant lesions in lung squamous cell carcinomas identifies mechanisms involved in stepwise carcinogenesis. Cancer Prev Res 2014;7:487-95. Gustafson AM et al. Airway PI3K pathway activation is an early and reversible event in lung cancer development. Sci Trans Med 2010;2:26ra25. Spira A et al. Airway epithelial gene expression in the diagnostic evaluation of smokers with suspect lung cancer. Nat Med 2007;13:361-6. Silvestri GA et al. A Bronchial Genomic Classifier for the Diagnostic Evaluation of Lung Cancer. N Eng J Med 2015 May 17 [Epub ahead of print]. Zhang X et al. Similarities and differences between smoking-related gene expression in nasal and bronchial epithelium. Physiol Genomics 2010;41:1-8.

Background:
Nedaplatin (N) is a second-generation platinum compound with lower nausea/vomiting and nephrotoxicity than cisplatin (C). Nedaplatin plus docetaxel (ND) showed a promising efficacy with acceptable toxicity for advanced squamous cell lung cancer (SqLC) in the previous phase II study.

Methods:
Eligible patients (pts) were those with pathologically proven SqLC with stage IIIB/IV or postoperative recurrence, aged 20-74 years and ECOG PS 0-1. Pts were randomized 1:1 to ND (N 100 mg/m[2] and docetaxel (D) 60mg/m[2] iv, q3w, up to 6 cycles) or C plus D (CD) (C 80 mg/m[2] and D 60mg/m[2] iv, q3w, up to 6 cycles) according to stage, gender and institution. The primary endpoint was overall survival (OS), and secondary endpoints included progression-free survival (PFS), response rate (RR) and adverse events (AEs). Target sample size of 350 provided 90% statistical power to detect a hazard ratio of 0.71 with one-sided type I error of 0.05.

Results:
Between July 2009 and July 2012, 355 pts were randomized. Of 349 for efficacy analysis (ND 177; CD 172), baseline characteristics were well-balanced between two arms. ND had a significantly longer OS (p=0.037, one-sided stratified log-rank test). The OS HR was 0.81 (90%CI, 0.67-0.98) with a median OS of 13.6 months [m] for ND and 11.4 for CD. ND had a longer PFS (p=0.050) with a HR of 0.83 (0.69-1.00) and a median PFS of 4.9 m in ND and 4.5 in CD. RR was 54.5% in ND vs 52.9% in CD (p=0.829). Grade 3 or higher AEs of nausea (4.0% vs 14.3%), fatigue (3.4% vs 10.9%), hyponatremia (13.6% vs 30.3%) and hypokalemia (2.3% vs 8.6%) are more frequent in CD. Grade 3 or higher AEs of neutrophils (82.5% vs 70.3%) and platelets (9.0% vs 0.0%) are more frequent in ND, but there was no difference in grade 3 or higher febrile neutropenia (13.6% vs 15.4%). Treatment related deaths occurred in 4 and 3 pts in ND and CD, respectively.

Conclusion:
ND showed a significantly longer OS as compared to CD with different toxicity profile. ND will be considered as a new standard treatment for advanced or relapsed SqLC. Clinical trial information: UMIN000002015.

Background:
Sacituzumab govitecan (IMMU-132) is a new Antibody Drug Conjugate (ADC) comprising SN-38, the active metabolite of the topoisomerase I inhibitor, camptothecin (irinotecan), conjugated to an anti-Trop-2 humanized antibody at a high drug-antibody ratio (7.6). In vitro and in vivo preclinical data suggest that IMMU-132 delivers up to 136-fold more SN-38 than its parental drug, irinotecan, in a human cancer xenograft. Trop-2 is widely expressed in most epithelial cancers, including non-small and small-cell lung cancers (NSCLC and SCLC).The safety and efficacy of this new ADC is being examined in advanced metastatic lung cancers.

Methods:
A Phase II clinical trial (ClinicalTrials.gov, NCT01631552) is ongoing in subsets of previously-treated patients with metastatic lung cancer, administering IMMU-132 on days 1 and 8 of 21-day treatment cycles. A phase 1 run-in phase selected 8 and 10 mg/kg weekly dosage as safe for tumor cohort phase 2 expansion. Treatment is continued based on tolerance or until progression, with safety and response assessments made every week and 8-12 weeks, respectively.

Results:
Forty-four lung cancer patients were given IMMU-132 doses at 8 mg/kg (N = 23) or 10 mg/kg (N = 21); 38 patients (18 NSCLC and 20 SCLC) are assessable for efficacy. Patients were heavily pretreated (median of 3 prior lines). Objective tumor responses (all partial responses by RECIST1.1) and median progression-free survival (PFS) are reported below per tumor. These studies are being expanded.

Tumor type	Prior lines of therapy: median (range)	Objective Response Rate (PR)	Median PFS (maturity) in months
NSCLC (N=18)	3 (1-8)	33%	5.4 (56%)
SCLC (N=20)	2.5 (1-7)	25%	2.4 (70%)

IMMU-132 was well tolerated with limited grade 3/4 toxicities above the 3% threshold per patient. Neutropenia was the only Grade 3/4 toxicity (G3, 14%; G4, 7%) together with hyponatremia (G3, 2%; G4, 2%). Other drug-related G3 toxicities included diarrhea (7%), anemia (5%), leukopenia (5%), hyperglycemia (5%) and atrial fibrillation (5%); no patient developed antibodies to the conjugate.

