Therapeutic value of para-aortic lymph node dissection in gastric cancer with extensive lymph node metastasis: Integrated analysis of three phase II trials (JCOG2212A).

419 Background: Gastric cancer (GC) sometimes forms metastatic bulky nodes around celiac artery or its branches (Bulky N), or metastasizes to para-aortic node (PAN) without any other distant metastasis. These severe nodal metastases are seldom cured by surgery alone. Japan Clinical Oncology Group (JCOG) defined these tumors as an extensive lymph node metastasis (ELM), and has developed multidisciplinary treatment through 3 phase II trials (JCOG0001, 0405, and 1002). JCOG0405 and 1002 reported the efficacy of D2 gastrectomy and PAN dissection (PAND) after neoadjuvant chemotherapy for GC with ELM with favorable survival outcomes. However, it remains unclear whether PAND truly contributed to the survival. The aim of this study was to clarify an optimal lymph node dissection area for GC with ELM. Methods: Therapeutic value index (TI) was investigated using an integrated dataset of JCOG0001, 0405, and 1002. Irinotecan and cisplatin, S-1 and cisplatin, or docetaxel, cisplatin and S-1 were used as neoadjuvant chemotherapy in JCOG0001, 0405, or 1002, respectively. Patients were classified into Bulky N group (only bulky N without PAN) and PAN group (PAN regardless of bulky N). TI was calculated by the following formula: proportion of metastasis (%) × 5-year relapse-free survival (5y-RFS, %) of patients with metastasis in the respective lymph node area / 100. Subgroup analysis by pathological response was also examined using the Japanese Classification of Gastric Carcinoma (JCGC) criteria. Grade ≥ 2a in JCGC criteria was defined as major pathological response (MPR). Results: A total of 122 patients were analyzed (Bulky N group: 68, PAN group: 54). Proportion of metastasis in peri-gastric (PG), supra-pancreatic (SP), and PAN areas were 80.9%, 54.4%, and 15.2% in Bulky N group, and 79.6%, 51.9%, and 44.4% in PAN group, respectively. 5y-RFS of patients with metastasis in PG, SP, and PAN areas were 38.2%, 29.7%, and 30% in Bulky N group and 23.3%, 7.1%, and 4.2% in PAN group, respectively. 30.9% of Bulky N group and 27.8% of PAN group showed MPR.TIs in each lymph node area are shown in the table. Conclusions: TI of PAN was extremely low in PAN group, especially in recent 2 trials (JCOG0405 and 1002) and when neoadjuvant chemotherapy was effective. These results suggest that advances in chemotherapy may make PAND unnecessary for GC with PAN metastasis in the future. Therapeutic value index in each lymph node area. PG area SP area PAN area Population N Bulky N group PAN group Bulky N group PAN group Bulky N group PAN group Overall 122 30.9 18.5 15.0 3.7 5.6 1.9 JCOG0001 36 25.0 12.5 18.2 12.5 4.5 6.3 JCOG0405 42 40.9 25.0 15.4 0.0 3.8 0.0 JCOG1002 44 26.9 16.7 16.2 0.0 4.5 0.0 MPR 36 19.0 33.3 4.8 0.0 4.8 0.0 PG, peri-gastric; SP, supra-pancreatic; PAN, para-aortic node; MPR, major pathological response.