682PPhase III, international, multicenter, randomized, open-label trial of adjuvant nab-paclitaxel plus gemcitabine (nab-P/G) vs gemcitabine (G) alone for surgically resected pancreatic adenocarcinoma (APACT): Subgroup analyses

Abstract Background The primary APACT analysis showed no significant difference in disease-free survival (DFS) by blinded, independent radiologic review with nab-P/G vs G in patients (pts) with surgically resected pancreatic cancer (PC). However, investigator-assessed DFS (prespecified sensitivity analysis) and interim overall survival (OS; secondary endpoint) trended in favor of nab-P/G. Here, we report interim OS exploratory subanalyses. Methods Treatment-naive pts with histologically confirmed PC, macroscopic complete resection (R0/R1), CA 19-9 < 100 U/mL, and ECOG PS 0 or 1 were enrolled. Stratification factors included resection (R0/R1), lymph node (LN) status (positive [+]/negative), and geographic region. Pts were treated ≤ 12 weeks after surgery with nab-P 125 mg/m2 + G 1000 mg/m2 or G 1000 mg/m2 alone on days 1, 8, and 15 every 28 days for 6 cycles. For the primary endpoint assessment, independent reviewers received only baseline clinical data and scans. OS and safety were secondary endpoints. Ten prespecified subanalyses were performed. Results 866 pts were randomized. Median age was 64 y (range, 34–86); most were male (56%) and white (78%) and had ECOG PS 0 (60%), LN+ status (72%), and R0 resection (76%). At the original data cutoff (31 December 2018; median follow-up, 38.5 mo), median OS (interim) trended in favor of nab-P/G vs G (40.5 vs 36.2 mo; HR 0.82; 95% CI, 0.680 - 0.996; nominal P = 0.045). Pts with poor characteristics had numerically longer median OS with nab-P/G vs G: 32.5 vs 27.0 mo in pts with R1 resection (n = 105 and 100) and 33.8 vs 28.9 mo in pts with LN+ status (n = 311 and 312). This is consistent with the median OS observed with nab-P/G vs G in pts whose tumors were < 1 mm to the closest margin (32.5 vs 29.7 mo; n = 114 and 112). Pts with both R1 resection and LN+ status had numerically longer median OS with nab-P/G vs G (30.7 vs 24.9 mo; n = 87 and 83). Conclusions Final OS data may clarify the role for adjuvant nab-P/G. Interim OS analyses suggest that continued investigation of adjuvant nab-P/G for pts with suboptimally resected PC or who may not tolerate FOLFIRINOX is warranted. Clinical trial identification NCT01964430. Editorial acknowledgement Rebecca Tweedell, MediTech Media, Ltd, funded by Celgene Corporation. Legal entity responsible for the study The authors. Funding Celgene Corporation. Disclosure M.A. Tempero: Advisory / Consultancy: AbbVie, Inc., Advance Medical, Inc., BioPharm Communications, Bristol-Myers Squ