Gene expression profiles in esophageal adenocarcinoma predict survival after resection

Objective

The incidence of esophageal adenocarcinoma is rapidly increasing in the western population. Despite aggressive treatment, survival after esophagectomy is suboptimal. The main objective of the present study was to evaluate the gene expression profiles in esophageal adenocarcinoma and determine their association with survival after resection.

Methods

We conducted a prospective National Institutes of Health/National Cancer Institute funded study to evaluate the prognostic significance of gene expression in patients with esophageal adenocarcinoma undergoing esophagectomy. Gene expression in tumor tissue was analyzed using high-throughput oligonucleotide arrays. The association of gene expression and overall survival was analyzed using the tail-strength statistic and Cox regression analysis. Gene signatures were constructed with semisupervised methods using principal components. A cross-validated risk score was devised by conducting 10-fold cross-validation, 100 times.

Results

We evaluated the gene expression in 64 patients with esophageal adenocarcinoma who underwent esophagectomy. The median overall survival was 27 months (95% confidence interval 22 to not reached). After filtering, 10,214 probe sets were used for survival analysis. The tail-strength statistic for these probe sets (0.318) indicated a significant association with overall survival. Patients were classified into high- and low-risk groups, according to the gene signature. High-risk patients had a predicted median survival of 19 months, but the median was not reached for the low-risk group (P < .05). On multivariate analysis, the gene signature was independently associated with survival (hazard ratio, 2.22; P = .04).

Conclusions

Global gene expression levels were significantly associated with overall survival after esophagectomy. Furthermore, individual genes could be successfully combined into a strongly predictive, internally cross-validated gene signature. If validated further, these results could help direct additional clinical trials of neoadjuvant and adjuvant therapies for esophageal adenocarcinoma.