Radiographic and Serologic Response to First-Line Chemotherapy in Unresected Localized Pancreatic Cancer

Background: This study aimed to determine the clinical relevance of putative radiographic and serologic metrics of chemotherapy response in patients with localized pancreatic cancer (LPC) who do not undergo pancreatectomy. Studies evaluating the response of LPC to systemic chemotherapy have focused on histopathologic analyses of resected specimens, but such specimens are not available for patients who do not undergo resection. We previously showed that changes in tumor volume and CA 19-9 levels provide a clinical readout of histopathologic response to preoperative therapy. Methods: Our institutional database was searched for patients with LPC who were treated with first-line chemotherapy between January 2010 and December 2017 and did not undergo pancreatectomy. Radiographic response was measured using RECIST 1.1 and tumor volume. The volume of the primary tumor was compared between pretreatment and posttreatment images. The percentage change in tumor volume (%Dvol) was calculated as a percentage of the pretreatment volume. Serologic response was measured by comparing pretreatment and posttreatment CA 19-9 levels. We established 3 response groups by combining these metrics: (1) best responders with a decline in %Dvol in the top quartile and in CA 19-9, (2) nonresponders with an increase in %Dvol and in CA 19-9, and (3) other patients. Results: This study included 329 patients. Individually, %Dvol and change in CA 19-9 were associated with overall survival (OS) (P#.1), but RECIST 1.1 was not. In all, 73 patients (22%) were best responders, 42 (13%) were nonresponders, and there were 214 (65%) others. Best responders lived significantly longer than nonresponders and others (median OS, 24 vs 12 vs 17 months, respectively; P,.01). A multivariable model adjusting for type of chemotherapy regimen, number of chemotherapy doses, and receipt of radiotherapy showed that best responders had longer OS than did the other cohorts (hazard ratio [HR], 0.35; 95% CI, 0.21–0.58 for best responders, and HR, 0.55; 95% CI, 0.37–0.83 for others). Conclusions: Changes in tumor volume and serum levels of CA 19-9—but not RECIST 1.1—represent reliable metrics of response to systemic chemotherapy. They can be used to counsel patients and families on survival expectations even if pancreatectomy is not performed.