Gene Expression Analysis of Diagnostic Biopsies Predicts Pathological Response to Neoadjuvant Chemoradiotherapy of Esophageal Cancer

作者：Stephen G. Maher, Charles Gillham, Shane Patrick Duggan, Paul Smyth, Nicola Miller, Cian Muldoon, Kenneth John O'Byrne, Orla M. Sheils, Donal P. Hollywood, John Vincent Reynolds · 发表于：Annals of Surgery · 年份：2009 · DOI：10.1097/sla.0b013e3181bce7e1 · 被引用次数：76 · 研究领域：Esophageal Cancer Research and Treatment、Gene expression and cancer classification、HER2/EGFR in Cancer Research

In Brief Objective: This study explored gene expression differences in predicting response to chemoradiotherapy in esophageal cancer. Purpose: A major pathological response to neoadjuvant chemoradiation is observed in about 40% of esophageal cancer patients and is associated with favorable outcomes. However, patients with tumors of similar histology, differentiation, and stage can have vastly different responses to the same neoadjuvant therapy. This dichotomy may be due to differences in the molecular genetic environment of the tumor cells. Background Data: Diagnostic biopsies were obtained from a training cohort of esophageal cancer patients (13), and extracted RNA was hybridized to genome expression microarrays. The resulting gene expression data was verified by qRT-PCR. In a larger, independent validation cohort (27), we examined differential gene expression by qRT-PCR. The ability of differentially-regulated genes to predict response to therapy was assessed in a multivariate leave-one-out cross-validation model. Results: Although 411 genes were differentially expressed between normal and tumor tissue, only 103 genes were altered between responder and non-responder tumor; and 67 genes differentially expressed >2-fold. These included genes previously reported in esophageal cancer and a number of novel genes. In the validation cohort, 8 of 12 selected genes were significantly different between the response groups. In the predictive model, 5 of 8 genes could predict response ...