Virtual Library

Abstract not provided

Abstract not provided

Abstract:
Lung cancer is still the leading cause of cancer-related deaths worldwide. Non-small cell lung cancer (NSCLC) remains the predominant form of the disease, with majority of patients being diagnosed at advanced stages. The survival benefit offered by chemotherapy regimens has reached a therapeutic plateau. Fortunately, mutation-specific targeted therapies directed against “actionable” genetic alterations have significantly improved the prognosis of advanced NSCLC. Epidermal growth factor receptor (EGFR) mutation is the most common targetable genetic alteration in NSCLC. The prevalence of EGFR mutations range from 10% in Caucasians, to 60% in never-smoking, Asian, adenocarcinoma patients. This presentation will focus on the first-line management of EGFR mutant NSCLC. Overview of EGFR-TKIs Nine large randomised controlled studies have established the superiority of EGFR tyrosine kinase inhibitors (EGFR-TKIs) against chemotherapy as first-line treatment in NSCLC harboring EGFR mutations in terms of progression-free survival (PFS), objective response rate (ORR) and quality of life (QoL) (Table 1). Several studies suggest there is no significant difference in efficacy between gefitinib and erlotinib. LUX-lung 7 (afatinib) and ACHER 1050 (dacomitinib) are two randomised head-to head trials comparing second-generation EGFR-TKIs to gefitinib, respectively. Although PFS is significantly improved with second-generation TKIs, the increased rates of grade 3 or higher adverse events such as rash and diarrhea result in more dose modification with second-generation TKIs. FLAURA study, which assesses the efficacy of third-generation EGFR-TKI osimertinib versus first-generation EGFR-TKIs as first-line treatment, has been accomplished recently. This trial may establish the role of osimertinib as first-line treatment for EGFR mutant NSCLC. Management of uncommon mutations Exon 19 deletion and L858R mutation account for about 90% of all EGFR mutations. The remaining 10% of EGFR mutations (uncommon mutations) are a heterogeneous group of molecular alterations. Results of retrospective studies and case reports of first-generation EGFR-TKIs show inconsistent responses in this population. According to the post-hoc analysis of LUX-Lung 2, LUX-Lung 3, and LUX-Lung 6, objective responses to afatinib are observed in certain types of uncommon mutations such as G719X (77.8%), L861Q (56.3%), and S768I (100%). However, Patients with de-novo Thr790Met and exon 20 insertion mutations are insensitive to afatinib. Jonathan Riess and colleagues report that osimertinib has in vivo activity across multiple exon 20 insertion mutations in preclinical study. Thus osimertinib warrants further study in patients with exon 20 insertion mutations. Combination treatment strategies Combinations of EGFR-TKIs with chemotherapy, anti-angiogenetic therapy, and immunotherapy have been explored in clinical trials. JMIT study is a randomized phase II trial comparing addition of pemetrexed to gefitinib with gefitinib alone as first-line therapy in EGFR mutant NSCLC. PFS was significantly longer with pemetrexed+gefitinib than with gefitinib (15.8 months vs 10.9 months; hazard ratio [HR], 0.68; 95% CI, 0.48 to 0.96; P = 0.029). According to JO25567 study, combination of erlotinib with bevacizumab also significantly prolongs PFS than erlotinib. Several phase III trials evaluating combination of EGFR-TKIs with anti-angiogenetic therapy are ongoing. Attention should be paid to adverse events such as interstitial lung disease when EGFR-TKIs are used with immunotherapy. Management of brain metastases Brain metastases is a major challenge when managing EGFR mutant NSCLC as up to 40% of patients would develop central nervous system (CNS) metastases during the course of their disease. Novel EGFR-TKIs provide new strategies for brain metastases treatment. AZD3759 is a CNS penetrable and reversible EGFR-TKI. The phase I study (BLOOM) of AZD3759 in TKI naïve EGFR mutant NSCLC with CNS metastases is reported in 2017 ASCO annual meeting. Intracranial response is observed in 83% of patients with measurable brain metastases lesions at baseline. The extracranial anti-tumor efficacy of AZD3759 is also promising. 13 out of 18 patients with extracranial lesions have confirmed objective response. CNS response to osimertinib in patients with T790M-positive advanced NSCLC from AURA3 study is also presented in 2017 ASCO annual meeting. In 30 patients with more than one measurable CNS metastases, the intracranial response is 70%. The median CNS PFS with osimertinib is 11.7 months. Furthermore, osimertinib shows encouraging activity in patients with leptomeningeal metastases. Table 1. First-line treatment of EGFR mutant NSCLC: EGFR-TKIs vs. Chemotherapy

Trial	TKI	PFS (month)			OS
		TKI	Chemo	HR (95%CI)	HR (95%CI)
IPASS	Gefitinib	9.5	6.3	0.48 (0.36-0.64)	0.78 (0.50-1.20)
First Signal	Gefitinib	8.4	6.7	0.61 (0.31-1.22)	0.82 (0.352-1.922)
NEJ002	Gefitinib	10.8	5.4	0.322 (0.236-0.438)	0.88 (0.634-1.241)
WJTOG3405	Gefitinib	9.6	6.6	0.52 (0.378-0.715)	1.185 (0.767-1.829)
OPTIMAL	Erlotinib	13.1	4.6	0.16 (0.10-0.26)	1.19 （0.83-1.71）
ENSURE	Erlotinib	11.0	5.5	0.34 (0.22-0.51)	0.91 （0.63-1.31）
EURTAC	Erlotinib	9.7	5.2	0.37 (0.25-0.54)	0.80 (0.47-1.37)
Lux-lung 3	Afatinib	11.1	6.9	0.58 (0.43-0.78)	0.88 （0.66-1.17）
Lux-lung 6	Afatinib	11.0	5.6	0.29 (0.20-0.33)	0.93 （0.72-1.22）

Abstract not provided

Abstract:
The identification of anaplastic lymphoma kinase (ALK) gene rearrangements as an oncogenic driver in NSCLC has radically changed the treatment of a subset of patients harboring this molecular alteration. [1] ALK mutations occur in 3-7% of NSCLCs and are more frequently associated with never/light smoker, younger age and adenocarcinoma histology. Crizotinib, an oral small-molecule multitargeted ALK/c-MET /ROS1 tyrosine kinase inhibitors, was the first-in-class agent approved from FDA for advanced, ALK-rearranged NSCLC. The accelerated approval in 2011 was granted on the basis of pronounced activity observed in early phase I and II clinical studies, coupled with a favorable toxicity-profile and concurrent development of a diagnostic test for ALK rearrangement. [2] More recently, the results from a front-line phase III trial in ALK-positive NSCLC, PROFILE 1014, revealed the superiority, in terms of progression free survival (PFS) and overall response rate (ORR), of crizotinib versus standard pemetrexed-platinum chemotherapy.[3] Based on these data, crizotinib represents standard fist- line therapy in patients with advanced ALK mutant NSCLC. [4] Despite marked and durable initial responses to crizotinib, most patients develop progressive disease after a median of 11 months, with the brain as a common site of relapse. This can be explained by pharmacokinetic limitations rather than a biologic resistance. Several acquired resistance mechanisms have been characterized, including secondary mutations in the ALK kinase domain and/or ALK copy number alterations. ALK-independent resistance mechanisms can also occur through activation of alternative bypass signaling pathways, such as EGFR activation, KIT amplification, KRAS mutation and IGF-R1 activation.[5] This evidence has prompted the development of increasingly potent, selective and brain-penetrant ALK inhibitors, with differential spectrum of activity against the most common resistance mutations. [6] Several next-generation ALK inhibitors, such as ceritinib, alectinib, brigatinib have demonstrated clinical benefit in patients with crizotinib-refractory NSCLC patients also at the central nervous system (CNS) level. This observation has supported the assessment of these drugs as frontline therapy in patients crizotinb-naïve with advanced ALK+ NSCLC.[7] Soria and colleagues have published the results of the ASCEND-4 trial, randomizing ALK+ treatment naïve patients to ceritinib or chemotherapy. Ceritinib treatment significantly has improved median PFS compared to chemotherapy (16.6 vs 8.1 months; hazard ratio [HR] 0.55,P<0.00001). This molecule was also associated with a better control of the disease in the brain (PFS 10.7 vs 6.7 months, HR 0.70, 95% CI: 0.44–1.12). Dose-limiting gastrointestinal adverse events were common with ceritinib at the starting dose of 750 mg daily and 80% of cases required dose reduction or interruption. Although ceritinib has not been compared head-to-head with crizotinib, data confirm ALK inhibitor superior efficacy compared to standard chemotherapy in the ALK-rearranged NSCLC and suggest ceritinib as another option for the front line management. [8] First Line Head to Head trials are ongoing or recently completed. Findings from J-ALEX trial, involving untreated Japanese patients with ALK-rearranged advanced NSCLC, have shown that alectinib induces longer durations of response compared to crizotinib. Median PFS exceeded 2 years in the alectinib group, compared with just over 10 months in the crizotinib group. [9] Recently, Peters et al. have presented the results of global ALEX study. Data are consistent with previous Japanese analysis: PFS was significantly improved with alectinib as compared to crizotinib (25,7 5% vs 10,4%, HR 0.5, p<0.0001). In addition, 12% of patients in alectinib arm vs 45% in crizotinib arm has experienced brain progression (cause specific HR 0.16; 95% CI, 0.10 to 0.28; P<0.001). Alectinib appeared to be better tolerated than crizotinib with grade 3 to 5 adverse events occurring in 41% vs 50% of patients, respectively.[10] Other studies with next-generation ALK inhibitors versus crizotinib—such as lorlatinib, brigatinib, are ongoing and they will help define optimal sequencing therapy for patients with ALK-rearranged NSCLC. To improve outcomes in this patient population, some studies are also currently investigating several combination strategies, including immunotherapy, anti-angiogenetic agents or radiotherapy approach in association with ALK inhibitors as shown in Table 1. Table 1

DRUGS (dose)	Clinical trial (phase)	Patient Number	Comparator	ORR (%)	PFS (mo)
Crizotinib (250 mg twice/day)	PROFILE 1014 (III)	343	Pemetrexed+platinum	74 vs45	10.9 vs 7.0
Ceritinib (750 mg/die)	ASCEND 4 (III)	376	Pemetrexed+platinum pemetrexed±maintenance	72.7vs27.3	16.6 vs 8.1*
Alectinib (300 mg twice/day)	J-ALEX (III)	207	crizotinib (250 mg twice/day)	85vs70	NEvs10.2
Alectinib (600 mg twice/day)	ALEX (III)	303	crizotinib (250 mg twice/day)	82.9vs75.5	25.7vs10.4*
Brigatinib (90 mg/die for 7 days, 180 mg/die)	ALTA 1L (III)	ongoing	Crizotinib (250 mg twice/day)	-
Lorlatinib (150 mg/die)	NCT03052608 (III)	ongoing	Crizotinib (250 mg twice/day)	-
Crizotinib+Bevacizumab (250 mg twice/day; 7.5 mg/kg every 3 wks)	CAMAR01 (II)	ongoing	_	_	_
Crizotinib+Pembrolizumab	NCT02511184 (I)	ongoing	_	_	_
ORR: overall response Rate, PFS Progression free survival, NE not evaluable, *independent review committee

Treatment paradigms continue to evolve for patients with advanced ALK-positive NSCLC subsequently to rapid development of ALK inhibitors history. It is expected that one of the next generation of ALK inhibitors will be used as first-line. In this landscape it is necessary to define the impact of first-line choice on patterns of progression and mechanisms of resistance. [11] It is uncertain if a specific sequence of therapeutic agents influences the biology of the cancer and therefore the clinical course of the patient. The spectrum of ALK resistance mutations varies according to ALK inhibitor and it is unclear if the mechanisms of resistance to these agents as the first ALK inhibitor will be similar to the mechanisms of resistance identified when they are used after crizotinib. Future efforts should be focused on determining the best treatment sequence to maximize clinical outcomes. Key factors to guide the selection of therapies could be include: molecular characteristics of the patient's tumor, different toxicity profile of different ALK inhibitors, availability of combinations/ multimodal therapy. References 1. Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer. Soda M et al. Nature 2007; 448:561–567 2. The continuum of care for ALK-positive NSCLC: from diagnosis to new treatment options-an overview Solomon & Soria, Ann Oncol Vol 27 Supp 3 2016 3. First-line crizotinib versus chemotherapy in ALK-positive lung cancer Solomon et al N Engl J Med 2014 Dec 4;371(23):2167-77 4. Metastatic non-small-cell lung cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Novello S et al. Annals of Oncology 27 (Supplement 5): v1–v27, 2016 5. Crizotinib resistance: implications for therapeutic strategies. Dagogo-Jack & Shaw Ann Oncol Supp 3 2016 6. Molecular Mechanisms of Resistance to First- and Second-Generation ALK Inhibitors in ALK-Rearranged LungCancer. Gainor et al Cancer Discov2016Oct;6(10): 1118-1133. 7. Ascending role of next-generation ALK inhibitors. Costa. Lancet Oncol.2017 Jul; 18(7):837-839. 8. First-line ceritinib versus platinum-based chemotherapy in advanced ALK-rearranged non-small-cell lung cancer (ASCEND-4): a randomized, open-label, phase 3 study. Soria et al Lancet 2017; 389: 917–29 9. Alectinib versus crizotinib in patients with ALK-positive non-small-cell lung cancer (J-ALEX): an open-label, randomized phase 3 trial Hida et al. Lancet 2017; 390: 29–39 10. Alectinib versus Crizotinib in Untreated ALK-Positive Non–Small-Cell Lung Cancer Peters et al for for the ALEX Trial Investigators NEJM N Engl J Med. 2017 Jun 6. 11. First-line treatment options for ALK-rearranged lung cancer. Solomon. Lancet Oncology 2017 Mar 4; 389(10072): 884-886

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Abstract not provided

Background:
Mutant neoantigens generated from genetic alterations that are exclusively present in tumors represent highly promising cancer vaccine targets. However, publically available neoantigen prediction algorithms only identify and utilize single nucleotide mutations (SNVs) but not short insertion and deletions (INDELs). Short INDELs can lead to the generation of novel junctional or frameshift neoantigens which may be more immunogenic than neoantigens that result from single nucleotide missense mutations.

