Abstract
PURPOSE: This research aimed to identify significant prognostic factors that interact with the treatment effect of capecitabine and oxaliplatin (CapeOX), based on individual patient data from the CLASSIC study (NCT00411229), which evaluated the efficacy of adjuvant chemotherapy for gastric cancer. METHODS: Stepwise variable selection of prognostic factors was performed using the Cox proportional hazards model, with patient data from 519 CapeOX-treated and 514 untreated patients. RESULTS: For all-cause mortality, older women (≥ 55 years) with a serum albumin level ≥ 4.0 g/dL were identified as significant prognostic factors interacting with CapeOX treatment, and unexpectedly, the treatment was associated with poor outcomes in this group. The prognostic significance of serum albumin levels was primarily attributed to the particularly poor survival outcomes observed in untreated patients with serum albumin < 4.0 g/dL. Tumor stage and lymph node status were prognostic factors that interacted with treatment for disease-free survival (DFS). The results showed that in patients with tumor stage (T) ≥ T3 and lymph node metastasis (N) < N2, improvement in DFS with CapeOX was not significant. CONCLUSION: As this was a post-hoc analysis, the results should be interpreted as hypothesis-generating rather than definitive. Nevertheless, the findings suggest the need for a more detailed consideration of patient baseline characteristics when determining the adjuvant chemotherapy strategy for gastric cancer.