Factors affecting resection rate in patients undergoing neoadjuvant therapy for pancreatic cancer.

Abstract

e14531 Background: Patients who derive benefit from neoadjuvant therapy in pancreatic cancer is unclear. A subset, borderline resectable (BR) has been defined in whom therapy may be of benefit. We evaluated factors affecting resection, in particular those patients deemed borderline resectable. Methods: This is a retrospective chart review of patients who received neoadjuvant therapy at the University of Oklahoma over a 5 year period. Pre-therapy scans were reviewed and classified as locally advanced (LA) or BR per the AHPRA/SSO/SSAT Consensus Conference. Unless otherwise speicified, p-values are exact calculations from Chi-square tests on the equality of two independent proportions. Results: Among 18 patients studied, 9 underwent surgery. Pathology was available for 8 patients. All underwent R0 (62%) or R1 (38%) resections. Amongst those with LA, 5/6 (83%) underwent surgery, compared with 4/12 (33%) with BR disease (p value Pearson chi-square=0.1312). Kaplan-Meier analysis estimated median survival at 24 months in those with LA and 19 months with BR disease (p from log rank test=0.3475). Seven of 12 (58%) of patients who received only chemotherapy underwent surgery, compared with 2 of 6 (33%) patients who received chemoradiation (exact p=0.6199). High baseline level of Ca19-9 (>100) was not predictive of resectability (exact p=1.0) in the 12 patients whose Ca19-9 expression was unaccompanied by high levels of total bilirubin (>1.4). Among these patients, mean Ca19-9 expression was nonsignificantly higher (exact p from Wilcoxon test=0.5417) among those with BR disease. Two of 7 (29%) patients ≥ 65 y/o underwent surgery, compared with 7 of 11 (64%) younger patients (exact p=0.3348). Outcome did not differ between patients who underwent surgery more than 90 days from start of therapy (8/14 or 57%) vs less than 90 days (1/4 or 25%; exact p=0.5765). Conclusions: A higher percentage of LA patients underwent resection than those with BR. This preliminary data underscores the importance of disease biology in this patient group and the need, in future multivariable analysis on more patients to adjust for factors like age, CA 19-9 levels and pathologic grade that may independently affect surgical candidacy and survival.