Method:
We developed a bioinformatics pipeline for neoantigen prediction using paired normal tissue and tumor exome sequencing, RNA sequencing and HLA binding prediction. 536 lung adenocarcinoma (LUAD) and 466 lung squamous cell carcinoma (LUSC) cases were analyzed using our bioinformatics pipeline. The non-synonymous somatic SNVs and short INDELs mutations were identified to generate a list of mutation neoantigen-derived and, when possible, their corresponding wild-type epitopes. Binding affinities of the paired wild-type and mutant peptides to HLA class I were then predicted and compared.

Result:
On average, 8.65 (range1-158) mutant neoantigen peptides per sample were identified in 395 out of 536 (73.6%) LUAD samples. Among them, 63.7% were SNVs and 36.3% were INDELs. On average, 8.54 (range 1-504) mutant neoantigen peptides per sample were identified in 360 out of 466 LUSC samples. Among those, 67% were SNVs and 33% were INDELs. Most neoantigen peptides are private in both LUAD and LUSC. The mutant neoantigen peptides identified from INDELs were predicted to have 3.9 (p = 2.42E-74) and 1.14 (p = 5.44E-67) fold higher HLA class I binding affinity than wild type peptide compared to those from SNVs in LUAD and LUSC respectively.

Conclusion:
Tumor INDELs may be a rich source of neoantigens with a higher predicted high HLA binding affinity in lung cancers that warrant consideration in development of a personalized cancer vaccine.

Background:
The genomic landscape of primary resistance to PD-1 blockade in lung adenocarcinoma (LUAD) is largely unknown. We previously reported that co-mutations in STK11/LKB1 (KL) or TP53 (KP) define subgroups of KRAS-mutant LUAD with distinct therapeutic vulnerabilities and immune profiles. Here, we present updated data on the clinical efficacy of PD-1/PD-L1 inhibitors in co-mutation defined KRAS mutant and wild-type LUAD patients and examine the relationship between genetic alterations in individual genes, tumor cell PD-L1 expression and tumor mutational burden (TMB) using cohorts form the SU2C/ACS Lung Cancer Dream Team and Foundation Medicine (FM).

Method:
The cohorts included 924 LUAD with NGS (FM cohort) and 188 patients with KRAS non-squamous NSCLC (SU2C cohort) who received at least one cycle of PD-1/PD-L1 inhibitor therapy and had available molecular profiling. Tumor cell PD-L1 expression was tested using E1L3N IHC (SU2C) and the VENTANA PD-L1 (SP142) assay (FM). TMB was defined as previously described and was classified as high (TMB-H), intermediate (TMB-I) or low (TMB-L).

Result:
188 immunotherapy-treated (83.5% nivolumab, 11.7% pembrolizumab, 4.8% anti-PD1/PD-L1 plus anti-CTLA-4) pts with KRAS-mutant NSCLC were included in the efficacy analysis. The ORR differed significantly between the KL (8.8%), KP (35.9%) and K-only sub-groups (27.3%) (P=0.0011, Fisher’s exact test). KL LUAC exhibited significantly shorter PFS (mPFS 1.8m vs 2.7m, HR=0.53, 95% CI 0.34-0.84, P<0.001, log-rank test) and OS (mOS 6.8m vs 15.6m, HR 0.53, 95% CI 0.34 to 0.84, P=0.0072, log rank test) compared to KRAS-mutant NSCLC with wild-type STK11. Loss-of function (LOF) genetic alterations in STK11 were the only significantly enriched event in PD-L1 negative, TMB-I/H compared to PD-L1 high positive (TPS≥50%), TMB-I/H tumors in the overall FMI cohort (Bonferroni adjusted P=2.38x10[-4], Fisher’s exact test) and among KRAS-mutant tumors (adjusted P=0.05, Fisher’s exact test) . Notably, PD-1 blockade demonstrated activity among 10 PD-L1-negative KP tumors, with 3 PRs and 4SDs recorded. In syngeneic isogenic murine models PD-1 blockade significantly inhibited the growth of Kras mutant tumors with wild-type LKB1 (K), but not those with LKB1 loss (KL), providing evidence that LKB1 loss can play a causative role in promoting PD-1 inhibitor resistance.

Conclusion:
Loss of function genomic alterations in STK11 represent a dominant driver of de novo resistance to PD-1/PD-L1 blockade in KRAS-mutant NSCLC. In addition to tumor PD-L1 status and tumor mutational burden precision immunotherapy approaches should take into consideration the STK11 status of individual tumors.

Background:
Baseline high derived NLR (dNLR>3, neutrophils/(leucocytes-neutrophils) ratio) has recently correlated with no benefit to immune checkpoint inhibitors (ICI) in advanced NSCLC, but the dynamic monitoring of dNLR has not been assessed in this population.

Method:
dNLR at baseline, at 2[nd] cycle and at progressive disease were retrospectively collected in advanced NSCLC patients treated with ICI from November 2012 to April 2017, in a multicentric cohort (N= 292) from 4 European centers. The primary endpoint was overall survival (OS), and secondary endpoints were progression free survival (PFS), response rate (RR) and disease control rate (DCR).

Result:
Out of 292 patients (67%) were males, 264 (92%) smokers and 239 (83%) with PS ≤1, with median age 64 years; 153 (52%) had adenocarcinoma and 114 (30%) squamous; 44 (15%) were KRASmut, 11 (4%) EGFRmut and 3 (1%) ALK positive. PDL1 was ≥ 1% by immunohistochemistry in 67 (76%), negative in 21 (24%) and unknown in 204 patients. The median of prior lines was 1 (0-10). The median follow-up was 12 months (m) [11-14]. The median PFS and OS were 4m [3-5] and 11m [9-15]. Baseline dNLR was>3 in 106 patients (36%) and at 2[nd] cycle in 90 patients (32%). dNLR>3 at baseline and at 2[nd] cycle were associated with poor PFS (p<0.0001 and p=0.0008, respectively), poor OS (both p<0.0001) and progressive disease (p=0.002 and p=0.005, respectively). At 2[nd] cycle of ICI, the dNLR status (> high or ≤ 3 low) changed in 63 patients: in 38 (14%) dNLR decreased; in 25 (9%) dNLR increased. According to the dNLR monitoring (combining dNLR at baseline et at 2[nd] cycle), the median OS was 17m (95%CI 13-NA) when dNLR remained low (n=153), 10m (95%CI 7-NA) when dNLR changed (n=64) and 4m (95%CI 3-7) when dNLR remained high (dNLR>3, n=64, p<0.0001).The dNLR monitoring was also associated with PFS (p=0.002), RR and DCR (p=0.003 and p=0.013, respectively).

Conclusion:
Monitoring dNLR at baseline and at 2[nd] cycle could be a routinely tool to early assess benefit to ICI in NSCLC patients on treatment. The dNLR monitoring showed a strong correlation with OS and PFS. Modification of dNLR between baseline and 2[nd] cycle impacts outcomes in NSCLC patients treated with ICI.

Background:
Recent clinical trials have demonstrated the efficacy of immune checkpoint inhibitors in advanced non-small cell lung cancer (NSCLC). However, not all the patients receive survival benefit from these immunotherapies. In an attempt to refine the current strategy of cancer immunotherapy to treat NSCLC, we examined the influence of tumor-infiltrating lymphocytes (TILs) on postoperative survival.

Method:
We evaluated the prognostic significance of TILs (CD4[+], CD8[+], and FOXP3[+]) comprehensively by immunohistochemical (n = 234) and immune-related gene expression analysis (n = 58), and explored the relationship between immune features and clinical characteristics including histological types, smoking habit, epidermal growth factor receptor mutation, and postoperative survival.

Result:
Compared with non-adenocarcinoma (non-AD) patients, adenocarcinoma (AD) tumors had significantly higher number of tumor-infiltrating CD4[+] T cells (P < 0.05) but lower CD8[+] T cells and FOXP3[+] T cells (P < 0.01). We found higher accumulation of CD8[+] T cells in non-AD patients was correlated with longer survival, indicating it is a better prognostic factor (P < 0.02). On the contrary, high accumulation of CD8[+] T cells and FOXP3[+] T cells were identified as unfavorable prognostic factors (P < 0.05) in AD patients, particularly in AD nonsmokers (P < 0.02). The expression of activated T cell-related genes including interferon gamma and granzyme was associated with CD8[+] T-cell accumulation in non-AD patients, but not in AD patients, especially in AD nonsmokers. Infiltrating CD8[+] T cells were significantly less activated in immunosuppressive microenvironment with high expression of immunoregulation related genes including GATA3, IL13, CCR4 and CCL17 in AD nonsmokers (P < 0.05). In AD nonsmokers, there are possibly immunodysfunctional CD8[+] GATA3[+] T cells (P < 0.01) and immunoregulatory CD8[+] FOXP3[+] T cells (P < 0.01), accompanied by immunoregulatory CD4[+] FOXP3[+] CCR4[+] T cells (P < 0.01) that may be recruited by CCL17 produced by tumor-associated CD163[+] macrophages (P < 0.05) in IL13-associated tumor microenvironments (P < 0.05).

Conclusion:
In contrast to presence of activated CD8[+] T cells in non-AD, CD8[+] T cells are not activated, and may include dysfunctional and immunoregulatory T cells, accompanied by FOXP3[+] regulatory T cells and M2-like macrophages in IL13-associated tumor microenvironment of AD nonsmokers. Our study suggests that modulation of such immunosuppressive condition may be an attractive strategy for treatment of AD nonsmokers including immune-checkpoint blockade.

Abstract not provided

Background:
The checkpoint inhibitors programmed cell death and its ligand (PD-L1) antibodies are promising treatment agents for patients with advanced non-small cell lung cancer (NSCLC). Their clinical efficacy is predicted by drug-tailored PD-L1 immunohistochemistry (IHC) assays. We aimed to identify the similarity and distinction of 22C3 and 28-8 IHC tests.

Method:
Three hundred and ninety consecutive cases of completely resected NSCLC between January 2009 and September 2014 that had adequate tissue samples were investigated. From the archived samples, 5-μm thick sections were cut and stained with PD-L1 IHC 22C3 PharmDx and 28-8 PharmDx (Dako, Santa Clara, CA). The staining and evaluation in 22C3 and 28-8 test were performed by two separate laboratories. PD-L1 expression and high PD-L1 expression were defined as ≥1% and ≥50% of tumor cells stained, respectively. Statistical significance was defined as a p-value of <0.05.

Result:
The study population included 288 patients with adenocarcinomas, 70 with squamous cell carcinomas, 18 with large cell carcinomas, 9 with adenosquamous carcinoma and 5 with pleomorphic carcinoma. Two hundred and ninety-three patients had pStage I; 47, pStage II; and 46, p Stage IIIA tumors. Two hundred and twenty-nine specimens showed no PD-L1 expression with either 22C3 or 28-8. The detection rate of PD-L1 expression was 36.9% (144 cases) with 22C3 and 35.6% (139 cases) with 28-8, respectively (p= 0.710). The detection rate of high PD-L1 expression was 16.9% (66 cases) with 22C3 and 9.0% (35 cases) with 28-8 (p= 0.0013). The Spearman correlation coefficient was 0.866 (95% confidence interval, 0.838–0.890; p< 0.0001). Figure 1



Conclusion:
22C3 IHC assay may be more sensitive to detect high PD-L1 expression in NSCLC compard to 28-8 IHC assay, whereas 22C3 and 28-8 showed no significant difference to detect PD-L1 expression. Further investigation is necessary to reveal clinical, pathological and molecular background. This approach will help better interpretation of PD-L1 IHC results.

Background:
Tumor PD-L1 IHC relates moderately with treatment outcome following anti-PD1 therapy in pts with NSCLC and single biopsies do not account for tumor heterogeneity. Aim: 1. Assess safety of the PET procedures. 2. Quantify [89]Zirconium-labeled nivolumab ([89]Zr-nivo) and [18]F-labeled BMS-986192 ([18]F-PD-L1) uptake. 3. Assess tracer uptake heterogeneity. 4. Correlate tracer uptake with PD-1/PD-L1 IHC in tumor, stroma and with treatment outcome.

Method:
NSCLC pts eligible for treatment with nivolumab were included. Pts received whole body [18]F-PD-L1 and [89]Zr-nivo PET scans. Baseline tumor biopsy was required to assess PD-(L)1 IHC status (28.8 assay). SUV~peak~ was calculated for delineable lesions and correlated to PD-(L)1 IHC and response after 12 wks of nivolumab treatment.

