A randomized 2X2 phase II trial comparing two and four courses of S-1/cisplatin (SC) and paclitaxel/cisplatin (PC) as neoadjuvant chemotherapy for locally resectable advanced gastric cancer: Survival results of COMPASS.

Abstract

111 Background: The prognosis for stage III gastric cancer is not satisfactory even by S-1 adjuvant chemotherapy. Neoadjuvant chemotherapy is a promising approach but its optimal duration and regimen have not been established yet. Methods: We conducted a randomized phase II trial to compare two or four courses and regimen of SC or PC using a two-by-two factorial design.Key eligibility criteria was (i) T2-3/N+ or T4aN0 in case of schirrhous or junctional tumors, T2-3 with N+ to the major branched artery, T4aN+, T4b, para-aortic nodal metastases, or resectable minimal peritoneal metastases confirmed by laparoscopy and (ii) no other distant metastasis. Patients received S-1 (80 mg/m2 for 21 days with 1 week rest) / cisplatin (60 mg/m2 at day 8) or paclitaxel / cisplatin (80 mg/m2 and 25 mg/m2, respectively, on days 1, 8, and 15 with 1 week rest) as neoadjuvant chemotherapy. Then, patients received D2 gastrectomy with curative intent. The primary endpoint was 3-year overall survival. The planned sample size was 80 eligible patients in total so that the treatment group with the superior observed 3-year OS rate by 10% increase was to be selected with a probability of 88% or higher. Results: Eighty-three patients were assigned to SC (n=41, two courses in 21 and four courses in 20) and PC (n=42, two courses in 21 and four courses in 21). Pathological response rate was 42% (17/41) in SC and 33% (14/42) in PC, and 36% (15/42) in the two courses and 39% (16/41) in the four courses. Pathological CR was 0% (0/42) in the two courses and 10% (4/41) in the four courses. Grade 3/4 adverse events by chemotherapy and grade 3/4 surgical morbidities defined by Clavien-Dindo classification were both less than 10% in each arm without treatment-related death. The 3-year OS was 60.9% (95% CI, 44.3-73.9%) in SC and 64.3% (95% CI, 47.9-76.7%) in PC, and 64.3% (95% CI, 47.9-76.7%) in the two courses and 61.0% (95% CI, 44.4-74.0%) in the four courses. Conclusions: Two courses of SC as neoadjuvant chemotherapy is recommended for a test arm of future phase III study for patients with locally resectable advanced gastric cancer. Clinical trial information: UMIN000002595.