Proposal of a Scoring Scale to Estimate Risk of the Discontinuation of S-1 Adjuvant Monotherapy in Patients with Stage II to III Gastric Cancer: A Multi-Institutional Dataset Analysis

Discontinuation of postoperative S-1 adjuvant monotherapy is a frequent problem in the management of patients with gastric cancer. A total of 355 stage II/III gastric cancer patients who underwent gastrectomy and adjuvant S-1 were retrospectively analyzed using a multicenter dataset. We randomly assigned patients into either discovery or validation cohort in a 2:1 ratio. In the discovery cohort, 29 parameters were assessed as candidate factors to predict discontinuation of S-1 adjuvant within 6 months. A scoring system was designed using independent risk factors identified by the multivariate analysis. Reproducibility was tested in the validation cohort. Overall, 92 patients (25.9%) discontinued the treatment within 6 months because of adverse effects. Age, preoperative urea nitrogen (UN) and the preoperative albumin-to-bilirubin index (ALBI) showed the highest area under the curve (AUC) for the discontinuation of S-1 adjuvant within 6 months in each category: body status, blood tests and indices. In the multivariate analysis, age ≥ 64 years, preoperative UN ≥ 15.2 mg/dl and preoperative ALBI ≥ −0.265 were identified as independent risk factors. A scoring scale consisting of these three factors was developed for the prediction of drug discontinuation and demonstrated a greater AUC (0.728) than that of each of the three constituents. The time to treatment discontinuation decreased incrementally as the risk score increased. The reproducible findings were confirmed in the validation cohort. We identified risk factors and developed a scoring scale to predict S-1 adjuvant monotherapy discontinuation in patients with stage II/III gastric cancer.