Prospective Phase II Trials Validate the Effect of Neoadjuvant Chemotherapy on Pattern of Recurrence in Pancreatic Adenocarcinoma

Abstract

The objective of this study was to characterize the patterns of first recurrence after curative-intent resection for pancreatic adenocarcinoma (PDAC). We evaluated the first site of recurrence after neoadjuvant treatment as locoregional (LR) or distant metastasis (DM). To validate our findings, we evaluated the pattern from 2 phase II clinical trials evaluating neoadjuvant chemotherapy (NAC) in PDAC. We identified site of first recurrence from a retrospective cohort of patients from 2011 to 2017 treated with NAC followed by chemoradiation and then an operation or an operation first followed by adjuvant therapy, and 2 separate prospective cohorts of patients derived from 2 phase II clinical trials evaluating patients treated with NAC in borderline-resectable and locally advanced PDAC. In the retrospective cohorts, 160 out of 285 patients (56.1%) recurred after a median disease-free survival (mDFS) of 17.2 months. The pattern of recurrence was DM in 81.9% of patients, versus LR in 11.1%. This pattern was consistent in patients treated with upfront resection and adjuvant chemotherapy (DM 83.0%, LR 16.9%) regardless of margin-involvement (DM 80.1%, LR 19.4%). The use of NAC did not alter pattern of recurrence; 81.7% had DM and 18.3% had LR. This pattern also remained consistent regardless of margin-involvement (DM 94.1%, LR 5.9%). In the Phase II borderline-resectable trial (NCI# 01591733) cohort of 32 patients, the mDFS was 34.2 months. Pattern of recurrence remained predominantly DM (88.9%) versus LR (11.1%). In the Phase II locally-advanced trial (NCI# 01821729) cohort of 34 patients, the mDFS was 30.7 months. Although there was a higher rate of local recurrence in this cohort, pattern of first recurrence remained predominantly DM (66.6%) versus LR (33.3%) and remained consistent independent of margin-status. The pattern of recurrence in PDAC is predominantly DM rather than LR, and is consistent regardless of the use of NAC and margin involvement.