Result:
10 pts (3 ≥50%, 5 ≥1%, 5 negative by PD-L1 IHC) were enrolled and 37 lesions analysed. No toxicity related to radiotracer was observed. Tumor uptake of both tracers was visualized in all pts, but not in all lesions. Tracer uptake varied among pts with mean [18]F-PD-L1 SUV~peak~ 4.6, range 0.5 - 14.4 and mean [89]Zr-nivo SUV~peak~ 5.0, range 1.6 – 11 (p=0.03) and within pts with mean SUV~peak~ difference 3.6-fold (±2.1) and 2.4-fold (±0.77) between lesions for [18]F-PD-L1 and [89]Zr-nivo, respectively. For lesions with ≥50% PD-L1 IHC, mean [18]F-PD-L1 SUV~peak~ was 8.0 (±4.7) as compared to 3.5 (±1.6) for lesions with <50% PD-L1 IHC (p=0.03). For tumors with high TIL/ stromal PD-1 expression, mean [89]Zr-nivo SUV~peak~ was 8.6 (±2.4) as compared to 6.1 (±2.1) for lesions with low PD-1 expression (p=0.1). Mean SUV~peak ~for [18]F-PD-L1 was 8.4 (±5.4) for pts with PR and 4.5 (±2.9) for pts with PD/SD (p=0.3). Mean SUV~peak~ for [89]Zr-nivo was 7.8 (±1.8) for pts with PR and 5.4 (±2.2) for pts with PD/SD (p=0.2).

Conclusion:
1. PET-imaging with both tracers is safe and feasible, with good tumor-to-normal tissue contrast. 2. Tumor uptake showed heterogeneity among pts and among tumors within pts. 3. Pts with ≥50% tumor PD-L1 expression showed higher [18]F-PD-L1 uptake. 4. Pts with high PD-1 expression showed higher [89]Zr-nivo uptake, and pts with PR demonstrated higher [18]F-PD-L1 and [89]Zr-nivo tracer uptake than pts with PD/SD, although these are without statistical significance which may be due to the small dataset.

Background:
Immunotherapy has become one of the options among the treatments of lung cancer. Nivolumab was first proven to have the utility as a second line treatment for non-small cell lung cancer. However, predictive factor of its efficacy is unknown. In recent years, studies have evolved on circulating tumor DNA (ctDNA). Clinical applications expanded and included prediction of prognosis, monitoring treatment effects and acquired resistance of driver genes, and assessment of residual tumor burden of resected cancer. In this study, we took cases in which tumor tissue was surgically resected or obtained by biopsy and the corresponding somatic mutations in plasma were studied. Then, we used these somatic mutations presumably derived from original tumor tissue as “tumor markers”. We took serial blood samples before and after starting nivolumab and examined to see whether early change of the level of ctDNA can predict long term treatment outcomes.

Method:
Fourteen patients who were treated by nivolumab from February 1st to September 30th in 2016 were studied. Peripheral blood samples were collected from the patients before, 1, 2, 4, 6 and 8 weeks after the initiation of nivolumab treatment. To identify somatic mutations in tissue and total plasma DNA, we performed targeted sequencing using “lung cancer panel” spanning whole exons of these 53 genes, and next generation sequencing was performed. Gene mutations detected in both tumor tissue and plasma were defined here as circulating tumor DNA (ctDNA). Early response of the level of ctDNA after starting nivolumab was evaluated to see whether it could predict treatment outcome.

Result:
Of 14 cases, 6 cases were Responders, and 8 Non-responders. ctDNA was detected more often in the serial plasma samples of patients carrying high tumor burden (p=0.02). In addition, basal and serial ctDNA analysis revealed that decrease of allelic frequency (AF) level within 2 weeks correlated with the good durable response, and on the contrary, the increase with no or poor response.

Conclusion:
In patients carrying high tumor burden, plasma analysis of ctDNA which was validated by tumor tissue, revealed the durable good response of nivolumab could be predicted within 2 weeks.

Background:
Association between T-effector (Teff) gene expression (GE), a marker of pre-existing immunity, and OS benefit with atezolizumab (anti–PD-L1) was demonstrated in the Phase II study POPLAR of atezolizumab vs docetaxel in 2L+ NSCLC. We analyzed Teff GE association with atezolizumab efficacy in a larger Phase III study, OAK.

Method:
Patients with 2L+ NSCLC were randomized to receive atezolizumab or docetaxel. Teff signature was defined by 3 genes (PD-L1, CXCL9, and IFNγ), and Teff GE was measured by averaging the normalized expression of each gene. Teff GE subgroups were defined by quartiles. PD-L1 expression was assessed using the SP142 IHC assay; the TC1/2/3 or IC1/2/3 subgroup had ≥ 1% PD-L1 expression on tumor cells (TC) or tumor-infiltrating immune cells (IC).

Result:
753 of 850 patients from the OAK primary analysis constituted the biomarker evaluable population (BEP) for Teff GE. Expression of the Teff signature was associated with PD-L1 expression by IHC (P = 7.3×10[−45]). Although no significant PFS benefit with atezolizumab vs docetaxel was observed in the BEP (HR, 0.94 [95% CI: 0.81, 1.10]) or the TC1/2/3 or IC1/2/3 subgroup (HR, 0.93 [95% CI: 0.76, 1.15]), a gradient of improved PFS benefit with atezolizumab was observed with increasing Teff GE. Significant PFS benefit occurred with ≥ median Teff GE cutoff (HR, 0.73 [95% CI: 0.58, 0.91]; Table). Teff GE also enriched for improved OS; however, a trend toward OS benefit was still observed in patients with low Teff GE (Table).

Table. PFS and OS with atezolizumab vs docetaxel by PD-L1 IHC and Teff GE subgroups
	PFS, HR (95% CI)	OS, HR (95% CI)
OAK primary population (N = 850)[a]
ITT[a]	0.95 (0.82, 1.10)	0.73 (0.62, 0.87)
TC1/2/3 or IC1/2/3[a ](n = 463)	0.91 (0.74, 1.12)	0.74 (0.58, 0.93)
TC2/3 or IC2/3[a] (n = 265)	0.76 (0.58, 0.99)	0.67 (0.49, 0.90)
OAK BEP for Teff GE (N = 753)
BEP	0.94 (0.81, 1.10)	0.71 (0.59, 0.85)
TC1/2/3 or IC1/2/3 (n = 420)	0.93 (0.76, 1.15)	0.74 (0.58, 0.95)
Teff GE subgroups
≥ 25% (n = 570)	0.91 (0.76, 1.09)	0.67 (0.54, 0.83)
< 25% (n = 183)	1.11 (0.82, 1.49)	0.87 (0.63, 1.21)
≥ 50% (n = 379)	0.73 (0.58, 0.91)	0.59 (0.46, 0.76)
< 50% (n = 374)	1.30 (1.05, 1.61)	0.87 (0.68, 1.11)
≥ 75% (n = 190)	0.66 (0.48, 0.91)	0.60 (0.42, 0.87)
< 75% (n = 563)	1.10 (0.92, 1.31)	0.76 (0.62, 0.92)
[a]Rittmeyer A. et al. Lancet, 2017;389:255-265. NCT02008227.

Conclusion:
This is the first demonstration of the association between markers of Teff biology and clinical outcomes with cancer immunotherapy in a randomized Phase III trial. Teff GE may reflect pre-existing immunity and be a more sensitive biomarker compared with PD-L1 IHC, identifying more patients (50% prevalence) likely to experience PFS benefit with atezolizumab.

Abstract not provided

Background:
Immunotherapy has become a promising recourse for lung cancer therapy. The endothelium separates circulating immune cells and the tumor microenvironment, and it is necessary for immune cells to penetrate this barrier to accost the tumor. This requires cell-matrix interactions via endothelial adhesion molecules(EAM) such as selectin and integrin. While it is expected that higher expression of EAM is linked to greater immune cell infiltration in general, little is known as to its actual effect on various types of immune cells in human lung cancer.

Method:
Based on the TCGA database, mRNA-seq values of genes related to the leukocyte recruitment cascade were analyzed in 504 patient samples with lung squamous cell carcinoma(SCC) and 522 patient samples with lung adenocarcinoma. The genes were divided into 3 sets: rolling, firm adhesion, and transmigration. Immune cell infiltration of each set was analyzed using Gene Set Enrichment Analysis(GSEA), and p values were derived from Fisher’s exact and Chi-squared tests.

Result:
In lung SCC, overexpression(z score>2.0) of the above genes was statistically significantly correlated with decreased infiltration of activated CD4/CD8 T cells, but increased infiltration of activated B/Treg cells (Figure1). Similar trend was also observed in lung adenocarcinoma. Macrophage, dendritic cells, and natural killer cells showed increased infiltration in the EAM overexpression groups of both SCC and adenocarcinoma. Overall survival showed no significant difference in all three EAM gene overexpression groups in both types of lung cancer.Figure 1



Conclusion:
We demonstrate for the first time that overexpression of EAM genes is linked to differential infiltration of various immune cells (including decreased CD4/CD8 T cells and increased activated B/Treg cells) in human lung cancer tissue. Recruitment of B/Treg cells by EAM may have an impact on inactivation of infiltrated T cells in the tumor microenvironment. This suggests that EAM status may serve as a predictive biomarker for T cell-mediated immunotherapy.

Background:
Lung cancer development is driven by the expression of mutant oncogenes, with EGFR and KRAS the most frequent in lung adenocarcinoma. However, these mutations alone are not sufficient for tumorigenesis suggesting additional factors influence tumour development and progression, including the balance of anti-tumour immune effector cells and pro-tumorigenic immune suppressor cells. Tumour cells can evade immune surveillance by producing cytokines to recruit immune modulatory cells that promote an immune suppressive environment, such as regulatory T cells (Tregs). We hypothesized that oncogene signaling regulates the production of cytokines by tumour cells in order to recruit immune suppressive cells and promote lung tumour development.

Method:
We used CIBERSORT to quantify 22 immune cell types in over 300 human lung adenocarcinoma and 100 matched normal lung tissues, and validated findings with immunohistochemistry. Cells expressing doxycycline inducible EGFR[L858R] and KRAS[G12V]were analyzed for cytokine production using a multiplex assay (LUMINEX). EGFR (Afatinib) and MEK (Trametinib) inhibitors were used in lung cancer cell lines harbouring EGFR or KRAS mutations and cytokine production was quantified using ELISA. Conditioned media from EGFR[L858R] and KRAS[G12V] expressing cells were used in a trans-well assay to determine if secreted cytokines could induce Treg migration. Transgenic mouse models of lung adenocarcinoma and bronchoalveolar lavage (BAL) from patients with and without lung cancer were used to assess CCL5 and Tregs in vivo.

Result:
Treg cells were significantly enriched in lung tumours and not normal tissue. CCL5 production is increased rapidly upon oncogene induction and subsequent transformation of normal cells and is dependent on sustained ERK signaling for continued expression. Conditioned medium from EGFR[L858R] expressing cells increased Treg migration, which was mitigated by an anti-CCL5 antibody. Transgenic mice expressing EGFR[L858R ]or KRAS[G12D] in the lung epithelium recruited Tregs to the lung upon tumor induction. Assessment of CCL5 in BAL from patients with and without lung cancer is currently in progress.

Conclusion:
Oncogene driven ERK signaling may regulate expression of CCL5 from lung tumour cells, and oncogene induced CCL5 production stimulates Treg migration ex vivo. These data suggest CCL5-mediated Treg recruitment to lung tumours may occur in early stages of lung tumour development and that targeted inhibition of CCL5 or ERK signaling may represent therapeutic strategies to block recruitment of immunosuppressive Tregs by lung tumours.

Background:
Tumor-associated macrophages (TAMs) with immunosuppressive and tumor promoting features are attractive targets for immunotherapy. MARCO is a scavenger receptor expressed on a subpopulation of macrophages in secondary lymphoid organs. A recent study performed in animal models concluded that treatment with an anti-MARCO antibody results in reprogramming of the TAMs and inhibition of tumor growth and metastatic spread. The expression and function of MARCO in lung cancer TAMs is not known.

Method:
The infiltration of TAMs expressing MARCO, CD68, CD163 and MSR1, in the tumor and stromal compartments, was analyzed by immunohistochemistry in a non-small cell lung cancer (NSCLC) cohort (n=352). In addition, PD-L1 expression was assessed on tumor cells. Immunofluorescence was performed on selected cases to evaluate marker co-expression. Associations to immune cells and regulatory inflammatory pathways were studied in a subset of cases (n=174) with available RNA-seq data.

Result:
A large variance in TAM density could be observed between cases as well as a strong correlation between CD68 and CD163, indicating a pro-tumor phenotype of infiltrating macrophages. Correlation to clinical data showed a trend towards worse survival for patients with high macrophage infiltration. TAM expression of MARCO was seen on a subpopulation of pro-tumor macrophages. The majority of MARCO expressing TAMs were found to be located within tumor cell nests. Interestingly, stromal macrophages expressing MARCO tended to aggregate in close proximity to the tumor nests. On the transcriptomic level, increased MARCO gene expression correlated to genes linked to immunosuppressive TAMs, T-cell infiltration and immune checkpoint molecules like PD-L1 and CTLA-4. The association between macrophage infiltration and tumor cell PD-L1 expression was confirmed by immunohistochemistry. Also, co-expression of PD-L1 and MARCO could be detected on certain macrophages within the tumor cell nests.

Conclusion:
MARCO expression characterizes a specific subpopulation of pro-tumor macrophages that are enriched in PD-L1 positive NSCLC cases. Patients with significant infiltration of MARCO positive TAMs could benefit from treatment with anti-MARCO antibodies, possibly in combination with available immune checkpoint inhibitors.