Conclusion:
Repeated cycles of IMMU-132 monotherapy are well tolerated. Objective response rate and progression-free survival data in previously-treated metastatic lung cancer (5.4 months in NSCLC) are encouraging and warrant further evaluation of IMMU-132 in these lung cancers.

Background:
Vintafolide (folic acid-vinca alkaloid conjugate) binds to the folate receptor (FR), which is overexpressed in approximately 80% of patients with NSCLC, including patients with squamous cell and adenocarcinoma. Using the molecular imaging agent 99mTc-etarfolatide for SPECT imaging, the FR status of malignant lesions can be determined. Vintafolide has demonstrated single agent activity in patients with advanced NSCLC whose tumors all expressed FR [FR(100%)] compared to patients not FR(100%) (Edelman et al, 2012).

Methods:
This study randomized patients with advanced NSCLC whose tumors were FR(100%) to vintafolide, vintafolide + docetaxel, or docetaxel. Key eligibility criteria: age ≥18 years; 1 prior systemic therapy for advanced disease; ECOG PS 0-1. Patients underwent [99m]Tc-etarfolatide SPECT screening for FR assessment. Vintafolide (2.5 mg) was administered on days 1, 4, 8, 11 every 21 days and docetaxel (75 mg/m[2]) on day 1 every 21 days. The primary endpoint was progression-free survival (PFS). Pre-specified PFS comparisons were performed for vintafolide vs docetaxel and vintafolide+docetaxel vs docetaxel in all patients as well as those with adenocarcinoma. Significance testing for each PFS analysis was one-sided without adjustment for multiplicity (alpha=0.10). Overall survival (OS) was a secondary endpoint.

Results:
Over 14 months, 199 FR(100%) patients were randomized and treated (vintafolide: 63; vintafolide+docetaxel: 68; docetaxel: 68). Patient and disease characteristics were well-balanced between arms. The vintafolide+docetaxel arm met the primary endpoint of superior PFS over the docetaxel arm in all patients regardless of histology (17.0% censored; unstratified Cox model hazard ratio [HR] =0.75; unstratified one-sided p-value=0.0696) as well as in the prespecified 133 patient adenocarcinoma subgroup (18.8% censored; HR=0.73; p-value=0.0899). Trends in OS favored the vintafolide+docetaxel arm over the docetaxel arm in all patients (37.7% censored; HR=0.88; p-value=0.2874) and showed the greatest benefit in the adenocarcinoma subgroup (42.8% censored; HR=0.70; p-value=0.1018). The single-agent vintafolide arm was not superior to docetaxel. Vintafolide+docetaxel treatment was associated with more neutropenia (all grades: 77% versus 62%), febrile neutropenia (13% versus 6%), and peripheral neuropathy (34% versus 21%) compared to docetaxel alone.

Conclusion:
The addition of vintafolide to docetaxel resulted in a statistically significant improvement in PFS in FR(100%) NSCLC patients regardless of histology (PFS HR= 0.75) and in the adenocarcinoma subset (PFS HR= 0.73). Additionally, there was a trend towards improvement in OS in all patients regardless of histology (OS HR= 0.88) and in the adenocarcinoma subset (OS HR= 0.70). Vintafolide +docetaxel was generally well tolerated, although rates of neutropenia, neutropenic fever, and neuropathy were higher than with docetaxel alone. Final survival results will be presented at the conference.

Abstract not provided

Background:
L‑DOS47, a cancer therapeutic designed to exploit the acidic tumour extracellular environment, is a protein conjugate consisting of a urease conjugated to a camelid monoclonal antibody (AFAIKL2) that is targeted to the CEACAM6 antigenic tumour marker. The AFAIKL2 antibody serves as a targeting agent to deliver the enzyme to the tumor sites while the urease enzyme converts urea, an abundant natural metabolite, into ammonia and generates a local pH increase. The combined effect of ammonia toxicity and pH increase is cytotoxic to cancer cells in culture and in xenograft models. This first in human study of L‑DOS47 was designed to define the maximum tolerated dose of multiple doses of L-DOS47 administered intravenously to patients with non-squamous NSCLC when given as a monotherapy.

Methods:
Stage IIIb or IV histologically confirmed non-squamous NSCLC patients (aged ≥18 yrs, ECOG PS ≤2) receive multiple cycles of L-DOS47 during the study treatment period. L-DOS47 is administered once weekly over 14 days followed by 7 days rest in each treatment cycle. Patients are recruited into cohorts and received the same dose of L-DOS47 on Days 1 and 8 of each treatment cycle. Dose levels of L-DOS47 are escalated in further cohorts following a review of safety data by the Trial Steering Committee.

Results:
Thirty-three (33) pts (median age 61, 58% male) were enrolled in the first ten cohorts (dose levels: 0.12, 0.21, 0.33, 0.46, 0.59, 0.78, 1.04, 1.38, 1.84, 2.45 µg/kg) in four Polish centers. L-DOS47 was well tolerated at the dose levels reviewed. No DLTs were reported. Adverse events reported to date were expected for the population under study. None of the patients treated to date have had a partial or complete response as defined by RECIST v1.1. Sixteen (16) patients had an overall response of stable disease after completing two cycles of L-DOS47. One patient in cohort 9 was dosed for 9 cycles without disease progression.

