4118 Background: S100A2, a calcium-binding protein, has been validated as a prognostic marker in a retrospective cohort of patients with PC who did not receive adjuvant therapy.
METHODS
A secondary analysis of S100A2 expression on archival tissue from a subset of patients with head of pancreas lesions from a prospective adjuvant trial, RTOG 9704, was performed. Immunohistochemistry was performed on a tissue microarray (TMA) including triplicate samples from each specimen. Tissue was scored for % labeled cells and intensity to establish 4 levels of expression. Disease-specific survival (DSS) was estimated univariately with the Kaplan-Meier method and multivariate Cox proportional hazards models were used to determine if there is a correlation between S100A2 expression and DSS.
RESULTS
The analysis was based on 150 specimens. For DSS, the hazard ratio (HR) for no, low, mod, or high expression for each subgroup is as follows: 1.0, 1.22 (95% C.I. = [0.70, 2.12]), 1.36 (95% C.I. = [0.77, 2.41]), and 1.37 (95% C.I. = [0.95, 1.99]), respectively. For high vs. low/no expression, the curves separated at 1 year but were not significantly different (p=0.09) (Table). Interestingly, the 2-year DSS in the retrospective S100A2 cohort was only 17% compared to 29% in the prospective RTOG cohort, while the 2-year DSS of the S100A2 negative cohorts were similar in both studies. In this clinical setting, S100A2 expression was not an independent prognostic factor based on multivariate analysis.
CONCLUSIONS
Failure to validate S100A2 as a prognostic marker in RTOG 9704 may be a function of small number of archival samples available for analysis. In addition, comparison to the historical retrospective untreated cohort suggests that S100A2 may also be a predictive biomarker for response to treatment. S100A2 deserves further study as a prognostic and predictive biomarker. This project was supported by the RTOG U10 CA21661, U24 CA114734, and CCOP U10 CA37422 grants from the National Cancer Institute (NCI). [Table: see text].