Predictors of Undergoing Surgical Resection after Neoadjuvant Chemoradiotherapy for Borderline Resectable and Resectable Pancreatic Adenocarcinoma

Abstract

<h3>Purpose/Objective(s)</h3> Neoadjuvant chemoradiotherapy (CRT) serves to downstage and increase the rate of surgical resection for borderline resectable and resectable pancreatic adenocarcinoma ((B)RPC). Although conventional staging provides prognostic information, it remains a challenge to predict which patients will proceed with resection. We hypothesize that the severity of vessel involvement and the trajectory of CA19-9 levels after neoadjuvant treatment would associate with proceeding to resection. <h3>Materials/Methods</h3> Medical records were retrospectively reviewed at a single center for patients diagnosed with (B)RPC and treated with neoadjuvant CRT between 1/2005 and 12/2020. Progress notes were evaluated to determine if a patient proceeded to surgery and the trajectory of CA 19-9 levels in response to treatment. Radiologist notes were evaluated to determine the extent of mesenteric vessel (superior mesenteric artery, celiac axis, portal vein, hepatic artery, and superior mesenteric vein) involvement. A vessel involvement score was calculated as the sum of all vessels in contact with tumor: 1 point for each vessel, with partial (< 180 degrees) encasement of artery counting as 2 points. Odds ratios (OR) of proceeding to resection were computed with 95% confidence intervals (CI) for CA19-9 response to treatment. Ordinal logistic regression models were fit to estimate the impact of greater vessel involvement on the odds of NOT proceeding with surgery. <h3>Results</h3> Forty-two (n = 42) patients were included in the analysis. The median age was 65 years (IQR: 61 to 71). 73% of patients were treated with 3600 cGy in 15 fractions (fx), 24% of patients were treated with 4500-5040 cGy in 25-28 fx, and 3% of patients were treated with 3300 cGy in 5 fx. 55% of patients were treated with neoadjuvant FOLFIRINOX and 45% of patients were treated with other regimens. 88% of patients were considered borderline resectable and 12% were resectable at the time of diagnosis. Patients with any decrease in CA19-9 levels after CRT had higher odds of proceeding with resection (OR = 21.0, 95% CI: 3.2 to 139.2). However, decrease in CA19-9 levels after upfront chemotherapy was not significantly associated with proceeding with resection (<i>P</i> = 0.91). The odds of NOT proceeding with resection after neoadjuvant CRT were 1.81 times higher (95% CI: 0.97 to 3.39) for each increase in vessel involvement score, however these findings were not statistically significant (<i>P</i> = 0.062). <h3>Conclusion</h3> Patients diagnosed with (B)RPC proceeding with CRT have different odds of proceeding to surgery based on their response to CRT and extent of vessel involvement. Post-CRT CA19-9 can help predict which patients may proceed to resection. Future investigations will develop nomograms for estimating odds of proceeding to surgery in larger datasets to help aid in patient counseling and multidisciplinary discussion.