Conclusion:
L-DOS47 monotherapy is well tolerated at dose levels up to 2.45 µg/kg. ClinicalTrials.gov identifier: NCT02340208

Background:
Platinum-based doublet chemotherapy is the standard chemotherapeutic regimen for treatment-naïve advanced non-small cell lung cancer (NSCLC). S-1, an oral fluoropyrimidine, combined with carboplatin or cisplatin (CDDP) has demonstrated the non-inferiority to the standard platinum doublet chemotherapy in Japanese NSCLC patients. However, its effectiveness in Chinese NSCLC patients is uncertain. The purpose of this study is to compare the efficacy and safety of these chemotherapeutic regimens in Chinese NSCLC patients.

Methods:
We did this randomized controlled study in 21 sites in China. Eligible patients were those aged 18-70 years who was histologically or cytologically confirmed with locally advanced or metastatic NSCLC with no prior radiotherapy, molecular targeted therapy or chemotherapy. Patients were randomized to receive either S-1 orally 80 mg/m[2]/day (40 mg/m[2]2 b.i.d., 80–120 mg/day) with 60 mg/m[2] CDDP on day 8 every 5 weeks (SP) or docetaxel and CDDP (both 75 mg/m[2]) on day 1 every 3 weeks (DP) for up to 6 cycles. Randomisation was stratified by centre, pathological classification, disease stage and gender. The primary endpoint was progression free survival (PFS), analyzed in the full analysis set. The study is registered at ClinicalTrials.jp, number Japic CTI-111479.

Results:
Between March 2011 and November 2012, 246 patients from 21 institutions in China were randomly assigned and received SP or DP treatment (124 vs 122) with 18-month follow-up period from the last patient randomized. In the SP and DP group, median PFS was 5.9 and 5.7 months (HR=0.68; 95% CI, 0.48-0.96) respectively, median overall survival was 19.1 and 14.8 months, respectively (HR=0.84; 95% CI, 0.61-1.14). The most common grade 3 or worse adverse events in both treatment groups were neutropenia 3.3% vs 55.1%, leukopenia 1.7% vs 39.0%, and febrile neutropenia 0.8% vs 5.9%, of 121 patients in the SP group and of 118 patients in the DP group, respectively.

Conclusion:
The efficacy of SP was non-inferior to DP with a better safety profile. SP would be a new standard first-line chemotherapy regimen for Chinese patients with advanced NSCLC.

Abstract not provided

Background:
The purpose of this prospective multicenter study was to evaluate the natural course of progression of pulmonary subsolid nodules.

Methods:
Eight facilities participated in this prospective study. This study was conducted with the approval of the institutional review board of each of the participating institutions. Written informed consent was obtained from all the patients. A total of 845 patients with 1325 pulmonary subsolid nodules were registered, of whom 795 patients (341 men, 454 women; mean age, 62 years [range, 31-88]) with 1238 subsolid nodules were selected as being eligible for this study. In this study, the pulmonary subsolid nodules were classified into three categories: pure ground-glass nodules (hereafter abbreviated as PGGNs), heterogeneous GGNs (solid component detected only in the lung window setting; hereafter abbreviated as HGGN), and part-solid nodules (solid component also detected in the mediastinal window setting). The CT images of the nodules that showed progression were reviewed by an expert radiologists’ panel. Pathological specimens of the resected nodules were reviewed by an expert pathologists’ panel.

Results:
The mean prospective follow-up period was 4.3 ± 2.5 years (range, 0.2–12.1; median, 3.5 [IQR, 2.4–6.0]). After exclusion of 9 resected nodules (2 no-lung-cancer nodules and 7 lung cancers not reviewed by the expert pathologists’ panel), the pulmonary subsolid nodules were classified as follows at the baseline: 1046 PGGNs, 81 HGGNs, and 102 part-solid nodules. Among the 1047 PGGNs, 13 (13/1046; 1.2%) developed into HGGNs, and 56 (56/1046; 5.4%) developed into part-solid nodules. Among the 81 HGGNs, 16 (16/81; 19.8%) developed into part-solid nodules. Thus, the subsolid nodules were classified as follows at the time of the final follow-up: 977 PGGNs, 78 HGGNs and 174 part-solid nodules. Of the 977 PGGNs, 35 (3.6%) were resected; from the histopathologic standpoint, the 35 resected PGGNs consisted of 9 minimally invasive adenocarcinomas (MIAs), 21 adenocarcinomas in situ (AISs), and 5 atypical adenomatous hyperplasias (AAHs). Of the 78 HGGNs, 7 (9%) were resected; from the histopathologic standpoint, the 7 HGGNs consisted of 5 MIAs and 2 AISs. Of the 174 part-solid nodules, 49 (28.2%) were resected; from the histopathologic standpoint, the 49 part-solid nodules consisted of 12 invasive adenocarcinomas, 26 MIAs, 10 AISs, and 1 AAHs. In total, 12 (12/1229, 1%) invasive adenocarcinomas, 40 (40/1229; 3.3%) MIAs, 33 (33/1229; 2.7%) AISs, and 6 (6/1229; 0.5%) AAHs were resected as of December 31, 2013; For the PGGNs, the mean period to development into part-solid nodules was 3.8 ± 2.0 years (range, 0.5-8.7; median, 3.4 [IQR, 2.0–5.2]); for the HGGNs, the mean period to development into part-solid nodules was 2.1 ± 2.3 years (range, 0.2–8.8; median, 1.0 [IQR, 0.7–3.4]) (P=0.0004).

Conclusion:
Our prospective multicenter study revealed the frequency and period of development from PGGNs and HGGNs into part-solid nodules. Invasive adenocarcinomas were only diagnosed in the part-solid nodules. The findings of the study may contribute to the development of guidelines for follow-up of pulmonary subsolid nodules.