Background:
STING is a protein that promotes type I IFN production (IFNα/β) essential for activation of dendritic cells and antigen presentation and priming of T-cells. The cytoplasmic DNA sensor cGAS (cGAMP Synthase) is able to detect tumor DNA, and in response will synthesize cGAMP. cGAMP binds STING specifically, resulting in production of type I IFN. STING is therefore referred to as an adaptor protein essential for immune signaling following detection of tumor DNA. Analysis of the TCGA database indicates decreased survival in lung adenocarcinoma patients lacking STING expression. STING expression is decreased in tumor tissues and can be lost as the tumor progresses. One reported mechanism of loss of STING or cGAS in tumors is due to hypermethylation, a common occurrence in lung cancer. Agonists of STING show potent immune response and are currently in clinical trials. Importantly, recent studies show that expression of STING and cGAS proteins are essential for response to PD-1:PD-L1 blockade.

Method:
Section not applicable

Result:
We analyzed 55 NSCLC and 39 SCLC cell lines, and 317 NSCLC and 78 SCLC tissues for STING and cGAS expression using immunohistochemistry. 14/55 (25.45%) NSCLC cell lines and 25/39 (64.10%) SCLC cell lines showed no STING expression. Separated in to adenocarcinoma (AC) and squamous cell carcinoma (SCC) subsets, STING expression in AC shows loss of STING as tumor stage increases (Positive: 70% Stage I, 65% Stage II, 52% Stage III, 40% Stage IV, 71% total; n=156) while STING expression is low at all stages of SCC (Positive: 29% Stage I, 18% Stage II, 36% Stage III, 13% Stage IV, 27% total; n=161). SCLC tissues stained showed widespread loss of STING (Positive: 40% Stage I, 27% Stage II, 31% Stage III, 100% Stage IV, 37.18% total, n=78). Expression of cGAS was higher in AC (94%) than SCC (75%) and showed no correlation with stage. TCGA analysis of STING methylation shows hypermethylation in AC (0.15- ± 0.13 tumor vs 0.05 ± 0.02 normal, n=422) and SCC (0.23 ± 0.16 tumor vs 0.04 ± 0.03 normal, n=359). cGAS shows slight methylation in AC (0.05 ± 0.07 tumor vs 0.05 ± 0.01 normal, n=422) but a large increase in SCC (0.19 ± 0.24 tumor vs 0.04 ± 0.01 normal, n=359).

Conclusion:
This study indicates drastic differences in STING and cGAS expression in AC, SCC, and SCLC. Differential expression of these proteins could impact the efficacy of STING agonists, radiation therapy, and immunotherapy in lung cancer.

Abstract not provided

Background:
Anti-PD1-monotherapy has activity in NSCLC with improved outcomes compared to chemotherapy. Preclinical, and early clinical data, suggests that combining immune checkpoint inhibitors and platinum-based chemotherapy may be synergistic. We examined the cohort of patients (pts) with metastatic NSCLC, with no prior therapy for advanced disease, from this multicentre phase Ib trial. The primary objective was to establish the recommend phase II doses of durvalumab (Du) ± tremelimumab (Tr) in combination with platinum-doublet chemotherapy. Secondary objectives included assessing safety, tolerability, and antitumour activity of the 4-drug combination.

Method:
Pts were treated with Du±Tr at one of 4 dose levels (DL) concomitantly with either pemetrexed (Pem) +cisplatin/carboplatin followed by maintenance Pem for nonsquamous histology or gemcitabine (Gem) +cisplatin/carboplatin for squamous tumours. At DL0 Du 15 mg/kg IV q3wks was added; DL1 added Du 15mg/kg q3wk + Tr 1mg/kg x1 dose; DL2a added Du 15mg/kg q3wk + Tr 1 mg/kg q6wk x multiple doses; DL2b added Du 15mg/kg q3wk + Tr 3 mg/kg q6wk (1 dose with cycle 1 and 2 doses with maintenance pem), DL3 and DL4 added a fixed doses of Du 1125mg/Tr 56 mg and Du 1500 mg/ Tr 75 mg q3wk respectively. Du could continue until 1 year or unacceptable toxicity.

Result:
To date, 45 pts (median age=62 (range 36-78); 44% male, 100% ECOG PS≤1) have received 346 cycles in the Pem-platinum cohort while 9 pts (median age=64 (range 57-80); 78% male, 100% ECOG PS≤1) have been received 55 cycles in the Gem-platinum group. Most adverse events were ≤grade 2 and attributed to chemotherapy. Immunotherapy-related adverse events (irAEs) ≥ grade 3 were observed in 27% of patients and were more commonly reported with the addition of Tr. In the Pem-platinum cohort, diarrhea (n=5), fatigue (n=4) and elevated lipase (n=4) were the most come irAEs ≥ grade 3, while rash, pneumonitis and hypothyroidism occurred in 1 pt each after the introduction of Tr. In the Gem-platinum cohort, irAEs ≥ grade 3 were fatigue and rash (1 pt each at DL2b) and elevated lipase (1 pt at DL3). The objective response rate in 35 evaluable patients receiving Pem-platinum was 57.1% (95% CI=39.4, 73.7 ) and 37.5% (95% CI=8.5, 75.5) in the 8 evaluable patients receiving Gem-Platinum.

Conclusion:
The combination of Du±Tr can be safely administered with platinum-doublet chemotherapy with encouraging preliminary response data. Adding Tr may increase ≥ grade 3 irAE hence patient selection and early intervention is key to managing irAEs. *contributed equally

Background:
Chemotherapy may enhance immunotherapeutic effects by causing tumor antigen release, which primes the immune system to kill tumor cells. Early clinical data on nab-paclitaxel + carboplatin in combination with immune checkpoint inhibitors (ICI) demonstrated promising activity without compounding toxicities in patients with non-small cell lung cancer (NSCLC). ABOUND.2L+ evaluated nab-paclitaxel–based regimens in previously treated patients with advanced NSCLC. Here we report the efficacy and safety of nab-paclitaxel + durvalumab as second/third-line treatment.

Method:
Patients with advanced NSCLC were assigned to receive second/third-line (immunotherapy allowed in prior line, including platinum doublet combination) nab-paclitaxel 100 mg/m[2] on days 1 and 8 + durvalumab 1125 mg on day 15, in 21-day cycles, administered until unacceptable toxicity/progression per immune-related RECIST v1.1. Primary endpoint is progression-free survival (PFS). Secondary endpoints are overall survival (OS), overall response rate (ORR), disease control rate (DCR), and safety.

Result:
Seventy-nine patients were enrolled. Median age was 63 years, 68% of patients were male, 23% had Eastern Cooperative Oncology Group performance status of 0, and 70% had nonsquamous NSCLC; 11% of patients received prior ICIs. Median PFS (Table) and OS were 4.5 (3.4-5.8) months and NE (7.3-NE). ORR was 27% (1 complete response) and DCR was 71%. Grade 3/4 treatment-emergent adverse events of special interest occurring in ≥ 5% of patients included neutropenia (6%) and dyspnea (5%); grade 3/4 peripheral neuropathy and anemia each occurred in 4% of patients. Median treatment duration was 24 weeks; median number of treatment cycles was 7. For nab-paclitaxel and durvalumab, median dose intensities were 59.05 mg/m[2]/week and 326.61 mg/week, respectively; median percentages of per-protocol dose were 88.58% and 87.10%.

Conclusion:
The combination of durvalumab with nab-paclitaxel demonstrated antitumor activity with manageable toxicity in the second/third-line setting. Further details will be presented. NCT02250326

Nab-P Durva	Median PFS (range), months
Overall (n = 79)	4.5 (3.4-5.8)
ICI pretreated (n = 9)[a] ICI naive (n = 69)[a]	6.9 (1.4-NE) 4.4 (3.0-5.7)
Squamous (n = 23)[a] Nonsquamous (n = 55)[a]	5.9 (3.0-7.8) 4.2 (2.9-5.7)
ICI, immune checkpoint inhibitor; NE, not estimable; PFS, progression-free survival. [a] Data pending for 1 patient.

Background:
Atezolizumab (anti–PD-L1) has demonstrated OS benefit over docetaxel in a randomized Phase II study, POPLAR, in patients with advanced NSCLC. This benefit has been confirmed in the randomized Phase III study OAK (Rittmeyer, 2017). The 3-year survival analysis of the POPLAR study presented here describes the longest survival follow-up reported to date of an all-comer randomized PD-L1/PD-1 immunotherapy trial in 2L+ NSCLC.

Method:
Patients were randomized 1:1 to receive atezolizumab (1200 mg) or docetaxel (75 mg/m[2]) IV q3w. Tumors were prospectively evaluated for tumor cell (TC) or tumor-infiltrating immune cell (IC) PD-L1 expression using the VENTANA SP142 IHC assay. Landmark OS was estimated using the Kaplan-Meier method. Data cutoff, April 7, 2017; minimum follow-up, 3 years.

Result:
The 2-year and 3-year survival with atezolizumab vs docetaxel were 32.2% vs 16.6% and 18.7% vs 10.0%, respectively. The long-term OS benefit of atezolizumab vs docetaxel was observed across histology and PD-L1 expression subgroups (Table). While the TC3 or IC3 subgroup derived the greatest OS benefit, the TC0 and IC0 subgroup also had improved long-term OS with atezolizumab vs docetaxel. The ITT ORR was 15% in both atezolizumab and docetaxel arms, but the median duration of response was 3 times longer with atezolizumab (22.3 months [95% CI: 11.6, 31.1] vs 7.2 months [95% CI: 5.8, 12.2] with docetaxel). Seven of the 11 docetaxel-arm 3-year survivors received subsequent non-protocol therapy with anti–PD-L1/PD-1 agents. Atezolizumab had a favorable safety profile compared with docetaxel that was consistent with previous reports.

Conclusion:
Atezolizumab demonstrates superior 2-year and 3-year OS benefit compared with docetaxel, and this benefit is observed across histology and PD-L1 expression subgroups (including TC0 and IC0). Atezolizumab is well tolerated, and responses are highly durable. These results are consistent with long-term OS results from OAK, presented separately.

Table. Landmark OS in the ITT, PD-L1 expression, and histology subgroups in POPLAR
Population (n, atezolizumab vs docetaxel)	2-year OS rate, %			3-year OS rate, %
	Atezolizumab	Docetaxel	P value[a]	Atezolizumab	Docetaxel	P value[a]
ITT (144 vs 143)	32.2%	16.6%	0.0027	18.7%	10.0%	0.0419
PD-L1 Expression Subgroups
TC3 or IC3 (24 vs 23)	41.7%	19.9%	0.1003	37.5%	14.9%	0.0724
TC2/3 or IC2/3 (50 vs 55)	36.1%	13.8%	0.0082	21.2%	9.9%	0.1166
TC1/2/3 or IC1/2/3 (93 vs 102)	36.0%	19.8%	0.0124	18.0%	11.0%	0.1759
TC0 and IC0 (51 vs 41)	25.0%	6.8%	0.0202	20.5%	6.8%	0.0693
Histology Subgroups
Non-squamous (95 vs 95)	32.2%	21.1%	0.0960	23.3%	12.4%	0.0585
Squamous (49 vs 48)	32.7%	7.8%	0.0020	9.4%	5.2%	0.4603
[a ]For descriptive purpose only. TC3 or IC3 = PD-L1 ≥ 50% TC or 10% IC; TC2/3 or IC2/3 = PD-L1 ≥ 5% TC or IC; TC1/2/3 or IC1/2/3 = PD-L1 ≥ 1% on TC or IC; TC0 and IC0 = PD-L1 < 1% on TC and IC. NCT01903993.

Abstract not provided

Background:
Bavituximab targets exposed phosphatidylserine (PS) in the tumor microenvironment, resulting in repolarization of myeloid suppressor cells/M2 macrophages to M1, production of pro-inflammatory cytokines such as IFNγ and IL-12, dendritic cell maturation, and tumor specific cytotoxic T-cell activation. SUNRISE was a Phase III trial of docetaxel with bavituximab (D+B) or placebo (D+P) in patients with treated Stage IIIb/IV non-squamous NSCLC. Recent correlative analyses from SUNRISE suggest bavituximab is more active in PD-L1 negative, immune cold tumors and thus may complement PD-1/PD-L1 ICI.

Method:
This subgroup analysis included all patients who received subsequent ICI after discontinuing SUNRISE study drug. We calculated overall survival (OS) both from randomization and start of subsequent ICI.

Result:
Ninety-three of 597 randomized patients (16%) received ICI as next line of therapy after SUNRISE assigned treatment. Baseline characteristics were balanced between the treatment groups and consistent with the ITT population. From randomization, mOS was not reached (95% confidence interval [CI], 15.2-NA) in D+B (N=46) and 12.6 months (95% CI, 10.4-17.8) in D+P (N=47) (hazard ratio [HR], 0.46; P=0.006) (Figure). From start of ICI, mOS was not reached (95% CI, 10.2-NA) in D+B and 6.2 months (95% CI, 3.9-8.7) in D+P (HR, 0.42; P=0.002). The mPFS was 6.0 months (95% CI, 3.5-6.5) in D+B and 4.4 months (95% CI, 2.6-6.3) in D+P (HR, 1.00; P=0.991). ORR was 20% vs. 13% (Odds ratio 0.6; P=0.41) for D+B and D+P, respectively. The safety profile was similar between groups and no immune related (IR) toxicities (colitis, pneumonitis, hypothyroidism) were reported.

Conclusion:
Within the limits of a subgroup analysis, a significant improvement in OS was observed for patients previously treated with D+B. Furthermore, bavituximab has not been associated with IR toxicities and might serve as a useful drug in combination with ICI for the treatment of immune cold tumors. Figure 1

Background:
Nivolumab monotherapy has shown survival benefit in patients (pts) with different tumors, including melanoma, lung cancer, renal cell carcinoma, head and neck cancer and Hodgkin lymphoma. Controlled clinical trial setting differs from what experienced by pts and physicians in routine clinical practice. Here, we report efficacy and safety results of nivolumab in pts with non-squamous non-small cell lung cancer (Non-Sq-NSCLC) treated in the Expanded Access Programme in Italy.

