We appreciate the letter from Avan et al regarding our study examining the prognostic value of excision repair cross-complementing gene-1 (ERCC1) and ribonucleoside reductase subunit M1 (RRM2) expression in patients undergoing resection for pancreas adenocarcinoma. Observational studies have value in identifying potentially prognostic biomarkers that may help to risk stratify patients enrolled in trials and/or to guide the use of adjuvant therapy. A limitation of biomarker development using retrospective observational studies, as highlighted by Avan et al, is tissue heterogeneity and differences in techniques used for gene/protein analysis. In addition, the best method of assessment may differ by biomarker.
We agree with our colleagues that biomarker studies mandate expert selection of representative sections of tumor, as was done in our study and their previous works. Immunohistochemistry (IHC), although widely available and cost effective, is subject to inter-rater variability. We attempted standardization by using a semiquantitative scoring system, as detailed in the article. Evaluating mRNA levels using a quantitative technique such as reverse transcriptase-polymerase chain reaction (RT-PCR) is attractive but is limited by the potential difference between mRNA levels and protein expression. A recent study of ERCC1 in nonsmall cell lung cancer compared RT-PCR with IHC in patients treated on trial. The authors observed no correlation between RT-PCR results and IHC results, and IHC-based ERCC1 expression accurately discriminated survival.
Our study identifies ERCC1 and RRM2 as potential prognostic markers; however, given the heterogeneity of adjuvant therapy regimens, further work is required to determine the predictive value of these biomarkers (and others; ie, hENT1) for response to therapy. We congratulate our colleagues on their work in this field regarding hENT1. We strongly support incorporating tissue-based correlative studies into clinical trials in pancreas cancer and other gastrointestinal malignancies, as well as collaborative efforts to thoroughly investigate both prognostic and predictive biomarkers.