Background:
A wait-and-see principle is not commonly used when lung cancer is suspected, because of the aggressiveness of the disease. In-vivo information on growth patterns of lung cancers, from small nodules barely detectable by imaging techniques to histologically proven lung cancers, is therefore scarce. In low-dose computed tomography (LDCT) lung screening, lung nodules, usually benign, are found in the majority of screenees. Follow-up CT examinations are performed to determine nodule growth, in order to differentiate between benign and malignant nodules. Growth is often defined in terms of volume-doubling time (VDT), under the assumption of exponential growth. However, this pattern has never been quantified in actual patient data. Our purpose was to evaluate and quantify growth patterns of lung cancers detected in LDCT lung cancer screening, in order to elucidate the development and progression of early lung cancer.

Methods:
The study was based on data of the Dutch-Belgian randomized lung cancer screening trial (NELSON trial). Solid lung cancers detected at ≥3 LDCT examinations before referral and diagnosis were included. Nodule volume was calculated by semi-automated software (LungCARE, Siemens, Erlangen). We fitted lung cancer volume (V) growth curves with a single exponential, expressed as V=V~1~exp(t/τ), with t time from baseline (days), V~1~ estimated volume at baseline (mm[3]) and τ estimated time constant. Overall VDT per lung cancer for all time points combined was calculated using τ*log(2). We used R[2] coefficient of determination as a measure for goodness of fit, where a perfect fit results in R[2]=1. A normalized growth curve for all lung cancers combined was created by plotting normalized volume (V/V~1~), on a logarithmic y-axis as a function of normalized time, t*=t/τ. Statistical analyses were performed using SPSS 20.0 and Octave (www.octave.org).

Results:
Forty-seven lung cancers in 46 participants were included. Seven participants were female (13.0%); mean age 61.7 ±6.2 years. Median follow-up time before lung cancer was diagnosed, was 770 days (IQR: 383-1102 days). One cancer (2.1%) was diagnosed after six LDCTs, six (12.8%) after five LDCTs, 14 (29.8%) after four LDCTs, and 26 cancers (55.3%) after three LDCTs. Most lung cancers were stage I disease (35/47, 74.5%) at diagnosis. The majority concerned adenocarcinoma (38/48, 80.9%). Median overall VDT was 348 days (IQR: 222-492). Overall VDT for adenocarcinomas versus other histological cancer types were similar (median 338 days [IQR: 225-470 days] versus 348 days [IQR: 153-558 days], respectively [p=NS]). Good fit to exponential growth was confirmed by the high R[2] coefficient of determination for the individual cancer growth curves (median 0.98; IQR: 0.94-0.99). After normalization, we found linear growth on a logarithmic scale, according to exponential growth, for almost all nodules. Not all cancers showed an exponential growth immediately from baseline; five cancers were identified with constant (low) volume for >500 days before growth expansion occurred. However, when these dormant lung cancers started growing, they followed the exponential function with excellent fit (median 1.00; IQR: 0.98-1.00).

Conclusion:
Screen-detected lung cancers usually evolve at an exponential growth rate. This makes VDT a powerful imaging biomarker to stratify prevalent lung nodules to growth rates.

Background:
Lung cancer in never-smokers is recognised as a distinct entity. Many are expected to present late. As there are no established aetiological factors, identification of patients at risk is challenging. The aim of the study is to define the incidence and clinical features of never-smokers presenting sufficiently early for surgery to determine if it is possible to identify patients at risk.

Methods:
We retrospectively analysed data from a prospectively collected database of patients who underwent surgery at our institution. The incidence was defined as number of never-smokers versus current and ex-smokers by year. Clinical features at presentation were obtained and collated as frequency (percentage).

Results:
A total of 2170 patients underwent surgical resection for lung cancer from March 2008 to November 2014. The annual incidence of developing lung cancer in never-smokers increased from 13, 15, 18, 19, 20, 20 to 28 percent respectively, attributable to an absolute increase in number and not a change in the ratio of never smokers to current and ex-smokers. A total of 436 (20%) patients were never smokers. The mean age at presentation was 60 (16 SD) years and 295 (67%) were female. Good lung function was observed with mean predicted FEV1 of 90% (23 SD) and FVC of 97% (25 SD). The majority histological types were adenocarcinoma 54% and carcinoid 27%. The main presenting features were non-specific consisting of cough in 142 (34%), chest infections in 75 (18%) and haemoptysis in 46 (11%). Recurrent chest infections were predominantly a symptom of central carcinoid tumours (30 versus 15 percent; P=0.004). A total of 59 (14%) were detected on incidental chest film, 127 (30%) on incidental CT, 32 (7%) on incidental PET/CT and 4(1%) on incidental MRI.

Conclusion:
We observed more than double the annual incidence of never smokers presenting with non small cell lung cancer, in the last 7 years, increasing from 13 to 28 percent, and hypothesise that this is representative of the UK, as we are one of the highest surgical volume centres in our country. Patients present with non-specific symptoms and the majority were detected on incidental imaging. We conclude that imaging is likely to play a more important role and further efforts need to be expended on early detection of lung cancer in this increasing cohort without any observable risk factors.

