Primary analysis results of a phase II trial of laparoscopic/robotic total gastrectomy with spleen-preserving splenic hilar dissection for locally advanced proximal gastric cancer involving the greater curvature: JCOG1809.

363 Background: Advanced proximal gastric cancer (APGC) involving the greater curvature sometimes metastasized to the splenic hilar lymph node (No.10 LN), so total gastrectomy with splenectomy has been the standard treatment aiming complete removal of No.10 LNs; however, the high incidence of relevant complications (e.g. pancreatic fistula) has been a serious concern. Laparoscopic or robotic-assisted spleen preserving splenic hilar dissection (LRSHD) has the potential to reduce complications and become a new treatment option. We conducted a single-arm phase II trial to evaluate the safety and feasibility of LRSHD. Methods: The main eligibility criteria were clinical T2-4a resectable APGC involving the greater curvature without obvious No.10 LN metastasis. Protocol treatment was LRSHD (D2 + No.10 LN dissection) in which the spleen is preserved with skeletonizing the splenic vessels. Quality control of surgery was assured by surgeon’s qualification and central peer review of intraoperative photos. The primary endpoint was the proportion of postoperative pancreatic fistula and/or intra-abdominal abscess formation with Clavien-Dindo grade III or higher (within 30 days after surgery) for all resected patients. The sample size was set as 85 patients to obtain 80% power, with the hypothesis that the primary endpoint would have an expected value of 7% and a threshold value of 16% in one-sided alpha of 0.1. Results: Between August 2019 and December 2024, 85 eligible patients (T2/T3/T4a = 32/38/15; laparoscopic/robotic = 29/56) were enrolled from 24 centers. Four did not undergo resection due to P1 and/or CY1. Resection was performed in 81, and protocol treatment was completed in 75 since 4 palliative or 2 other organ combined resection resulted in incomplete protocol treatment. In all 81 resected patients, the median operation time was 379 min (IQR 323–448). The median blood loss was 30 ml (IQR 15–94). The median total number of retrieved LNs was 55, and the median number of retrieved No.10 LNs was 3. No.10 LN metastasis was observed in 12.5%. Overall grade III or higher complications (within 30 days) occurred in 9.9%. Pancreatic fistula and/or intra-abdominal abscess (primary endpoint) occurred in 1.2% (80% CI: 0.1-4.7%), which was significantly lower than prespecified threshold of 16% (one-sided p < 0.0001). Conclusions: This trial confirmed the safety of LRSHD. After further analysis of survival outcomes, LRSHD is expected to be established as one of the standard treatments for APGC involving the greater curvature. Clinical trial information: UMIN000037580 .