Early results of a randomized phase II, compass trial to compare regimen and duration of neoadjuvant chemotherapy for gastric cancer.

Abstract

4102 Background: Prognosis for stage III gastric cancer was not satisfactory even by D2 gastrectomy and adjuvant chemotherapy. Neoadjuvant chemotherapy is another promising approach. This study investigated the outcomes of two and four courses of neoadjuvant S-1/cisplatin (SC) and paclitaxel/cisplatin (PC) using a two-by-two factorial design for locally advanced gastric cancer. Methods: Patients with stage II schirrhous/junctional tumors, stage III, or resectable stage IV, 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). The primary endpoint was 3-year OS. Key secondary endpoints included pathological/clinical response, R0 resection, and adverse events. Sample size was set at 60 to 80 to achieve 10% improvement of 3-year OS by four courses or by PC with approximately 80% probability of the correct selection. Results: Between Oct 2009 and July 2011, 83 patients were assigned to arm A (2 courses of SC, n=21), arm B (4 courses of SC, n=20), arm C (2 courses of PC, n=21), and arm D (4 courses of PC, n=21). Clinical response (arm A/B/C/D) was 29%/40%/33%/24%. R0 resection (arm A/B/C/D) was 76%/75%/57%/76%. Pathological response (arm A/B/C/D), defined as tumor regression more than two third in the primary tumor, was 43%/40%/29%/38%. Pathological complete response (arm A/B/C/D) was 0%/10%/0%/10%. Major grade 3/4 toxicities (arm A/B/C/D) were anemia (14%/15%/0%/28.6%), neutropenia (10%/15%/14%/33%), nausea (0%/10%/5%/5%), and appetite loss (5%/10%/0%/5%). Pathological complete response by per-protocol analysis (arm B/D) was 17% and 12%. Treatment discontinuation (number of patients, arm A/B/C/D) was disease progression (1/3/0/1), toxicities (1/4/0/3), and others (0/1/0/0). No surgical mortality was observed. Grade 3 morbidity classified by Clavien-Dindo was leakage in 5% (arm A), pancreatic fistula in 5% (arm C), and postoperative hemorrhage in 5% (arm B). Conclusions: Pathological complete response could be induced by four courses of neoadjuvant chemotherapy without a marked increase of toxicities, regardless of a SC or PC regimen. Clinical trial information: UMIN000002595.