Abstract not provided

Background:
The National Lung Screening Trial (NLST) found a 20% reduction in lung cancer-specific mortality using low dose CT vs chest radiography for screening. The magnitude of mortality benefit has been questioned given that a higher proportion of tumors in the CT arm were diagnosed as “bronchioloalveolar cell carcinoma”. Subsequent to the initiation of the NLST, the pathological classification of lung cancer was revised to take into account the reported favorable outcome for solitary in situ nodules <3 cm. The term “bronchioloalveolar carcinoma” (BAC) was eliminated in favor of the more explicit terms adenocarcinoma in situ (AIS), microinvasive adenocarcinoma (MIA), and invasive carcinoma with various predominant histological patterns. To better assess the impact of these recent changes in the Pathological classification of lung cancer on possible over-diagnosis in the NLST, we have reviewed the histology of lung tumors detected through the ACRIN-NLST trial and reclassified them according to the most recent WHO pathology classification.

Methods:
Histology was initially classified by the pathologists at sites where NLST participants were managed. Representative slides of 192 surgical resection specimens and 15 non-surgical biopsies from 207 patients were collected from 19 participating institutions. Digital images were prepared from 533 glass H&E stained slides using an Aperio digital slide imager. Digital images were examined by three pulmonary pathologists (WAF, DTM and JDH) and reclassified according to criteria and nomenclature of the recently published 2015 edition of the WHO classification.

Results:
There was 92% concordance between submitting and reference pathologists when cases were grouped into the broad categories of adenocarcinoma, squamous carcinoma, neuroendocrine and large cell lung carcinoma (LCLC). The WHO classification permitted a more detailed analysis of the tumors. Invasive adenocarcinoma was the largest tumor category comprising 61% (127) of all tumors and included 70 acinar tumors, 23 solid, 13 papillary, 8 micropapillary, 5 mixed mucinous/non-mucinous, 4 invasive mucinous, 3 lepidic and 1 adenocarcinoma that could not be further classified. There were 48 (23%) squamous tumors, 10 (5%) LCLC, 15 (7%) neuroendocrine tumors including 6 (3%) small cell lung carcinomas. Finally, one tumor had sarcomatoid histology and an additional tumor was classified at sclerosing pneumocytoma. On reclassification, only 5 of the 26 tumors originally referred to as BAC or as having BAC features by submitting pathologists met criteria for adenocarcinoma in situ or minimally invasive carcinoma. Twenty-one of these 26 tumors were reclassified as invasive adenocarcinoma, most frequently acinar pattern predominant (8 cases).

Conclusion:
Reclassification of tumors identified through low dose CT screening in the National Lung Screening Trial permitted a detailed analysis of histological features and should permit a more nuanced assessment of biology and prognosis of this important cohort than has been available to date. Reclassification of BAC mainly as invasive adenocarcinoma conflicts with the suggestion that much of the benefit in the NLST CT screening trial was derived from surgical removal presumably non-invasive low grade tumor. *ACRIN received funding from the National Cancer Institute through the grants U01 CA079778 and U01 CA080098.

Background:
Screening for lung cancer with low-dose computed tomography (LDCT) was shown to reduce lung cancer mortality. However, lung cancer screening also detects indolent cancers of unclear clinical significance, which generally belong to the adenocarcinoma spectrum. The individualized management of these more indolent cancers may be facilitated by non-invasive risk stratification. We present our validation study of CANARY (Computer-Aided Nodule Assessment and Risk Yield), a novel LDCT-based software, used to stratify adenocarcinoma nodules in three groups with distinct outcomes.

Methods:
All individuals in the LDCT arm of the National Lung Screening Trial (NLST) with adenocarcinoma were identified. The last LDCT data available were analyzed blinded to clinical data. Using CANARY, all lung adenocarcinoma nodules were classified as Good (G), Intermediate (I) and Poor (P) based on previously established radiologic signatures. This classification was then used for survival analysis using progression-free survival

Results:
LDCT datasets of 294 patients with resected adenocarcinomas with available outcome data were included in the blinded CANARY analysis. Kaplan-Meier analysis of all the 294 adenocarcinoma nodules stratified into G, I and P CANARY classes yielded distinct progression-free survival curves (P < 0.0001). A similar separation was seen with adjusted progression-free survival curves, after adjustment for, age, gender, race and smoking status for all pathological stage I cases.

Conclusion:
CANARY allows the non-invasive risk stratification of lung adenocarcinomas into three groups with distinct post-surgical disease-free survival. Our results suggest that CANARY could facilitate individualized management of incidentally- or screen-detected lung adenocarcinomas.

Background:
Lung cancer screening identifies cancers with heterogeneous behaviors. In addition to screen-detected incidence lung cancers, screening also identifies prevalence cancers at the baseline screen and interval lung cancers diagnosed following a negative screen at any time point prior to the next screening round. To date, few studies have performed a comprehensive analyses comparing prevalence and interval lung cancers and screen-detected lung cancers based on sequence of screening results in the NLST.

Methods:
The entire CT arm of the NLST was reconstructed according to baseline and follow-up screening results (positive vs. negative screen). Lung cancers immediately following a positive baseline (T0), and prior to the T1 screen, formed the prevalence cancers (PC); interval cancers (IC) were defined as lung cancers diagnosed following a negative screen at any point prior to the next screening round. Two screen-detected lung cancer (SDLC) cohorts were identified based on one (SDLC1) or two (SDLC2) prior positive screens and two screen-detected lung cancer cohorts following one (SDLC3) or two (SDLC4) prior negative screens. Differences in patient characteristics, progression-free survival (PFS), and overall survival (OS) were assessed.

