Virtual Library
Start Your Search
H. Ma
Author of
-
+
P3.04 - Poster Session/ Biology, Pathology, and Molecular Testing (ID 235)
- Event: WCLC 2015
- Type: Poster
- Track: Biology, Pathology, and Molecular Testing
- Presentations: 1
- Moderators:
- Coordinates: 9/09/2015, 09:30 - 17:00, Exhibit Hall (Hall B+C)
-
+
P3.04-085 - Anti-Glycan Antibody Profiling in De-Novo Stage IV Non-Small Cell Lung Cancer: A Pilot Study (ID 1447)
09:30 - 09:30 | Author(s): H. Ma
- Abstract
Background:
NSCLC is a heterogeneous disease with marked molecular and genomic variability that usually presents in advanced stage. Aberrant glycosylation occurs early during malignant transformation producing tumor-associated carbohydrate antigens (TACAs). Immune response to TACAs can be evaluated by immunoprofiling serum anti-glycan antibodies (AGAs) using printed glycan arrays (PGA), a new biomarker-discovery platform. Control and stage I NSCLC immunoprofiles were evaluated in a parallel study involving 70 subjects with high risk for developing lung cancer enrolled in low-dose CT lung cancer screening trial, 20 of who subsequently developed stage I NSCLC and a putative signature for early stage NSCLC has been obtained. The objective of this study was to obtain a putative signature of de-novo stage IV NSCLC patients using serum AGA immunoprofiles.
Methods:
18 patients were enrolled in this prospective study. Data collected included demographics, tumor histology, EGFR mutational status, and cancer treatment details. Blood sample was collected at each visit. Response was assessed after every two cycles of first-line chemotherapy regimen by a radiologist using RECIST 1.1 and the best overall response and time-to-progression (TTP) were recorded. Patients were dichotomized as “good” or “poor” responders by median TTP. AGAs were immunoprofiled in 30 microliters of serum using PGA with 382 glycans. The raw PGA data were screened to remove glycans with critically low signal intensities, low Intra-array Correlation Coefficient (ICC) and high coefficient of variation (CV) of on-slide replicates. The raw data were normalized with intra-slide linear normalization, and log-transformed. The putative glycan signature was obtained by our novel Compound ImmunoRuler (CIR) algorithm based on bootstrap aggregation of multiple signatures derived from correlation-adjusted Wilcoxon ranking using projections based on multivariate logistic regression. The training of CIR was performed on the baseline immunoprofiles of de novo stage IV samples with control immunoprofiles from the screening trial.
Results:
Study population included 11 males and 7 females. Mean age was 62 years (range 47-80). 15 patients (83%) had adenocarcinoma, two squamous, and one poorly differentiated. Two had EGFR mutation. Most common regimen was platinum/pemetrexed (n=14) followed by platinum/gemcitabine (n=2) and erlotinib (n=2). One patient had a complete response (CR); 6 PR and 4 had stable disease while 7 progressed. The median TTP was 140 days (41-387). For PGA data analyses, 278 glycans with ICC>80% were used for downstream analysis which delivered 3 glycans-based putative signature of Stage IV de novo NSCLC with AUC value 0.956; specificity 86%, and sensitivity 88.9%. Glycans in signatures of stage I and de-novo stage IV NSCLC were distinctly different. When immunoprofiles of stage IV NSCLC were projected on stage I ImmunoRuler, 13 out of 18 (72%) stage IV patients were accurately recognized as malignant. AGA binding to 4 glycans was significantly different between the sub-groups of “poor” vs. “good” responders.
Conclusion:
Uniquely different serum AGA-signatures of early stage and de novo stage IV lung cancer may provide basis for minimally-invasive test for early detection of risk for these malignancies and for better understanding their underlying pathologies. Further study in a larger population should help validate these findings.