Identifying potential beneficiaries of neoadjuvant therapy in borrmann type III and IV gastric cancer: A treatment heterogeneity analysis using causal survival forests.

[BACKGROUND] Neoadjuvant chemotherapy (NAC) is a standard treatment for locally advanced gastric cancer, but its efficacy for Scirrhous-type Gastric Cancer (SGC) remains uncertain. This study evaluated the effectiveness of NAC and identified characteristics of potential beneficiaries within this patient population.

[METHODS] A retrospective single-center analysis included patients with large SGC who underwent either NAC followed by surgery (NAC group) or upfront surgery (SUR group). Kaplan-Meier methods and Cox regression were used for survival analysis. Causal survival forest (CSF) models assessed individual treatment effects (ITE) and treatment heterogeneity. Survival differences were validated in potential beneficiaries based on CSF-derived criteria.

[RESULTS] The study included 60 patients in SUR group and 79 in NAC group. The 5-year overall survival (OS) rates were 39.2 % for the SUR group and 40.9 % for the NAC group (p = 0.53). CSF analysis revealed heterogeneity in ITE, with BMI, age, and distal gastric tumors associated with ITE. In the subgroup of patients younger than 58 years with a BMI <25.9 kg/m (benefit group), the 5-year OS rate was significantly higher in the NAC group (47.7 % vs. 22.9 %; HR = 0.428, 95 % CI: 0.205-0.892; p = 0.023). Further restricting the analysis to distal gastric cancer reduced the HR for NAC (HR = 0.299, 95 % CI: 0.094-0.959; p = 0.033).

[CONCLUSION] NAC efficacy for large SGCs shows significant heterogeneity. Patients younger than 58 years, with a BMI <25.9 kg/m and distal tumors may derive survival benefits from NAC.