Results:
Since there were no differences in patient characteristics and outcomes between SDLC1 and SDLC2 and between SDLC3 and SDLC4, the four screen-detected cancer case groups were combined into two combined SDLC case groups (SDLC1/SDLC2 and SDLC3/SDLC4). The lung cancer-specific death rate was higher for SDLC3/SDLC4 compared to SDLC1/SDLC2 lung cancers (136.6/1,000 person-years vs. 71.3/1,000 person-years, P < 0.001). PFS and OS were significantly lower for SDLC3/SDLC4 than SDLC1/SDLC2 (P < 0.004; P < 0.002, respectively). Overall, PFS and OS were highest in SDLC1/SDLC2 and lowest in the interval cancers (Figure 1); PFS and OS for the prevalence cancers were intermediate between SDLC1/SDLC2 and SDLC3/SDLC4. All findings were consistent when stratified by stage and histology. Multivariable Cox proportional models revealed that the SDLC3/SDLC4 case groups were associated with significantly poorer PFS (HR=1.72; 95% CI 1.19-2.48) and OS (HR=1.62; 95% CI 1.08-2.45) compared to SDLC1/2 lung cancers (HR=1.00). Figure 1



Conclusion:
This post hoc analysis reveals novel insight to the heterogeneity of lung cancers diagnosed in a screening population. As with interval cancers diagnosed following a negative screen, lung tumors that arise in a lung environment ostensibly free of lung nodules are likely more rapidly growing and aggressive which results in significantly poorer outcomes. Additional research will be needed to understand the potential translational implications of these findings and to reveal biological differences of screen-detected tumors.

Abstract not provided

Background:
Currently, a blood test for lung cancer does not exist. Low-dose spiral computed tomography (CT) has been proposed as an early detection screening tool. However, despite its high sensitivity, the specificity of CT in lung cancer detection is poor. In addition, the necessity for repeated CT scans to determine growth rates over time can expose patients to potentially harmful radiation. Therefore, a minimally-invasive biomarker-based test that could further characterize CT-positive patients based on risk of malignancy would greatly enhance its clinical efficacy.

Methods:
From 2009 through 2013, six hospitals enrolled 1148 patients with lung cancer, 889 blood donors as healthy participants and 36 patients with other lung diseases. The lung cancer associated biomarker panels were identified from the pretreated serum samples and subsequently analyzed in three randomly determined subgroups, the discovery cohort (40 patients with lung cancer, and 45 healthy participants), test cohort (204 patients with lung cancer, and 120 healthy participants), and validation cohort (904 patients with lung cancer, 724 healthy participants, and 36 patients with other lung diseases). Finally the panel of biomarkers were used to predict 12 prospective patients with a suspicious pulmonary nodule by CT images.

Results:
The discovery cohort demonstrated that 4 serum biomarkers (C-reactive protein, prolactin, hepatocyte growth factor, and NY-ESO-1 autoantibody) were significantly higher in patients with lung cancer compared to healthy controls. The 4-biomarker panel was independently investigated in the test cohort and validation cohort. The test characteristics were area under the curve (AUC) of 0.835 (95% CI 0.79-0.873, sensitivity 70.1%, specificity 88.3%) in the test cohort, and 0.818 (95% CI 0.798-0.836, sensitivity 70.0%, specificity 79.6%) in the expanded validation cohort. The 4 biomarkers had no discriminatory power for detecting other benign lung diseases. The performance of the panels in patients with stage I-II lung cancer was AUC of 0.774 (95% CI 0.746-0.801) in the combined test and validation cohorts. Remarkably, analysis model generated by the biomarkers correctly predicted 7 out of 9 prospective patients having lung cancer, and 2 out of 3 patients as benign, which were further verified by the pathologist.

Conclusion:
This study identified four diagnostic biomarkers in serum samples with the potential to distinguish patients with lung cancer from healthy controls. This panel of serum proteins is valuable in suggesting the diagnosis of patients with an indeterminate pulmonary lesion, and potentially in predicting individuals at high risk for lung cancer. Further research is necessary to understand whether these have clinical implications for early detection of lung cancer.

Background:
Early detection represents an important opportunity for decreasing lung cancer mortality. Lung cancer screening with low-dose CT scanning is plagued by a high false-positive rate and non-invasive adjuncts that improve diagnostic accuracy or serve as a pre-screen may be helpful. This study evaluated the performance of a sputum based lung cancer detection (LCD) test that utilizes fluorescence in-situ hybridization (FISH) to detect chromosomal alterations at the 3p22.1 and 10q22.3 loci caused by a cancerous process.

Methods:
At 5 international centers, between March 2012 and July 2014, induced sputum samples were collected from 173 subjects with 8-30 mm solitary pulmonary nodules, where imaging and other subject characteristics mandated biopsy. At least 50 lower respiratory tract cells were required for analysis. The LCD Test, performed at one of 3 reference labs, enabled a combined analysis of sputum cytology and Target-FISH analysis on the same cell using an FDA approved imaging analysis system (BioView Duet™). The LCD test was considered positive if at least 7.5% of the target cells had an abnormal FISH pattern. The results of the LCD were then compared to the clinical pathology. Subjects with an initial non-surgical negative biopsy result were followed for up to 2 years to determine their final diagnosis.

