CA 19-9 Response: A Surrogate to Predict Survival in Patients With Metastatic Pancreatic Adenocarcinoma

OBJECTIVE: The objective of this study was to determine the features of carbohydrate antigen (CA) 19-9 decline that correlates best with survival benefit in patients with metastatic pancreatic cancer. METHODS: This is a retrospective study of 225 patients with metastatic pancreatic cancer receiving first-line therapy. Analysis was performed by the Kaplan-Meier method and Cox-proportional hazards ratios. CA 19-9 decline was grouped into quartiles within different CA 19-9 baseline groups. Time to nadir and CA 19-9 decline at month-2 (M2) of therapy were evaluated for patients with a baseline level >/=1000 U/mL. RESULTS: No significant trend in survival was observed across baseline CA 19-9 levels. The greatest survival benefit was seen with a >/=75% decline to nadir. Among those with a >/=75% decline and baseline >/=1000 U/mL, 43 of 57 patients had a >50% decline at M2 of therapy and additional survival benefit was observed with a slower decline to nadir. Small sample sizes limited analysis of other baseline groups. CA 19-9 decline at M2 was not predictive. CONCLUSIONS: In patients with a CA 19-9 >/=1000 U/mL, serial CA 19-9 levels may be considered as a surrogate for serial imaging to evaluate treatment response, with a >/=75% decline indicating the greatest survival benefit. Survival was improved further in the setting of a slower decline to nadir with the highest benefit seen in patients with a nadir occurring at 4 months or longer.