Added value of CA19-9 response in predicting resectability of locally advanced pancreatic cancer following induction chemotherapy

Abstract

Background: Determining the resectability of locally advanced pancreatic cancer (LAPC) after induction chemotherapy is complex since CT-imaging cannot accurately portray tumor response. We hypothesized that CA19-9 response adds to RECIST-staging in predicting resectability of LAPC. Methods: Post-hoc analysis within a prospective study on LAPC (>90° arterial or >270° venous involvement). CA19-9 response was determined after induction chemotherapy. Surgical exploration was performed in RECIST-stable or -regressive disease. The relation between CA19-9 response, resectability and survival was assessed. Results: Restaging in 54 patients with LAPC after induction chemotherapy (mostly FOLFIRINOX) identified 6 RECIST-regressive, 32 RECIST-stable, and 16 patients with RECIST-progressive disease. The resection rate was 20.3% (11/54 patients). Sensitivity and specificity of RECIST-regression for resection were 40% and 87% whereas the positive predictive value (PPV) and negative predictive value (NPV) were 67% and 68%. Using a 30% decrease of CA19-9 as cut-off, 9/10 patients were correctly classified as resectable (90% sensitivity, PPV 43%) and 3/15 as unresectable (20% specificity, NPV 75%). In the total cohort, a CA19-9 decrease ≥30% was associated with improved survival (22.4 vs. 12.7 months, p = 0.02). Conclusion: Adding CA19-9 response after induction chemotherapy seems useful in determining which patients with RECIST non-progressive LAPC should undergo exploratory surgery.