Method:
Nivolumab was available upon physician request for pts aged ≥18 years who had relapsed after a minimum of one prior systemic treatment for stage IIIB/stage IV non-Squamous NSCLC. Nivolumab 3 mg/kg was administered intravenously every 2 weeks to a maximum of 24 months. Pts included in the analysis had received at least one dose of nivolumab and were monitored for adverse events (AE) using Common Terminology Criteria for Adverse Events.

Result:
Overall, 1588 pts were enrolled in the EAP across 168 Italian centers. Baseline characteristics of pts were representative of Non-Sq-NSCLC population, in the advanced disease setting. As of March 2017, median overall survival (OS) was 11 months (10.0-12.0), with a median follow-up of 7.8 months (1-21.9) and a median of 7 doses (1-46). The best overall response rate (BORR) was 18%, including 10 pts (<1%) with complete response and 280 pts (17.6%) with partial response. Stable disease has been defined for 414 pts (26.0%) and totally 274 (17.2%) patients were treated beyond progression. Response rates and survival were comparable among pts regardless age (< and ≥ 75 years), presence of brain metastasis and number of prior therapies. Overall, among 1588 pts, 1254 discontinued treatment for any reason, with only 93 (7%) pts who discontinued treatment due to adverse events, in line with what observed in previous studies.

Conclusion:
To date, this is the largest clinical experience with nivolumab in a real-world setting and these EAP data are in line with what reported in the registrative phase 3 clinical trial. According to these results, nivolumab seems to be an effective and safe therapy for pre-treated patients with non-squamous NCSLC, supporting its use in current clinical practice.

Background:
Immunotherapy is revolutionising the way many cancer types are treated. The immunosenescence and the high heterogeneity of the elderly raises questions on the benefits with such treatments and real-world data is lacking.

Method:
ELDERS is a prospective, observational pilot study on the use of checkpoint inhibitors in patients with advanced/metastatic non-small cell lung cancer (NSCLC) or malignant melanoma. The study was designed with 2 arms, the elderly (≥ 70 years) and the non-elderly (45-69 years) with 2 co-primary endpoints, immune-related toxicity (irAE) profile and health-related quality of life (HRQoL) through the EORTC QLQ-C30. A comorbidity score (CIRS) was applied at baseline (score 0-56) and serial geriatric assessments were performed for stratification with a screening tool (G8) and further geriatric assessment as needed. A total of 110 patients of a planned 120 have been recruited. This interim analysis is of the NSCLC cohort with a minimum of 3 months on study/follow-up.

Result:
32 patients were included, with 96% treated with pembrolizumab (9% first-line) and 40.6% enrolled on the elderly arm. In both arms, 45% had a tumour PD-L1 expression of ≥50%. The elderly arm had more advanced disease with 69% staged M1b vs. 42.1% in younger arm (p=0.05). 69% of patients, in both arms, were performance status 0/1 at the start of treatment. The median CIRS total score was 12 for the elderly and 7 for the younger arm. 46% of elderly patients had an abnormal geriatric screening (G8≤14), requiring further assessments. With a median follow-up of 6 months, the objective response rate (ORR) was overall 19% with a median time to response of 8 weeks. The ORR was numerically higher in the elderly with 30.8% vs. 10.5% (p=0.09). 9.4% of patients on study had a grade 3/4 irAE, with no difference between study arms. Elderly patients had a numerically higher rate of admissions, 53.8% vs. 36.8% (p=0.18). No statistically significant correlation was identified between higher comorbidity score or abnormal geriatric assessment and the incidence of irAEs. No significant negative impact on the global HRQoL was detected in either arm during treatment with immunotherapy.

Conclusion:
Despite the small number of patients and the limited follow-up time, there is no signal in this interim analysis to indicate that elderly patients have less benefit or higher risk of irAE compared with younger patients, despite more comorbidities and geriatric syndromes. These results help to inform clinical practice in the absence of trials dedicated to the elderly population.

Abstract not provided

Background:
T follicular helper cells (Tfh) are an antigen-experienced CD4+ T cell subset that have been found to play crucial roles in the development of humoral immunity. In particular, their presence in the B cell-rich germinal centre of secondary and tertiary lymphoid tissue aids in B cell maturation through selection of B cells producing high-affinity antibodies. Tfh cells have known roles in autoimmune disease and B cell malignancies; however, their role in many solid tumours, including those of the lung, remains unclear.

Method:
We analyzed 83 paired tumour-normal lung adenocarcinoma samples from the BC Cancer Agency (BCCA) as well as 576 unpaired samples from The Cancer Genome Atlas (TCGA). Relative immune cell content was obtained from gene expression data using a linear support vector regression deconvolution approach (CIBERSORT). Spatial positioning of B and T cells within selected tumour sections was examined through IHC. The impact of Tfh infiltration on patient survival was analyzed using a Cox Proportional Hazard model.

Result:
T follicular helper cells are increased in tumour tissue, accompanied by global upregulation of Tfh markers PDL1 and CXCR5 in both the BCCA and TCGA cohorts. Histological analysis revealed localization of Tfh cells within tertiary lymphoid organs, with direct contact with B cells in the follicular zone observed. Importantly, Tfh recruitment appears to be an early event in tumour development and a function of neoantigen exposure, indicative of an active anti-tumour response rather than a result of chronic inflammation of the tumour microenvironment.

Conclusion:
T follicular helper cells are required for B cell maturation and subsequent antibody responses. As such, it is not surprising that Tfh infiltration in tumour-resident ectopic lymphoid structures correlates with patient survival in various cancer types. Given the importance of tumour-specific antibody responses in natural and therapeutic immunity, further investigation of Tfhs may show prognostic utility and be a marker of early-stage lung tumours.​

Background:
There is growing appreciation for the role of tumor-draining lymph nodes (TDLN) in the dynamic of immunoediting orchestrated by non-small cell lung cancers (NSCLC). By comparing T-cell subsets and gene expression in TDLN and non-draining lymph nodes (NDLN), we aim to determine whether there is tumor-regional variation in immunophenotype.

Method:
Patients undergoing endobronchial ultrasound-guided transbronchial needle aspiration for the diagnosis/staging of NSCLC were recruited. Aspirates were obtained from TDLN (N1/N2 nodes with increased fluorodeoxyglucose-F-18 (FDG) avidity and/or enlarged >1cm) and NDLN (non-enlarged/non-FDG-avid N2/N3 nodes) along with peripheral blood. Samples were stained with fluorophore-conjugated antibodies (CD4-FITC, CD8-V450, CD25-PECy7, CD127-APCR700, CD45RO-PECF594) and analyzed by flow cytometry. CD4+CD25- and CD8+ effector T-cells (Teff) were sorted. Gene expression profiling was performed on sorted Teff using the Nanostring™ platform to measure differential expression between TDLN and NDLNs.

Result:
We compared T-cell subpopulations in TDLN and paired NDLN from 16 subjects. There were significantly fewer CD4+ T-cells in TDLN vs NDLN (10.1% vs 28.9%, p=0.0039), with more Tregs (12.1% vs 7.3%, p=0.1563) suggesting a pattern of tumor-regional immunosuppression in the TDLN. This was more consistent when tumor histology was adenocarcinoma compared to squamous cell cancer with respect to both depletion of Teff and higher proportion of Tregs (Figure 1). A more immunosuppressive TDLN phenotype was also observed with high tumor PD-L1 expression (>50%), with 36% fewer CD4+ T-cells in TDLN relative to paired NDLN when PD-L1 expression was high relative to just 3.2% fewer CD4+ T-cells with low PD-L1 expression. Gene expression in Teff has preliminarily demonstrated upregulation of genes mediating T-cell exhaustion (CTLA-4, PD-1, TGF-β) and downregulation of co-stimulatory/recruitment factors (CD28, ICOS, ICAM2) in TDLN suggesting impaired activation of tumor-regional Teff. Figure 1



Conclusion:
Our findings suggest that TDLNs in patients with NSCLC display a tolerogenic phenotype, with more marked immunosuppression in the setting of adenocarcinoma and high tumor PD-L1 expression.

Background:
Using Tumor Growth Rate (TGR), HPD was identified in 9% of 131 advanced cancer pts, treated with IO in a single institution (Champiat et al. 2017). In this study, we explored HPD in a large, multicenter cohort of advanced NSCLC pts treated with IO.

Method:
We performed a retrospective analysis of consecutive NSCLC pts treated with IO, in 8 institutions, between November 2012 and April 2017. Eligibility criteria required, for each patient: 2 CT scans performed before starting IO and one during IO, an interval between two CT scans ≥2 weeks or 3 months (m) and presence of target lesions. CT scans were centrally assessed according to RECIST 1.1 criteria. We calculated TGR before IO (TGR pre-IO) and during IO (TGR IO); patients were defined HPD if they had progression disease (PD) at first evaluation during IO and a ≥ 2-fold increase in the TGR IO compared to TGR pre-IO. Median overall survival (mOS) was estimated using Kaplan-Meier method for the total population and HPD pts.

Result:
Among 419 eligible pts, 86 were excluded for inadequate intervals between CT scans. Among 333 evaluable pts, 63% were male, 46% ≥65 years, 43% smokers; 12% had PS ≥ 2, 65% adenocarcinoma, 45% ≥3 metastatic sites, 22% KRAS mutation, 4% EGFR mutation, 1% ALK rearrangement; 21% had PD-L1 positive status, 10% negative, 69% unknown, >90% received single agent PD-1 inhibitor in ≥ 2 line. Response rate (RR) to IO was 18%, median follow up was 12 m [10-14]. 33% of pts had TGR IO ≥1 (not regressing tumors), 25% had TGR IO ≥ 2-fold TGR pre-IO and 54 pts (16%) had HPD. 15 pts (4%) had confirmed pseudoprogression, 3 were initially qualified as HPD. Compared to not-HPD, HPD pts had more frequently ≥ 3 metastatic sites at baseline (59% vs 43% p=0,02) and more new lung lesions during IO (34% vs 17% p=0,007). PD-L1 negative status was more common among HPD pts but the association was borderline significant (53% vs 28% p=0,05). Age, clinical, molecular characteristics, RR to treatment before IO, baseline tumor burden, liver or brain new lesions during IO were not different according to HPD status. mOS was 13 m [10-17] in the total population, 5 m [3-8] in HPD pts.

Conclusion:
HPD occurred in 16% of advanced NSCLC pts treated with IO and was associated with plurimetastatic disease and appearance of new lung lesions. Further work is needed to characterize HPD prognostic value.

