Identification of laminin γ2 as a prognostic and predictive biomarker for determining response to gemcitabine-based therapy in pancreatic ductal adenocarcinoma

BACKGROUND: Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal malignancies. While the extracellular matrix (ECM) components plays an integral role in PDAC pathogenesis and mediating chemoresistance, its role in predicting response to chemotherapy in PDAC patients remains unclear.
METHODS: We performed a systematic biomarker discovery by analyzing genomewide transcriptomic profiling data from 423 patients (GSE71729, GSE21501 and TCGA) for predicting overall survival (OS). This was subsequently validated in two independent clinical cohorts of 270 PDAC patients (training cohort; n=121 and validation cohort; n=149). In addition, we investigated EUS-FNA biopsy specimens from 51 PDAC patients with an unresectable cancer for predicting therapeutic response to gemcitabine-based therapy.
RESULTS: Following rigorous bioinformatic analysis, we identified LAMC2 to be a significant prognostic factor in all three PDAC datasets (GSE71729, HR=2.04, P=0.002; GSE21501, HR=2.17, P=0.031; TCGA, HR=2.57, P<0.001). High LAMC2 expression in PDAC patients associated with a significantly poor OS and relapse-free survival (RFS) in both training (P<0.001, P<0.001) and validation cohorts (P=0.001, P=0.003). More importantly, LAMC2 expression robustly identified PDAC patients with unresectable disease and those who responded to gemcitabine-based therapy (AUC= 0.79; 95%CI, 0.65-0.89). The univariate logistic regression analysis revealed that high LAMC2 expression was the only factor that predicted poor response to gemcitabine in PDAC patients (Odds Ratio [OR]=4.90; 95% CI, 1.45-16.6; P=0.011).
CONCLUSION: We conclude that LAMC2 is a novel prognostic and predictive biomarker for gemcitabine-based therapy in both adjuvant and palliative setting; which could have significant impact in precision and individualized treatment of PDAC patients.

© 2021. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/