Results:
There were 116 subjects who met the inclusion criteria, had a pathologic diagnosis of lung cancer if the nodule was malignant, and produced adequate sputum for analysis. Seventy-two subjects were diagnosed with lung cancer from the initial biopsy, 7 had definitive negative surgical biopsies, and 37 subjects were classified as indeterminate due to non-surgical negative biopsies. Initial positive concordance was 86.1% (62/72) and initial negative concordance was 71.4% (5/7). From the initial 37 indeterminate negative subjects, additional clinical analyses during the follow up period enabled a definitive classification for 23 subjects: 11 were diagnosed with lung cancer and 12 were reclassified as definitive negative. From this group the LCD test had a positive concordance of 81.8% (9/11) and a negative concordance of 91.7% (11/12). Overall, sensitivity was 85.5% (71/83), specificity was 84.2% (16/19), positive predictive value was 95.9%, and negative predictive value was 57.1%. Fourteen indeterminate negative subjects are still being clinically monitored. The test performance for nodules of 8-20mm was as good as the results for 21-30mm nodules.

Conclusion:
In a cohort of patients with a high risk of lung cancer, the LCD test had a high positive predictive value. A positive LCD test could potentially lead to an earlier intervention in a nodule that might otherwise have been monitored for growth. An adequate cancer resection might then be accomplished by segmental resection rather than lobectomy in smaller lesions. The LCD test may be useful as a decision support tool at critical points in the management of solitary pulmonary nodules detected by screening CT scans in subjects at high risk for lung cancer.

Background:
Lung cancer is conventionally diagnosed by confirmatory tissue biopsy, an invasive procedure that involves waiting time, costs and complications. The development push for a blood based liquid biopsy is a less invasive, more readily acceptable means to expedite the diagnosis and management of cancer. Circulating tumour DNA is promising in this regard as cancer specific genetic mutations are not usually found in the circulation of healthy individuals. The aim of our study is to report the performance of a three gene signature in for the diagnosis of cancer.

Methods:
Pre-operative blood samples were obtained from patients undergoing surgery for known or suspected lung cancer and 1ml aliquots of plasma were extracted from 9ml of EDTA preserved blood. DNA was extracted from the plasma using the QIAamp DNA blood mini kit. High resolution melt analysis was undertaken to identify mutations in hotspots of the TP53, KRAS and EGFR genes in the ctDNA from plasma as well as matching FFPE tissue. A positive test result was defined as a mutation identified in the plasma ctDNA and compared against the reference clinical histopathology report of the resected lung abnormality. Clinical test performance was quantified and reported conventionally using sensitivity and specificity.

Results:
Pre-operative blood was analysed in a blinded manner from 223 patients undergoing surgery at our institution, and the pathology reports were issued blinded to the blood test results. In total, 116 (52%) had primary lung cancer, 64 (29%) had secondary cancer, 6 (3%) had primary thoracic (not lung) cancer and 35 (16%) did not have any evidence of cancer. Of the 186 patients with confirmed cancer, a mutation was identified in the FFPE sections of the primary tumour of 113 (61%) and in the plasma ctDNA in 127 (68%) with substantial agreement of 85% and a kappa statistic of 0.70 (P<0.001). The clinical test performance for the blood based diagnostic signature was a sensitivity of 68% (95% CI 61-75), specificity of 91% (77 to 98), positive predictive value 98% (93-100) and a negative predictive value of 35% (25 to 46) when compared to conventional clinical histopathology reporting of the resected tissue.

Conclusion:
There is substantial agreement between the detection of ctDNA and FFPE tumour tissue mutations. We postulate higher mutation levels detected in the plasma is due to heterogeneity of tumour and FFPE sections in comparison to a global (plasma based ctDNA) estimate of mutation burden. Our results suggest blood based ctDNA analysis of cancer mutations is a specific, non-invasive test for the diagnosis of cancer. A positive test strongly rules in the diagnosis but a negative test does not have sufficient discriminatory ability to exclude the diagnosis of cancer.

Abstract not provided

Background:
Current clinical diagnostic methods lack the specificity in detecting lung cancer patients. The issue is critical for stage I & II patients as there are no available biomarkers to indicate which high-risk patients should undergo adjuvant therapy. There is considerable evidence that microRNA plays a very important role in lung carcinogenesis. We postulated that the expression pattern of multiple microRNAs (miRNAs) could aid clinicians in detecting cancer patients thus reducing the mortality of lung cancer.

Methods:
Differential expressed miRNAs were analyzed by miRNA microarrays in 15 paired non-small-cell lung cancer (NSCLC) tumors and distant normal tissues. The identified miRNAs were further validated by qRT-PCR using snap-frozen lung tissue samples collected from independent 22 patients with NSCLC. Classification analyses of miRNA expression data were performed by the Bayesian Compound Covariate predictor (BCCP). The expression levels of miR-141-5p, miR-301a-3p and miR-1244 were also analyzed by qRT-PCR in serum samples collected from 60 patients with lung cancer and 50 healthy controls.

Results:
A total of 41 miRNAs was identified with significantly elevated levels in patients with lung cancer by profiling microRNA array, of which 12 miRNAs were further validated in the independent sample cohort. Multiplexing analysis with the panel of 12 miRNAs generated the highest discriminatory power in separating NSCLC from normal tissues with an AUC of 0.915 (95% CI = 0.894-1.000; P <0.001). Leave-one-out cross-validation revealed the 85% sensitivity and 95% specificity at a cutoff score of 0.5. In addition, serum miR-1244 was significantly upregulated in an independent trial and could distinguish NSCLC from controls with 77.6% sensitivity and 78.7% specificity.

Conclusion:
Our 12-miRNA classifier might have potential clinical utility in discriminating NSCLC from healthy population.

Background:
Five studies have shown that microRNA levels in whole blood can be used to diagnose lung cancer. We conducted a large bi-institutional study to validate this finding.