Abstract not provided

Abstract not provided

Abstract:
Antigen-specific immunotherapy or cancer vaccination has been studied in several large phase III trials in NSCLC in different stages [1]. Agents consisted of two major components: immunogenic tumor-associated antigens, combined with a strong adjuvant to generate the tumor directed attack. Overall, the results of these trials have been disappointing. One of these phase III trials was MAGRIT, the largest therapeutic clinical trial ever performed in NSCLC. The MAGRIT trial (MAGE-A3 as Adjuvant Non-Small Cell LunG CanceR ImmunoTherapy[2]. Other major trials in this setting, such as RADIANT, studying adjuvant erlotinib [3], and ECOG 1505, studying the addition of bevacizumab cisplatin-based adjuvant chemotherapy [4], proved to be negative. As adjuvant cisplatin-based chemotherapy can be hard to tolerate for many patients [5], a less toxic and effective therapy to improve the outcome in this group was of great interest. Melanoma associated antigen (MAGE)-A3 was an interesting target because it is almost exclusively expressed on tumor cells, and not expressed in normal tissue (except in male germline cells, which however do not present the antigen). The MAGE-A3 vaccine was a recombinant protein antigen-based vaccine, containing the recombinant fusion protein (MAGE-A3 and protein D of Haemophilus Influenzae) in combination with an immune response enhancing adjuvant. Clear responses to this compound had been noted in early experience in patients with metastatic melanoma [6]. For NSCLC, the proof of concept study was a double-blind, placebo-controlled, randomized phase II trial [7]. Patients with completely resected MAGE-A3-positive stage IB-II NSCLC were randomly assigned to either MAGE-A3 vaccine (N=122) or placebo (N=60), 5 administrations q3 weeks followed by 8 administrations q3 months. No adjuvant chemotherapy was given, as this therapy was not established in the study interval. Disease-free interval (DFI) was the primary endpoint. After a median post-resection period of 70 months, there was a trend in favor of MAGE-A3, with a Hazard Ratio (HR) for DFI 0.78 (95%CI 0.49-1.24; two-sided P=0.295). No significant toxicity was observed, resulting in very high therapy compliance. Furthermore, a possible gene signature (GS), predictive of clinical activity of the MAGE-A3 vaccine in previous metastatic melanoma experience [6], could be validated in early-stage NSCLC [8]. Actively treated GS-positive NSCLC patients showed a favorable DFI compared with placebo-treated (HR 0.42, 95% CI 0.17-1.03; P=0.06), whereas among GS-negative patients, no such difference was found (HR 1.17, 95% CI 0.59-2.31; P=0.65). This led to the large double-blind, randomized, placebo-controlled phase III trial MAGRIT (ClinicalTrials.gov, number NCT00480025) [9]. MAGE-A3 positive patients with completely resected stage IB, II or IIIA NSCLC, and adjuvant chemotherapy as clinically indicated, were 2:1 randomly assigned to the MAGE-A3 vaccine or placebo. Randomization and treatment allocation was done centrally via internet with stratification for chemotherapy versus no chemotherapy. A minimization algorithm accounted for the number of chemotherapy cycles received, disease stage, lymph node sampling procedure, performance status score, and lifetime smoking status. The primary endpoint was broken up into three co-primary objectives: disease-free survival in the overall population, the no-chemotherapy population, and patients with a potentially predictive gene signature. Between Oct 18, 2007, and July 17, 2012, a total of 13,849 surgical patients in 443 centers in 34 countries were screened for MAGE-A3 expression, 4210 had MAGE-A3 expression, and 2,272 were treated (active vaccine 1,515; placebo 757). 784 patients in the MAGE-A3 group also received chemotherapy, as did 392 in the placebo group. At the time of the report, median follow-up was 38.1 months in the MAGE-A3 group and 39.5 months in the placebo group. In the overall population, median disease-free survival (DFS) was 60.5 months (95% CI 57.2–not reached) in the MAGE-A3 vaccine group and 57.9 months (55.7–not reached) in the placebo group (hazard ratio 1.02, 95% CI 0.89–1.18; P=0·74). In the predefined subgroup patients who did not receive chemotherapy, median DFS was 58.0 months (95% CI 56·6–not reached) in the MAGE-A3 group and 56.9 months (44.4–not reached) in the placebo group (hazard ratio 0.97, 95% CI 0.80–1.18; p=0·76). Because of the absence of treatment effect, the predictive gene signature could not be further studied. The frequency of grade 3 or worse adverse events was similar: 246/1515 (16%) in the MAGE-A3 group and 122/757 (16%) in the placebo group. It was concluded that adjuvant treatment with the MAGE-A3 vaccine did not increase DFS compared with placebo in patients with MAGE-A3-positive surgically resected NSCLC, and the further development of the MAGE-A3 vaccine for NSCLC was stopped. REFRERENCES 1. Decoster L, Wauters I, Vansteenkiste J. Vaccination therapy for non-small cell lung cancer: Review of agents in phase III development. Ann Oncol 2012; 23: 1387-1393. 2. The International Adjuvant Lung Cancer Trial Collaborative Group, Arriagada R, Bergman B et al. Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small cell lung cancer. N Engl J Med 2004; 350: 351-360. 3. Kelly K, Altorki NK, Eberhardt WE et al. Adjuvant erlotinib versus placebo in patients with stage IB-IIIA non-small cell lung cancer (RADIANT): A randomized, double-blind, phase III trial. J Clin Oncol 2015; 33: 4007-4014. 4. Wakelee HA, Dahlberg SE, Keller SM et al. Randomized phase III trial of adjuvant chemotherapy with or without bevacizumab in resected non-small cell lung cancer (NSCLC): Results of E1505. J Thorac Oncol 2015; 10 Suppl 2: 66S. 5. Alam N, Shepherd FA, Winton T et al. Compliance with post-operative adjuvant chemotherapy in non-small cell lung cancer. An analysis of National Cancer Institute of Canada and intergroup trial JBR.10 and a review of the literature. Lung Cancer 2005; 47: 385-394. 6. Kruit WH, Suciu S, Dreno B et al. Selection of immunostimulant AS15 for active immunization with MAGE-A3 protein: Results of a randomized phase II study of the European Organisation for Research and Treatment of Cancer melanoma group in metastatic melanoma. J Clin Oncol 2013; 31: 2413-2420. 7. Vansteenkiste J, Zielinski M, Linder A et al. Adjuvant MAGE-A3 immunotherapy in resected non-small cell lung cancer: Phase II randomized study results. J Clin Oncol 2013; 31: 2396-2403. 8. Ulloa-Montoya F, Louahed J, Dizier B et al. Predictive gene signature in MAGEA3 antigen-specific cancer immunotherapy. J Clin Oncol 2013; 31: 2388-2395. 9. Vansteenkiste JF, Cho BC, Vanakesa T et al. Efficacy of the MAGE-A3 cancer immunotherapeutic as adjuvant therapy in patients with resected MAGE-A3-positive non-small-cell lung cancer (MAGRIT): A randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 2016; 17: 822-835.

Abstract:

Abstract:
A well-known mechanism of resistance to EGFR-TKIs is MET upregulation. MET inhibitors were developed to overcome and prevent this resistance mechanism. Onartuzumab is a monoclonal antibody that binds the extracellular domain of MET. By blocking the interaction with its HGF ligand, activation of the MET signaling pathway does not occur and tumor growth is halted. The clinical evaluation of onartuzumab followed the traditional phase I, II and III registration pathway. In the randomized Phase II trial of erlotinib and placebo versus erlotinib and onartuzumab the trial failed to meet the co-primary endpoint of PFS in the intent to treat population but was positive for the co-primary endpoint of PFS in 66 patients with MET positive tumors defined as IHC > 2+ expression (HR, .53; P .04) (1). Overall survival was also significant (HR, .37; P .002). Objective response rate (ORR) was reported as 8.6% and 3.2% for onartuzumab versus placebo, respectively. Based on these results a randomized placebo controlled phase III trial was launched in patients with MET expressing tumors (2). Surprisingly this trial did not meet its OS primary endpoint and numerically favored the placebo arm. A total of 499 patients were enrolled. The median OS was 6.8 months for onartuzumab versus 9.1 months for placebo (HR, 1.27; 95% CI, 0.98 to 1.65; P = .067). Median progression-free survival was 2.7 months versus 2.6 months (HR, 0.99; 95% CI, 0.81 to 1.20; P = .92) and the ORR was 8.4% compared with 9.6% respectively. When a trial is negative scrutinizing all aspects of the trial and its predecessor trial to determine if there were instructive signals is needed. In this case, reported patient characteristics were similar between the two trials but other patient variables such as the frequency of patients with brain metastases, sites of metastases, and time from previous therapy were not provided. However a large magnitude of difference would be required to significantly alter the results which is unlikely. Adverse events profiles were unrevealing. There were differences related to the MET biomarker that may have influenced the phase III results. In the phase II trial MET IHC expression was retrospectively determined compared to its prospective determination in the phase III trial and its use as a stratification factor. Although the frequency of MET 2+ versus 3+ IHC expression was similar in the two trials, the retrospective nature of the analysis in the phase II trial with its inherent imbalance in patient characteristics may have been misleading especially in the context of the small sample size. Had this been a randomized biomarker selected Phase II trial with a larger sample size we might have seen a different outcome. The assay itself was not a factor. Rigorous validation of the MET IHC assay was conducted. The assay was performed at Genentech for the phase II study and these investigators carefully trained the central laboratories personnel performing the assay for the phase III trial. A robust quality check and audit program was followed. A frequently asked question is whether IHC accurately characterize drivers of MET dysregulation that would result in EGFR tyrosine kinase inhibitor (TKI) resistance and onartuzumab responsiveness. In an exploratory biomarker analysis from the Phase II study MET IHC remained the most robust predictor of efficacy for the combination (3). In the phase III trial a detailed EGFR and MET pathway analysis that included MET and EGFR FISH, EGFR amplification and EGFR, KRAS BRAF, PIK3CA mutational analysis failed to find a biological explanation for onartuzumab inactivity (3). An analysis of MET Exon 14 splicing mutations was not conducted because these mutations had not been discovered at the time of study conduct. Although it would be intriguing to know the frequency of these mutations and their association to onartuzumab activity this is unlikely to be performed. The subgroup results supported further investigation of onartuzumab in a MET positive population but the results were modest. The statistically significant 1.4 month improvement in median PFS is not clinically significant and objective response rates were similar between the arms. Driving the phase III design was the impressive 8.8 month improvement in median OS for the combination but without strong efficacy signals in ORR and PFS to account for this survival outcome suggests other factors were at play such as subsequent therapies and warrants caution. The dilemma with encouraging preliminary data is what is the optimal next study design especially in this instance where the findings were modest? Too many times we have seen positive phase II trials lead to negative phase III results. While a phase III trial is the quickest route to a definitive answer it is done at the price of hundreds of patients. This is particularly highlighted by this study where the combination was approaching an inferior overall survival. Alternative design strategies such as a randomized Phase II/III design that can better balance benefit and risk for our patients and still achieve the goal should be considered. Criteria should be established to help investigators select the appropriate design. References 1. Spigel DR, Ervin TJ, Ramlau RA, Daniel DB, Goldschmidt JH Jr, Blumenschein GR Jr, Krzakowski MJ, Robinet G, Godbert B, Barlesi F, Govindan R, Patel T, Orlov SV, Wertheim MS, Yu W, Zha J, Yauch RL, Patel PH, Phan SC, Peterson AC. Randomized phase II trial of Onartuzumab in combination with erlotinib in patients with advanced non-small-cell lung cancer. J Clin Oncol. 2013, 31:4105-14. 2. Spigel DR, Edelman MJ, O'Byrne K, Paz-Ares L, Mocci S, Phan S, Shames DS, Smith D, Yu W, Paton VE, Mok T. Results From the Phase III Randomized Trial of Onartuzumab Plus Erlotinib Versus Erlotinib in Previously Treated Stage IIIB or IV Non-Small-Cell Lung Cancer: METLung. J Clin Oncol. 2017, 35:412-20. 3. Koeppen H, Yu W, Zha J, Pandita A, Penuel E, Rangell L, Raja R, Mohan S, Patel R, Desai R, Fu L, Do A, Parab V, Xia X, Januario T, Louie SG, Filvaroff E, Shames DS, Wistuba I, Lipkind M, Huang J, Lazarov M, Ramakrishnan V, Amler L, Phan SC, Patel P, Peterson A, Yauch RL. Biomarker analyses from a placebo-controlled phase II study evaluating erlotinib±onartuzumab in advanced non-small cell lung cancer: MET expression levels are predictive of patient benefit. Clin Cancer Res. 2014, 20:4488-98.

Abstract not provided

Abstract:
In this talk I will cover reasons why phase III trials are "negative" and what can be done to increase the rate of true positive results. Reasons for failure in phase III include inactive agents, poor statistical design, and inadequate implementation. With targeted agents including immunotherapy, additional issues of statistical design arise, as well as assay reliability. I will review the trials discussed in this section (PROCLAIM, MAGRIT, METLung, CheckMate026) with a view towards assessing the reasons these trials might have been negative, and make some suggestions for future trials.

Abstract:
T4 lung cancers invading neighboring structures comprise a heterogenous group of locally invasive tumors. In a small subset of these localized tumors whose extrapulmonary invasion preceded any lymphatic extension, an adequate excisional procedure can achieve surprising long term survivals. The indications for such procedures and the anticipated outcomes should be weighed on a case by case basis, in terms of potential perioperative complications and expertise of the surgical team. Advanced surgical techniques are now being applied for T4 lesions due to improvements in surgery and anesthesiology and progress in combined treatment modalities. In the present staging, T4 tumors without mediastinal nodal metastasis are now considered to be potentially curable if complete resection is possible. A summary of the literature, under the light of personal experience allows a critical point of view, knowing that a surgical procedure which would not be reproducible in other centers would never be recognized as an option for practice. Therefore among the published series it is important to distinguish the real progress given by innovative techniques or procedures that could be applied throughout the world, even only in selected centers, from the simple reports of individual performances or exploits. Proximal tumors from the lower lobe may involve the atrial wall of the heart. In some cases a left atrial resection can be performed, followed by direct closure, or replacement of the atrial wall. Fukuse reported a series of 42 patients, from which left atrium was resected in 14 patients, Mortality rate was low (2.4%), regarding complexity of the procedures [1]. Low stages in nodal status were associated with increased survival (p = 0.0013). More recently was reported a series of 19 patients who underwent extended lung resection involving the left atrium without cardiopulmonary bypass [2]. An interatrial septum dissection is performed, thus increasing the length of the atrial cuff. R0 resections were observed in 89% of the patients. Ninety-day mortality rate was 16%. Five-year survival rate is 44%, and 3 patients (16%) are alive more than 6 years after surgery. Other reports advocate the use of cardiopulmonary bypass in these occasional situations [3-5]. Invasion of the superior vena cava (SVC) by a T4 non small cell lung c ancer (NSCLC) led surgical teams to attempt lobectomies or pneumonectomies extended to the vena cava [6]. Actually direct extension to the vessel by the tumor mass itself is a rare situation. SVC involvement generally results from a bulky disease, in which the nodal disease is the greatest component. The rationale of resecting SVC in N2 disease remains questionable, in view of the high potential of metastatic spread and the poor prognosis. Nevertheless, different techniques were proposed, including lateral clamping of SVC, partial, or total reconstruction. These procedures are associated with high morbidity rates. A multicentric international review of prosthetic replacement after SVC resection for nsclc in 28 patients (N2 involvement in 50%) showed morbidity and mortality rates of 39% and 14%, respectively. Overall 5-year survival rate was 15% [7]. Despite some reports who claim better survival rates, close to 50% at five years, the latter seems more realistic, and this warrants a thorough evaluation with the aim to preclude these patients from surgery in case of N2 involvement. A bronchial carcinoma extended to the tracheal bifurcation can be resected in selected patients [8]. A high rate of post-operative morbidity (10 to 30%), including bronchial dehiscences, jeopardizes the outcome, but long-term survivals have been observed in 15 to 23% of the cases. A meticulous mediastinal assessment is mandatory to eliminate invasion of the airway by a bulky disease. Only patients with T-invasion will be offered surgical resection. NNSCLC invading the thoracic inlet can easily penetrate spinal structures because of their particular anatomic situation. The best local control for resectable tumors is achieved by surgical operation, provided the resection is complete and respecting oncologic principles. Direct invasion of the vertebral body became an option following the first report in 1996 of a successful en bloc total vertebrectomy for lung cancer invading the spine [9]. Reported experiences from Europe, North-America and Asia demonstrate feasability and encouraging results of these challenging procedures, . Recently a comprehensive literature search, on a total of 1,001 abstracts and 93 articles found overall 5-year survival rates ranging from 37% to 59% and the mortality rate ranged from 0% to 6.9% [10]. Undoubtly enbloc resection for lung cancer invading the spine is reaching the stage of current practice in expert centers. This is probably due to a particular biology of these tumors which are peripheral and whose noisy symptoms lead to a relatively early diagnosis, thus permitting a high rate of complete resections. Evidence suggests that triple modality therapy with complete resection of locally advanced Pancoast tumors with involvement of the spine offers an advantage over other therapeutic modalities. Despite the absence of such an evidence in other T4 lung cancers, recent advances in patient's care and surgical techniques allowed surgeons to become more aggressive, and to propose occasionally extended resections with encouraging long-term survival rates to patients suffering from tumors invading the tracheal bifurcation, the left atrium, or the great vessels. The 8[th] edition of TNM classifies,T4N0-1 tumors in a "surgical" category, stage IIIA. 1. Fukuse T, et al. Extended operation for nsclc invading great vessels and left atrium. Eur J Cardiothorac Surg 1997;11:664–9 2. Galvaing G, et al. Left atrial resection for T4 lung cancer without cardiopulmonary bypass: technical aspects and outcomes.Ann Thorac Surg 2014;97:1708-13 3. Klepetko W, et al. T4 lung tumors with infiltration of the thoracic aorta: is an operation reasonable? Ann Thorac Surg 1999;67:340–4 4. De Perrot M, et al. Resection of locally advanced (T4) nsclc with cardiopulmonary bypass. Ann Thorac Surg 2005; 79:1691–6 5. Langer NB, et al. Outcomes after resection of T4 nsclc using cardiopulmonary bypass. Ann Thorac Surg 2016;102:902-10 6. Grunenwald DH. Resection of lung carcinomas invading the mediastinum, including the superior vena cava. Thorac Surg Clin 2004;14:255–63 7. Spaggiari L, et al. Superior vena cava resection with prosthetic replacement for nsclc: long term results of a multicentric study. Eur J Cardiothorac Surg 2002;21:1080–6 8. Mathisen DJ, Grillo HC. Carinal resection for bronchogenic carcinoma. J Thorac Cardiovasc Surg 1991;102:16−23 9. Grunenwald D, et al. Total vertebrectomy for en bloc resection of lung cancer invading the spine. Ann Thorac Surg 1996;61:723–6 10. Setzer M, et al. Management of locally advanced pancoast (superior sulcus) tumors with spine involvement. Cancer Control 2014;21:158-67

