A Randomized Phase II Trial of Neoadjuvant Chemotherapy Compared with Chemoradiotherapy in Locally Advanced Gastric Adenocarcinoma

Abstract

This phase II study was designed to evaluate whether NACRT was superior to NACT with both followed by surgery and postoperative chemotherapy for locally advanced gastroesophageal and gastric adenocarcinoma. Patients with resectable or unresectable gastric cancer (cT3-4NxM0 or cTxN1-3M0) were randomized to either NACT arm or NACRT arm in a 1:1 ratio with stratification by clinical T stage (cT1-3 vs cT4). NACT arm consisted of three cycles of SOX followed by radical surgery and another postoperative three cycles of SOX. NACRT arm received intensity-modulated radiation therapy with a simultaneous integrated boost (SIB-IMRT) to primary tumor (45.1Gy and 40.04Gy in 22 fractions) concurrently with S-1 followed by surgery and four to six cycles of SOX at the same dosage with NACT arm. The primary end point was surgical resection rate (NCT02301481). From January 2014 to October 2017, 75 patients were randomly assigned to two arms: 38 in NACRT and 37 in NACT arm. Since two of each arm refused surgery after neoadjuvant treatment (NA), 71 were finally eligible for evaluation. R0 resection rate and D2 lymphadenectomy were performed in 77.8% (28/36) and 92.9% (26/28) in NACRT and 77.1% (27/35) and 96.3% (26/27) in NACT arm. The severe and moderate pathologic responses were significantly achieved with a higher rate of 96.6% (27/28) in NACRT than in NACT arm (66.7%, 18/27, P=0.013), in which 14.3% (4/28) and 11.1% (3/27) had pCR (P=0.724). NACRT arm were observed with much less detected median total number of lymph nodes than NACT (25 vs 37, P<0.001). More grade 1 to 3 myelosuppression including thrombocytopenia (61.3% vs 5.3%, P<0.001), neutropenia (77.4% vs 28.9%, P=0.067) and liver dysfunction (16.1% vs 0, P=0.037) occurred in NACT and NACRT arm had more gastritis (39.5% vs 5.9%, P=0.008) and radiation esophagitis (26.7% vs 0, P=0.001). There were no toxic death or postoperative death in both arms. Postoperative complications were similar in the two treatment groups (NACRT vs NACT, 6.9% vs 7.1%). The compliance with NACRT was 97.4% for radiation and 89.5% for S-1. In NACT arm, 97.3% of patients completed three planned cycles preoperatively. However, only 62.1% and 66.7% completed planned cycles of adjuvant chemotherapy in NACRT and NACT arm. With a median follow-up of 27 months, the 2-year overall survival was not significant difference in all 71 patients (NACRT vs NACT, 75.4% vs 70.1%, P=0.455) and in patients with R0 resection (NACRT vs NACT, 91.3% vs 80.3%, P=0.113). However, patients with R0 resection treated with NACRT were associated with significantly higher disease-free survival (DFS, 87.1% vs 63.9%, p=0.050) and locoregional recurrence free survival (LRFS, 100% vs 79.3%, p=0.014) as compared with NACT. The design of preoperative concurrent SIB-IMRT with oral S-1 showed better DFS and LRFS with an acceptable toxicity profile, which encouraged future randomized phase III trials comparing NACRT with NACT for resectable or unresectable gastric adenocarcinoma.