Tumor markers in patients with pancreatic carcinoma

BACKGROUND: Tumor markers are putative prognostic indicators for patients with carcinoma, but have not heretofore been evaluated in patients with Stage II and III pancreatic carcinoma. METHODS: Patients with Stage II (n=9) and Stage III (n=25) unresectable regional adenocarcinoma of the pancreas were treated with combined modality therapy. Treatment consisted of split course radiotherapy and simultaneous combination chemotherapy with fluorouracil infusion, streptozotocin, and cisplatin. Prior to treatment, patients free of both infection and jaundice provided blood for CA 19-9, carcinoembryonic antigen (CEA) and CA 125 assays. RESULTS: The overall median survival of Stage II patients was 21.1 months. Due to the small number of Stage II patients with markedly abnormal assays, it was not possible to test for a statistically significant association between pretreatment tumor assays and survival. Among patients with Stage III pancreatic carcinoma, a CA 19-9 assay of 2000 u/ML or less identified a group of 16 patients with a median survival of 12.8 months. In contrast, 8 Stage III patients with a CA 19-9 assay of greater than 2000 u/mL had a median survival of 8 months and only 1 patient survived for 1 year (P=0.020, log rank test; P=0.010, Wilcoxon test). Among Stage III patients, a comparison of those with a normal assay versus any degree of abnormal assay failed to provide prognostic information. Analyses based on a combination of CA 19-9 and CA 125 assays provided additional powerful prognostic information: (P=0.002, log rank test; P=0.005, Wilcoxon test). CEA assays failed to provide information alone or in combination with the CA 19-9 assay. After adjusting for the CA 19-9 assay in multivariate analyses, neither performance status nor tumor size were significant prognostic variables for patients with Stage III cancers. CONCLUSIONS: Pretreatment CA 19-9 assays provide powerful independent and objective prognostic information.