Abstract not provided

Abstract:
Genotyping of plasma cell-free DNA (cfDNA) is rapidly changing our management approach for genotype-defined lung cancers. Widely available assays now have the ability to noninvasively identify driver mutations in cfDNA, monitor response to therapy, and characterize emerging resistance. However, such liquid biopsies also have clear limitations – existing assays are unlikely to replace tumor biopsies completely. This presentation will discuss an optimal approach for integrating liquid biopsies into lung cancer care, while also envisioning how liquid biopsies may become a routine part of lung cancer treatment in the years ahead.

Abstract:
Any Differences in the Management of Elderly Patients with Lung Cancer between East and West? Lung cancer is one of the commoner cancers in the world, accounting for 1.6 million cases annually. In many western countries, lung cancer rates have declined in contrast to the predicted increase in the incidence of lung cancer in Asia, especially among males. Many patients with lung cancer are older and this applies to all parts of the world. There are relatively few data for the older age group, since many studies have an age restriction, such that there can be uncertainty in the extrapolation of trial findings to the average older person. On the other hand it is clear that there are differences in lung cancer in the East and the West including genetic and biological differences such as smoking habits between genders, genetic aberrations such as the frequency of EGFR gene mutations, in addition to cultural and local preferences. Moreover clinical trials are often performed either in the East or the West, as are studies of lung cancer in the older population, such that extrapolation may be needed. In the work up of suspected lung cancer, the older patient may be less tolerant of diagnostic tests given the ageing process and potential concurrent co-morbid disease, especially smoking related diseases which can differ between East and West. In addition in terms of treatment apart from well-known somatic mutation differences, ethnic differences in efficacy and toxicity from therapies may differ, in addition to non-patient factors such as affordability, cultural sensitivities and preferences. Thus the challenge in interpreting and applying research and clinical trials data is includes the relative paucity of data on the older population as well as uncertainty whether the data that is available on the older population applies equally to those in the East and the West, even allowing for known biological differences. This presentation will examine knowledge gaps and differences that should be considered for the management of the older person with the lung cancer from the East and West, in the setting of cultural and genetic uniqueness.

Abstract:
Chemotherapy Treating the elderly with advanced NSCLC remains a major challenge. ELVIS (Elderly Lung Cancer Vinorelbine Italian Study) was the first elderly-specific phase III randomized trial to be conducted in advanced NSCLC; this study demonstrated improved survival and quality of life in patients 70 years of age and older, receiving weekly vinorelbine compared to best supportive care. The most studied non platinum-based regimen has been the combination of gemcitabine plus vinorelbine, which proved active and well tolerated in several phase II trials. However, in a large phase III randomized Multicenter Italian Lung cancer in the Elderly Study (MILES), which enrolled about 700 patients, combination chemotherapy with gemcitabine and vinorelbine failed to improve outcomes (response rate, time to progression, survival or quality of life) compared to single agent chemotherapy with either vinorelbine or gemcitabine. Based on these trials, single agent chemotherapy, until relatively recently, was considered the standard of care in PS 0-2 elderly advanced NSCLC patients. However, over the past two decades, retrospective analyses of randomized phase III trials and multiple phase II studies have confirmed the activity and tolerability of cisplatin-based and carboplatin-based chemotherapy in fit elderly patients. In aggregate, these trials suggested that the elderly fared as well or nearly as well as their younger counterparts with respect to response rate and survival, albeit with more toxicity. A randomized phase III trial led by Quoix et al in advanced NSCLC patients between the ages of 70 and 90 with PS 0-2 showed superiority in all outcome parameters for carboplatin and weekly paclitaxel versus either single agent gemcitabine or vinorelbine. Response rate was nearly triple at 29% vs 11% (p <0.0001) with progression-free survival of 6.1 mos vs 3.0 mos (p < 0.0001), median survival of 10.3 mos vs 6.2 mos (p=0.00004), and 1-year overall survival rate of 45% vs 26% (p < 0.001). Quality of life was preserved or improved with the combination regimen. Survival benefit was seen in those 80 years of age and older (n = 114; HR 0.56 [95% CI: 0.37 -0.85], p = 0.0067), and in individuals with PS 2 (n = 122; HR 0.65 [95% CI: 0.44 – 0.95] p =0.027). More recently, in a retrospective subset analysis of elderly individuals enrolled on a phase III study of carboplatin plus either conventional solvent based (sb) paclitaxel administered every three weeks or nanoparticle, albumin-bound (nab) paclitaxel given weekly, the latter regimen resulted in a significant improvement in overall survival at 19.9 mos vs 10.4 mos (p = 0.009). To date, there has been no formal, prospective, phase III comparison of weekly sb- vs weekly nab- paclitaxel in combination with carboplatin in advanced NSCLC in the elderly or any other population. At ASCO 2017, ABOUND 70+ showed that therapy with carboplatin and weekly nab-paclitaxel x 3, followed by a one week break (Arm B) was well tolerated with no dilution in efficacy compared to the “uninterrupted” weekly nab-pacliaxel regimen (Arm A) used in the phase III trial: median PFS (Arms A and B) 3.9 vs 7.0 mo (HR 0.49; 95% CI, 0.30 - 0.79; P = 0.003), confirmed ORR 23.9% vs 40.3% (P = 0.037), and median OS 15.2 vs 16.2 mo (HR 0.76; 95% CI, 0.46 - 1.26; P= 0.292). In addition, the feasibility of cisplatin-based chemotherapy was also reported during ASCO 2017. The results from two separate phase III randomized trials (MILES 3 and 4) led by Gridelli et al comparing single agent gemcitabine to cisplatin+gemcitabine in squamous cell histology or single agent pemetrexed to cisplatin+pemetrexed in non-squamous cell NSCLC, showed a significant improvement in ORR (15.5% vs 8.5%, p=0.02) and PFS (4.6 mos vs 3.0 mos, p=0.005, HR 0.76) and suggested a non-significant trend toward improved median OS (9.6 vs 7.5 mos, p= 0.136, HR: 0.86) in patients receiving the platinum-based combinations. In clinical practice, non-platinum monotherapy remains the standard treatment for unfit elderly patients with advanced NSCLC, but a carboplatin-based combination is a reasonable option for those who are fit enough. Bevacizumab and targeted therapies: E4599 showed a survival advantage for combination bevacizumab and paclitaxel/carboplatin (PCB) vs chemotherapy alone, though a subsequent analysis of those over 70 years of age suggested that this benefit was diluted in older patients. More nuanced analyses by Wakelee and Ramalingam suggested that males of any age sustained a survival benefit, while women up to the age of 60 also realized an OS advantage. It was only in women > 60 years of age where this benefit seemed to be lost, in part because the control arm performed better in this group. A more recent joint analysis pooling the data from the experimental arm of E4599 and the “control” arm of POINT BREAK, both featuring combination PCB, and comparing these data to PC alone (E4599), showed that a survival benefit was sustained up to age 75 with median survival (MS) of 13.4 vs 10.2 mos (HR, 0.78; 95% CI, 0.68–0.89). However, above the age of 75 in a limited cohort (n=157), combination PCB offered no advantage: MS of 9.6 mos vs 13.0 mos for those < 75 yo (HR, 1.05; 95% CI, 0.70–1.57). As for tyrosine kinase inhibitors (TKIs), a retrospective analysis of the elderly enrolled on BR 21, which compared erlotinib to placebo in the 2nd and 3rd line setting in advanced NSCLC showed a consistent response (OR%), progression-free survival (PFS) and OS benefit, albeit a bit more toxicity in older subjects. More recently, in patients with actionable mutations or translocations, phase III trials have demonstrated a consistent OR% and PFS advantage for erlotinib, afatinib, and crizotinib compared to standard front-line or second line chemotherapy regardless of age. Sub-analyses using 65 to 75 as cut points consistently show similar benefits, although the individual comparisons for the elderly have been frequently underpowered to demonstrate statistical significance. In the EURTAC trial comparing erlotinib to platinum-based chemotherapy in patients with exon 19 or 21 mutations, the magnitude of PFS benefit for the TKI was very similar in those above and below 65 years of age: amongst 88 individuals > 65, the HR was 0.26 (95% CI: 0.16 – 0.51), while it was 0.49 for 85 subjects < 65 years of age (95% CI: 0.25- 0.75). In Lux Lung 3, which compared afatinib to cisplatin/pemetrexed in advanced, treatment-naïve patients with EGFR mutations, PFS favored afatinib over chemotherapy in 135 patients 65 years of age and older (HR 0.64, 95% CI: 0.39 – 1.03), not too different from the benefit seen in 211 patients under the age of 65 (HR 0.53, 95% CI 0.36- 0.79). More recently, the J-ALEX trial demonstrated a statistically significant and clinically meaningful PFS advantage for alectinib vs standard crizotinib in TKI-naïve ALK (+) NSCLC. Amongst 207 enrollees, 22 subjects were 75 years of age and older; the hazard ratio for alectinib’s PFS benefit in this very small population was impressive at 0.28, but because of small numbers and an under-powered comparison, the 95% confidence intervals overlapped unity (0.06 – 1.19), and so the putative difference was not statistically significant. Immunotherapy Novel immunotherapy with check-point inhibitors nivolumab, pembrolizumab and atezolizumab have yielded an overall survival benefit compared to standard docetaxel in second-line setting of advanced NSCLC with less toxicity. As of 2017, PD1/PDL1 inhibitors have effectively replaced chemotherapy in the second line setting independent of histology. In these trials, the benefits have been confirmed consistently in subgroup analysis of the elderly, particularly in those between 65 and 75 years of age. Representation of those above 75 years of age in these studies, unfortunately, has been relatively sparse. In some series, there is no indication of increased toxicity of nivolumab in older patients. To date, there are no ongoing elderly-specific trials evaluating immunotherapy in advanced NSCLC. Summary: Elderly NSCLC patients who are fit for clinical trials do reasonably well in comparison to their younger counterparts. However, extrapolating clinical trial data to the general population, is problematic. Clearly, more evidence is required, particularly among octogenarians and patients with multiple chronic conditions. Older patients are at risk for tolerating chemotherapy poorly because of comorbidity and organ dysfunction; in this regard, modified comprehensive geriatric assessments may help facilitate appropriate treatment selection.

Background:
Retrospective studies indicate that selecting individuals for low dose computed tomography (LDCT) lung cancer screening based on a highly predictive risk model is superior to applying National Lung Screening Trial (NLST)-like criteria, which use only categorized age, pack-year and smoking quit-time information. The Pan-Canadian Early Detection of Lung Cancer Study (PanCan Study) was designed to prospectively evaluate whether individuals at high risk for lung cancer could be identified for screening using a risk prediction model. This paper describes the study design and results.

