A prospective trial for defining a subset of patients with limited metastatic gastric cancer who may be candidates for bimodal treatment strategies: FLOT3.

Abstract

4090 Background: The utility of surgery for metastatic gastric cancer is debated. A prospective trial was performed to evaluate a prognostic model for selecting patients (pts) treated with systemic chemotherapy (ct) who may also be candidates for surgical intervention. Methods: Using a predefined algorithm pts with untreated gastric cancer were prospectively stratified into 3 groups: operable (OD), limited metastatic (LD), or extensive metastatic (ED) disease and treated with 5-FU, oxaliplatin, leucovorin and docetaxel (FLOT). LD was defined as: distant intra-abdominal lymph node metastases only or/and a maximum of 1 organ involved, normal serum alkaline phosphatase, < 5 liver lesions, no visible carcinomatosis (peritoneum or pleura), and ECOG ≤ 1. All other metastatic pts were ED. Pts with OD received 4 preoperative ct cycles followed by surgery and 4 postoperative cycles. Pts with LD received 8 cycles with surgery allowed for complete macroscopic resection. Pts with ED received 8 cycles with surgery allowed for palliation only. The study had 80% power to detect a HR of 0.55 for overall survival in favor of the LD group (vs. ED group; 2-sided log-rank p=0.05). Results: 238 of 252 pts included were eligible (OD/LD/ED: 51/60/127). LD pts had distant lymph nodes only (41%), liver (22%), lung (17%), localized peritoneal involvement (7%), or others (13%). A median of 8 ct cycles was applied to all groups. Median OS was 22.9 vs. 10.7 months in pts with LD vs. ED, respectively (HR 0.37; 95% CI, 0.25 – 0.56; p <0.001). LD was the strongest predictor of OS in the multivariate analysis including all single determinants of LD status (p=.002). Surgical resection was conducted in 96%, 62%, and 12% in the OD, LD, and ED groups, of which R0 resection (primary) was achieved in 82%, 81% and 33%, respectively. Within the LD arm, operated pts had better outcome than non-operated pts (median OS 31.3 vs. 15.9 months; p=.004) and pts with lymph node only involvement had best outcome. Conclusions: This clinical model identifies a subset of pts with (limited) metastatic gastric cancer who have a favorable outcome and who may be candidates for bi-modal treatment strategies.