Methods:
PAXgene[TM] Blood miRNA System (Qiagen®) was used for peripheral venous blood collection and total RNA isolation for 85 pathologic stage IA-IIIB non-small cell lung cancer cases and 76 clinically-relevant controls who either had a high risk of developing lung cancer because of smoking and age >50 y, or had a benign pulmonary nodule. Cases and controls were accrued at two institutions in the United States, Roswell Park Cancer Institute, Buffalo and University of Pennsylvania, Philadelphia. MiRCURY™ microarrays (Exiqon®) with locked nucleic acid hybridization probes were used to quantify microRNAs in RNA isolates. Quantification was also performed using Taqman™ microRNA reverse transcription (RT)-PCR assays (ABI®) for five microRNAs whose lung cancer-diagnostic biomarker utility had been suggested by the five published studies.

Results:
Cases (n=85) and controls (n=76) were similar for age, gender, race, and blood hemoglobin and leukocyte but not platelet levels (Table 1). Of the 1936 human mature microRNAs detectable with the microarray platform, 586 (30%) were identified as expressed and reliably quantified among the study's subjects. However, none of the microRNAs was differentially expressed between cases and controls (P >0.05 in test using empirical Bayes-moderated t statistics and false discovery rate <5%). In classification analysis using the whole blood microRNA profiles with leave-one-out internal cross-validation, accuracy was 48% and 50% with the support vector machines and top-scoring pair methods, respectively. With RT-PCR assays, cases and controls did not differ for any of the five microRNAs whose biomarker potential had been suggested by previous studies.

Table 1. Characteristics of study groups; *Fisher's exact test for categorical variables, and t test for others; #blood values for 84 cases and 30 controls.

	Cases	Controls	P*
	85	76
Mean age, y (range, SD)	64 (41-83, 8)	61 (45-83, 9)	0.07
%male	49	51	0.87
%white	90	93	0.57
RPCI	42	32	0.43
U. Pennsylvania	43	44
Adenocarcinoma	43
Squamous cell	33
Other non-small cell	9
High-risk control		58
Nodule control		18
Leukocytes (x1000/µl; mean, SD)#	8.2 (2.6)	7.8 (2.1)	0.37
Platelets (x1000/µl; mean, SD)#	291.8 (114.3)	238.2 (50.2)	0.01
Hemoglobin (g/dl; mean, SD)#	13.4 (1.8)	13.9 (1.4)	0.15

Conclusion:
This study suggests that whole blood microRNA expression profiles may not be useful for developing biomarkers for use in non-invasive blood-based assays for generic screening of non-small lung cancer. Further studies are required to examine if whole blood microRNA diagnostic biomarkers may exist for use with specific types of lung cancer or non-cancer control conditions.

Background:
Bronchoscopy is often used for the diagnosis of lung cancer however its sensitivity is imperfect, especially for small and peripheral lesions. Adjunctive methods to improve the sensitivity of cancer detection would reduce the need for more invasive follow-up procedures when bronchoscopy is non-diagnostic. It has previously been shown that gene expression of cytologically-normal bronchial airway epithelial cells is altered in smokers with lung cancer. In this study we evaluated the performance of a bronchial genomic classifier to predict malignancy in an independent cohort of suspect lung cancer patients.

Methods:
A bronchial genomic classifier consisting of the expression of 23 genes measured in the airway epithelium was evaluated in a previously published, independent cohort (n=163) of current and former undergoing bronchoscopy for suspect lung cancer. In cases where bronchoscopy was non-diagnostic for malignancy, the performance of the classifier was evaluated using ROC-AUC, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

Results:
In the test set, bronchoscopy led to a diagnosis in 40 of 78 patients with cancer (sensitivity=51%, 95% CI 40-63%). The combination of the classifier with bronchoscopy improved the sensitivity to 96% (95% CI 89-99%; p <0.001); see Table. The prediction accuracy of the classifier was similar in lesions <3cm, as well as across cancer stage and histology. Among the 123 patients with a non-diagnostic bronchoscopy, 38 were ultimately diagnosed with lung cancer (prevalence of 31%). In this group of patients, the classifier had an AUC of 0.81 (95% CI, 0.73-0.88), accurately identifying 35 of the 38 lung cancer patients (sensitivity=92%; 95% CI, 78-98%), and 45 of 85 patients with benign lesions (specificity=53%; 95% CI, 42-63%). Of the 48 patients with a negative classifier result, 45 were diagnosed with benign lesions (NPV=94%, 95% CI 83-99%).

Table. Performance of bronchoscopy, classifier, and the combined procedures in the test set

Category	Bronchoscopy	Classifier[a]	Combined
Total, N	163	123	163
Lung Cancer, N	78	38	78
Benign Lesion, N	85	85	85
Sens. (95% CI)	51% (40-62%)	92% (78-98%)	96% (89-99%)
Spec. (95% CI)	100% (95-100%)	53% (42-63%)	53% (42-63%)
NPV (95% CI)	69% (60-77%)	94% (83-99%)	94% (83-98%)
PPV (95% CI)	100% (90-100%)	47% (36-58%)	65% (56-73%)

a) The performance of the classifier was evaluated for patients in whom bronchoscopy did not result in a finding of lung cancer (n=123).

Conclusion:
A gene expression classifier measured in bronchial epithelial cells is able to accurately identify those at low risk for lung cancer in patients who have undergone bronchoscopy with non-diagnostic results. Due to the high sensitivity and NPV of the classifier, it could potentially inform clinical decisions regarding the need for further invasive testing for lung cancer in patients whose bronchoscopy is non diagnostic.

Abstract not provided