Fluoropyrimidine plus platinum chemotherapy remains the standard first line treatment for gastric cancer (GC). Guidelines exist for the clinical interpretation of four DPYD genotypes related to severe fluoropyrimidine toxicity within European populations. However, the frequency of these single nucleotide polymorphisms (SNPs) in the Latin American population is low (< 0.7%). No guidelines have been development for platinum. Herein, we present association between clinical factors and common SNPs in the development of grade 3–4 toxicity. Retrospectively, 224 clinical records of GC patient were screened, of which 93 patients were incorporated into the study. Eleven SNPs with minor allelic frequency above 5% in GSTP1, ERCC2, ERCC1, TP53, UMPS, SHMT1, MTHFR, ABCC2 and DPYD were assessed. Association between patient clinical characteristics and toxicity was estimated using logistic regression models and classification algorithms. Reported grade ≤ 2 and 3–4 toxicities were 64.6% (61/93) and 34.4% (32/93) respectively. Selected DPYD SNPs were associated with higher toxicity (rs1801265; OR = 4.20; 95% CI = 1.70–10.95, p = 0.002), while others displayed a trend towards lower toxicity (rs1801159; OR = 0.45; 95% CI = 0.19–1.08; p = 0.071). Combination of paired SNPs demonstrated significant associations in DPYD (rs1801265), UMPS (rs1801019), ABCC2 (rs717620) and SHMT1 (rs1979277). Using multivariate logistic regression that combined age, sex, peri-operative chemotherapy, 5-FU regimen, the binary combination of the SNPs DPYD (rs1801265) + ABCC2 (rs717620), and DPYD (rs1801159) displayed the best predictive performance. A nomogram was constructed to assess the risk of developing overall toxicity. Pending further validation, this model could predict chemotherapy associated toxicity and improve GC patient quality of life.

中文翻译：

---

结合单核苷酸多态性和临床参数预测与氟嘧啶和铂类化疗相关的胃癌临床相关毒性的病例对照研究

氟嘧啶加铂类化疗仍然是胃癌 (GC) 的标准一线治疗方法。存在关于欧洲人群中与严重氟嘧啶毒性相关的四种 DPYD 基因型的临床解释指南。然而，这些单核苷酸多态性 (SNP) 在拉丁美洲人群中的频率很低 (< 0.7%)。尚未制定铂金指南。在此，我们介绍了临床因素与常见 SNP 在 3-4 级毒性发展中的关联。回顾性筛选了 224 份 GC 患者的临床记录，其中 93 名患者纳入研究。评估了 GSTP1、ERCC2、ERCC1、TP53、UMPS、SHMT1、MTHFR、ABCC2 和 DPYD 中具有 5% 以上次要等位基因频率的 11 个 SNP。使用逻辑回归模型和分类算法估计患者临床特征和毒性之间的关联。报告的 ≤ 2 级和 3-4 级毒性分别为 64.6% (61/93) 和 34.4% (32/93)。选定的 DPYD SNP 与较高的毒性相关（rs1801265；OR = 4.20；95% CI = 1.70–10.95，p = 0.002），而其他的则显示出较低毒性的趋势（rs1801159；OR = 0.45；95% CI = 0.0819 ; p = 0.071)。配对 SNP 的组合在 DPYD (rs1801265)、UMPS (rs1801019)、ABCC2 (rs717620) 和 SHMT1 (rs1979277) 中显示出显着关联。使用结合年龄、性别、围手术期化疗、5-FU 方案的多变量逻辑回归，SNP DPYD (rs1801265) + ABCC2 (rs717620) 和 DPYD (rs1801159) 的二元组合显示出最好的预测性能。构建列线图以评估发生总体毒性的风险。有待进一步验证，该模型可以预测化疗相关毒性并改善 GC 患者的生活质量。