559 Predictive Markers of Gemcitabine Treatment Response in Pancreas Cancer: A Pharmacogenomic Pathway Approach

Background: The development of effective non-Gemcitabine based treatment options for pancreatic cancer (e.g. FOLFIRONOX) opens the door for the clinical need for predictive markers of Gemcitabine response to assist in treatment selection. Our group previously validated hENT1, a Gemcitabine membrane transporter, as a predictive marker of Gemcitabine treatment response using RTOG 9704. Controversy exists about the predictive value of other components of the Gemcitabine metabolism pathway: DCK, RRM1, RRM2 and p53R2. We used the RTOG 9704 data set to evaluate the potential predictive value for Gemcitabine response of these markers. Methods: RTOG 9704 prospectively randomized 538 patients after pancreatic resection to receive either 5-FU or Gemcitabine. Tumor DCK, RRM1, RRM2, and p53R2 protein expression was analyzed in patients from this study using a Tissue Microarray (TMA) and IHC. Overall survival (OS) and disease-free survival (DFS) were estimated univariately with the Kaplan-Meier method and multivariate Cox proportional hazards models were used to determine if deoxycytidine kinase (DCK), ribonucleotide reductase M1 (RRM1), ribonucleotide reductase M2 (RRM2), and p53R2 protein levels correlated with OS and DFS. Results: 203 patients were eligible for analysis from both the 5-FU and Gemcitabine arm. DCK, RRM1, RRM2 and p53R2 protein expression was not associated with OS or DFS in the Gemcitabine treatment arm (Table 1). RRM1 and DCK expression did appear to be associated with OS in the entire treatment group and the 5FU treatment group, but not the Gemcitabine treatment arm. Conclusion: Despite limited data from nonrandomized treatment data suggesting a possible predictive role for DCK or RRM1 for determining response to Gemcitabine, our data from this prospective randomized treatment trial, which included a Gemcitabine alone arm, does not support the predictive value of DCK, RRM1, RRM2 or p53R2 in determining response to Gemcitabine. Efforts should focus on hENT1 as a valid predictive marker of treatment response of Gemcitabine in pancreas cancer. This project was supported by RTOG grant U10 CA21661, RTOG BsR grant U24 CA114734, and CCOP grant U10 CA37422 from the National Cancer Institute (NCI) and by the RTOG Translational Research Program. Overall Survival for all patients, Gemcitabine Treatment Arm and 5-FU Treatment Arm. Univariate Hazard Ratios (p-value)