Method:
2537 individuals were recruited through 8 centers across Canada based on a ≥2% of lung cancer risk estimated by the PanCan model, a precursor to the validated PLCOm2012 model. Individuals were screened at baseline and 1 and 4 years post-baseline.

Result:
At a median 5.5 years of follow-up, 164 individuals (6.5%) were diagnosed with 172 lung cancers. This was a significantly greater percentage of persons diagnosed with lung cancers than was observed in the NLST(4.0%)(p<0·001). Compared to 57% observed in the NLST, 77% of lung cancers in the PanCan Study were early stage (I or II) (p<0.001) and to 25% in a comparable population, age 50-75 during 2007-2009 in Ontario, Canada’s largest province, (p<0·001).

Conclusion:
Enrolling high-risk individuals into a LDCT screening study or program using a highly predictive risk model, is efficient in identifying individuals who will be diagnosed with lung cancer and is compatible with a strong stage shift – identifying a high proportion at early, potentially curable stage. Funding This study was funded by the Terry Fox Research Institute and Canadian Partnership Against Cancer. ClinicalTrials.gov number, NCT00751660

Background:
In lung cancer screening, a nodule management protocol describes nodule assessment and thresholds for nodule size and growth rate to identify patients who require immediate diagnostic evaluation or additional imaging exams. The NELSON and NLST clinical trials used different selection criteria and nodule management protocols. Several modelling studies have reported variations in screening outcomes and cost-effectiveness across selection criteria and screening intervals; however, the effect of variations in the nodule management protocol remains uncertain. This study evaluated the effects of the eligibility criteria and nodule management protocols on the benefits, harms, and cost-effectiveness of lung screening scenarios in a population-based setting in Germany.

Method:
We developed a modular microsimulation model: a biological module simulated individual histories of lung cancer development from carcinogenesis onset to death; a screening module simulated patient selection, screening-detection, nodule management protocols, diagnostic evaluation and screening outcomes. Benefits included mortality reduction, life years gained, averted lung cancer deaths. Harms were costs, false-positives, overdiagnosis. Comparator was no screening. Evaluated 57 screening scenarios included variations in selection criteria and thresholds for nodule size and growth rate.

Result:
Five years of annual screening resulted in an 11.3–12.6% lung cancer mortality reduction in the screened population. The efficient scenario included volumetric assessment, a threshold for a volume of 300 mm[3], and a threshold for a volume doubling time of 400 days. Assessment of volume doubling time is essential for reducing overdiagnosis and false-positives. Incremental cost-effectiveness ratios of the efficient scenarios were 19,389–23,804 Euro per life years gained and 178,673–285,630 Euro per averted lung cancer death.

Conclusion:
Lung cancer screening can be cost-effective in Germany. Along with the eligibility criteria, the nodule management protocol influences screening performance and cost-effectiveness. Definition of the thresholds for nodule size and nodule growth in the nodule management protocol should be considered in detail when defining optimal screening strategies.

Background:
Screening of high risk smokers with computed tomography (CT) aims to identify early stage lung cancers in screening participants amenable to curative surgery. The National Lung Screening Trial (NLST) demonstrated a 20% reduction in lung cancer specific mortality in the CT arm compared to chest x-ray (control) arm. European screening trial results to date have failed to show any evidence of a reduction in lung cancer mortality. Reduction in lung cancer mortality comes from the combined effects of successful surgical removal of life-threatening early stage lung cancers and post-operative survival. In screening participants of the NLST, who are older chronic smokers, there exists a balance between mortality from lung cancer and mortality from non-lung cancer related causes.

Method:
This study aimed to validate a gene-based risk tool for dying of lung cancer and examine the outcomes from screening according to tertiles of risk. It also aimed to establish the utility of adding SNP-based data to risk prediction and efficacy in identifying which screening participants get the best outcomes from screening. Using prospective data from the NLST-ACRIN cohort (N=10,054), we examined the utility of combining risk genotypes with clinical risk variables in our risk model for dying of lung cancer. We then stratified screening participants into risk tertiles according to our risk model and compared the outcomes from CT versus CXR screening

Result:
The addition of risk genotypes (combined genetic risk score) to our clinical risk model for dying of lung cancer was significantly improved (AUC increased from 0.61 to 0.66, P=0.014). We show that screening participants in the middle risk tertile achieves a lung cancer specific mortality reduction of 55% and all-cause mortality reduction of 21%. In this group the number of lung cancers averted is maximised (12/1000 person screened) and number needed to screen to avert one lung cancer reduced to 84. We show that this is achieved through minimising pre-existing co-morbid disease and by maximising screen detected lung cancers amenable to CT detection and successful surgical intervention. We believe genetic data provides useful information on lung cancer biology.

Conclusion:
The “Sweet spot” of CT screening comes from identifying high risk smokers optimised according to co-existing premorbid disease (especially COPD), early stage lung cancers amenable to surgical cure and least likely to die of other complications of smoking. Gene-based risk testing appears superior to just clinical risk models alone in prioritising high risk smokers for screening.

Background:
Lung cancer (LC) is the commonest cause of cancer-related death in the world. Screening with low-dose computer tomography (LDCT) had been shown to reduce LC specific and all-cause mortality. Benefit is greatest in those at highest risk, such as current smokers from areas of high socio-economic deprivation, yet participation in these ‘hard-to-reach’ populations remains a challenge and must be improved if we are to succeed with screening. The aim of this NHS implementation project was to assess LC screening within the community in deprived areas.

Method:
Ever smokers, aged 55-74, registered at 14 participating general practitioner (GP) practices in deprived areas of Manchester were invited to attend and have a free ‘Lung Health Check’ (LHC) in a mobile unit located at their local shopping centres. Lung cancer risk score (PLCO~M2012~), respiratory symptoms and spirometry were assessed as part of the LHC with results communicated back to the GPs. Those at high risk of LC, i.e. 6-year lung cancer risk ≥1.51%, were offered immediate LDCT in a co-located mobile CT scanner. These were all reported by thoracic radiologists with an interest in pulmonary oncology. Specifically designed nodule algorithms were followed in the reporting.

Result:
The maximum available capacity of the project was filled within days of going live. 2,541 individuals attended for a LHC and consented to data analysis. The mean age was 64.1±5.5, 51.0% (n=1,296) were female, 35.1% (n=891) were current smokers and 74.5% (n=1,893) ranked in lowest deprivation quintile. Of these 56.2% (n=1,429) qualified for a LDCT scan (PLCO~M2012~ risk score ≥1.51%). 46 lung cancers were detected in 42 individuals, a prevalence of 3.0%, of which 80% (n=37/46) were early stage (I+II). A treatment with curative intent was offered to 89.1% (n=41/46) of screen detected cancers and the surgical resection rate was 65.2%, which is almost fourfold the UK national average (16.8%).

Conclusion:
Taking lung cancer screening into the community can identify and target those at most risk, using the PLCO~m2012~ model, resulting in a significant stage shift in screen detected lung cancers in deprived populations.

Abstract not provided

Background:
Low-dose computed tomography (LDCT) lung cancer screening is recommended by US guidelines for high-risk individuals. New solid nodules are regularly found in incidence screening rounds and have a higher lung cancer probability at smaller size than do baseline nodules, leading to the proposal of lower size cutoffs at initial new solid nodule detection. However, currently there is no evidence concerning the risk-stratification of new solid nodules at first LDCT screening after initial detection.

Method:
In the ongoing, multicenter, randomized controlled Dutch-Belgian Lung Cancer Screening (NELSON) Trial, 7,295 participants underwent the second and 6,922 participants the third screening round. We included participants with solid non-calcified nodules, that were registered by the NELSON radiologists as new or smaller than 15mm[3] (study detection limit) at previous screens and received a follow-up or regular LDCT screening after initial detection; thereby excluding high-risk nodules according to the NELSON management protocol (nodules ≥500mm[3]). Nodule volume was generated semiautomatically. For assessment of the predictive performance, the area under the receiver operating characteristics curve (AUC) of nodule volume, volume doubling time (VDT), and VDT combined with a predefined 200mm[3] volume cutoff were evaluated with eventual lung cancer diagnosis as outcome.

Result:
Overall, 680 participants with 1,020 low and intermediate risk new solid nodules were included. A total of 562 (55%) new solid nodules were resolving, leaving 356 (52%) participants with a nonresolving new solid nodule of whom 25 (7%) were eventually diagnosed with lung cancer in such a nodule. At first follow-up or regular LDCT screening after initial new solid nodule detection, VDT, volume, and VDT combined with the predefined ≥200mm[3] volume cutoff had a high discriminative performance for lung cancer (VDT, AUC: 0.91; volume, AUC: 0.88; VDT and ≥200mm[3] combination, AUC: 0.94). A cutoff combination of ≤590 days VDT or ≥200mm[3] at first LDCT after initial new solid nodule detection, classifying a nodule positive when at least one criterion was fulfilled, provided 100% (95% confidence interval [CI] 84-100%) sensitivity and 84% (95%CI 80-87%) specificity for discriminating lung cancer, with positively classified nodules having a lung cancer probability of 27% (95%CI 19-37%).

Conclusion:
More than half of new solid nodules identified in LDCT lung cancer screening are resolving nodules. At first follow-up, a cutoff combination of ≤590 days VDT or ≥200mm[3] volume can be used for risk stratification.

Background:
New solid nodules detected in low-dose computed tomography (LDCT) lung cancer screening have a higher lung cancer probability at a smaller size than baseline nodules and lower size cutoff values for risk stratification at initial detection have been proposed. So far, it is unknown whether nodule characteristics, such as morphology or location, could improve risk stratification by size in new solid nodules.

Method:
This study forms part of the ongoing, randomized controlled Dutch-Belgian Lung Cancer Screening (NELSON) trial. This analysis included solid non-calcified nodules detected during the three incidence screening rounds and registered by the NELSON radiologists as new or previously below detection limit (15mm[3]). Nodule volume was generated semiautomatically. The predictive performance of nodule characteristics (location, distribution [peripheral, nonperipheral], shape [round, polygonal, irregular], margin [smooth, lobulated, spiculated, irregular], visibility <15mm[3] in retrospect) combined with previously established volume cutoffs (<30mm[3], low risk; 30-<200mm[3], intermediate risk; ≥200mm[3] high risk) was evaluated by multivariable logistic regression analysis with eventual lung cancer diagnosis as outcome. Discrimination of lung cancer based on volume, the final parsimonious model, and the model stratified into three risk groups (low, intermediate, high) was assessed through the area under the receiver operating characteristics curve (AUC) and compared using DeLong's Method.

Result:
Overall, 1,280 new nodules were included with 73 (6%) being diagnosed as lung cancer eventually. Of the new nodules visible <15mm[3] in retrospect and now ≥30mm[3], 22% (6/27) were lung cancer. Discrimination based on volume cutoffs (AUC: 0.80, 95% confidence interval [CI] 0.75-0.84) and continuous volume (AUC: 0.82, 95%CI 0.77-0.87) was comparable (P=0.14). After adjustment for volume cutoffs, only location in the right upper lobe (odds ratio [OR] 2.0, 95%CI 1.2-3.4), nonperipheral distribution (OR 2.4, 95%CI 1.4-4.2), and visibility <15mm[3] in retrospect (OR 4.7, 95%CI 1.7-12.8) remained significant predictors. Discrimination based on the model (AUC: 0.85, 95%CI 0.81-0.89) was superior to the volume cutoffs alone (P=0.0002), but when stratified into three risk groups (AUC: 0.82, 95%CI 0.78-0.86) discrimination was comparable (P=0.2).

Conclusion:
At initial detection, nodule volume is the strongest predictor for lung cancer in new nodules. Nodule characteristics may further improve lung cancer prediction, but only have limited incremental discriminatory value additional to volume cutoffs in a three-category stratification approach.

Background:
Substantial variation exists in the processes of care for potentially curable NSCLC. We examined the impact of thoroughness of staging for patients undergoing NSCLC surgery in a large, heterogeneous population within a lung cancer endemic region of the US.

Method:
We evaluated all surgically resected patients in the Mid-South Quality of Surgical Resection (MS-QSR) cohort from 2009-2017. MS-QSR is a population-based cohort including all curative-intent NSCLC resections at 11 hospitals in the mid-south US. Patients were classified into 8 groups based on use (Yes/No) of the following staging modalities: PET/CT, pre-operative invasive staging, operative mediastinal nodal examination (MLE). We compared stage distribution, adjuvant therapy, and overall survival outcomes across groups using the chi-square test and adjusted Proportional Hazards Models.

Result:
The 2,370 patients had a median age of 67 years, were 53% male. The racial distribution was: 70% White, 25% Black, 5% Other. Clinical N-stage was similar between the 8 groups. We found statistically significant differences in pathologic stage distribution, adjuvant therapy usage, and overall survival across the 8 groups (Table 1). Patients who received PET/CT, invasive staging, and MLE (Group 1) had significantly higher pathologic N-stage distribution compared to the other groups due to substantial nodal upstaging. Group 1 had 76% eligibility and 31% use of adjuvant chemotherapy compared to 51% and 8% in the Group 8 (No PET/CT, No Invasive Staging, No MLE). Use and eligibility for adjuvant radiation therapy was also highest in Group 1. There was an overall difference in survival across the groups (p-value=0.0019) which remained significant after adjusting for age, sex, race, histology, and path stage (p-value=0.0013). After adjustment, Group 8 had a 14% increased hazard of death compared with Group 1. Figure 1



Conclusion:
A less thorough approach to staging may lead to less nodal upstaging and less eligibility for adjuvant therapy, which could have implications for long term survival.

Abstract not provided