Abstract
BACKGROUND: Reliable biomarkers of apatinib response in gastric cancer (GC) are lacking. We investigated the association between early presence of common adverse events (AEs) and clinical outcomes in metastatic GC patients. PATIENTS AND METHODS: We conducted a retrospective cohort study using data on 269 apatinib-treated GC patients in two clinical trials. AEs were assessed at baseline until 28 days after the last dose of apatinib. Clinical outcomes were compared between patients with and without hypertension (HTN), proteinuria, or hand and foot syndrome (HFS) in the first 4 weeks. Time-to-event variables were assessed using Kaplan-Meier methods and Cox proportional hazard regression models. Binary endpoints were assessed using logistic regression models. Landmark analyses were performed as sensitivity analyses. Predictive model was analyzed, and risk scores were calculated to predict overall survival. RESULTS: Presence of AEs in the first 4 weeks was associated with prolonged median overall survival (169 vs. 103 days, log-rank p = 0.0039; adjusted hazard ratio (HR) 0.64, 95% confidence interval [CI] 0.64-0.84, p = 0.001), prolonged median progression-free survival (86.5 vs. 62 days, log-rank p = 0.0309; adjusted HR 0.69, 95% CI 0.53-0.91, p = 0.007), and increased disease control rate (54.67 vs. 32.77%; adjusted odds ratio 2.67, p < 0.001). Results remained significant in landmark analyses. The onset of any single AE or any combinations of the AEs were all statistically significantly associated with prolonged OS, except for the presence of proteinuria. An AE-based prediction model and subsequently derived scoring system showed high calibration and discrimination in predicting overall survival. CONCLUSION: Presence of HTN, proteinuria, or HFS during the first cycle of apatinib treatment was a viable biomarker of antitumor efficacy in metastatic GC patients.