Background and Objectives: FLOT is a highly effective first-line treatment for metastatic gastric cancer and offers a favorable safety profile. Clinical studies investigating the FLOT regimen have reported varying outcomes depending on the infusion duration and have highlighted possible differences in complication rates and the efficacy of neoadjuvant therapy. The choice between 24 h or 48 h infusion durations for fluorouracil can be influenced by several factors, such as the patient’s overall health status, their tolerance to treatment, and the specific treatment protocol determined by the medical team. In this study, we aimed to evaluate the effects of different infusion durations (24 and 48 h) on clinical response, toxicity, and survival in patients with gastric and gastroesophageal junction (GEJ) adenocarcinoma. Materials and Methods: This retrospective multicenter study included 113 patients with gastric or gastroesophageal junction adenocarcinoma who received neoadjuvant FLOT chemotherapy (24 h infusion: n = 28; 48 h infusion: n = 85). Propensity score matching (PSM) was performed to balance baseline characteristics, yielding a matched cohort of 90 patients. The primary endpoints were the pathologic complete response (pCR) and toxicity. Secondary endpoints included disease-free survival (DFS) and overall survival (OS). Results: Significant baseline imbalances existed (cT stage p 0.99. The median DFS was 27.4 mo (24 h) vs. NR (48 h), p = 0.847. The median OS was 32.8 mo in both, p = 0.797. Multivariate analysis (baseline variables) indicates that infusion duration is not prognostic (DFS HR = 0.77, p = 0.453; OS HR = 0.72, p = 0.328). Power was ~10% for a 1% pCR difference. Conclusions: The 24 h infusion protocol was associated with similar outcomes to the 48 h protocol after PSM adjustment. However, residual confounding persists (cT stage p = 0.009 despite PSM), and the combination of this study’s retrospective design and severe underpowering (~10%) precludes definitive conclusions. As a result, the findings are hypothesis-generating.
Demir et al. (Tue,) studied this question.