绝经前HR+乳腺癌患者内分泌治疗相关脂肪肝预测模型的研究

目的 探索绝经前HR+乳腺癌患者内分泌治疗相关脂肪肝的影响因素。方法 回顾性收集研究对象基本信息及脂肪肝发生信息，构建列线图并进行评价。结果 294例患者纳入最后的分析，其中服用SERM类药物和服用AI类药物者分别为247例(84.0%)和47例(16.0%)。Cox单因素分析发现BMI和确诊年龄与脂肪肝的发生相关。Cox多因素分析（逐步回归法）发现使用SERM类药物(P=0.002, 95%CI:1.27~2.78)、超重(P<0.001, 95%CI:1.40~2.82)或肥胖(P<0.001, 95%CI:2.08-5.57)及确诊年龄≥46.5岁(P=0.003, 95%CI:1.20~2.45)与脂肪肝发生相关。构建列线图并进行评价，一致性指数C-index为0.668(95%CI: 0.628~0.717)，列线图预测的率接近校准曲线，模型的准确度及符合度均较好。结论 绝经前HR+乳腺癌患者中，脂肪肝的发生与使用SERM类药物、超重或肥胖、确诊年龄≥46.5岁有关。列线图预测性能较好。

Predictive model of endocrine therapy related fatty liver in premenopausal women with HR+ breast cancer

To explore the influencing factors of endocrine therapy related fatty liver in premenopausal women with HR+ breast cancer, and to develop an effective nomogram for the prediction of fatty liver in this population. Methods The basic information of the study subjects and the information of fatty liver occurrence were collected retrospectively, and a nomogram was constructed and evaluated. Results A total of total of 294 patients were included in the final analysis. Among them, 247(84.0%) were taking SERM drugs and 47(16.0%) were taking AI drugs. Cox univariate analysis found that BMI and age at diagnosis were associated with the development of fatty liver. Cox multivariate analysis(stepwise regression) found that using SERM drugs(P=0.002, 95%CI: 1.27-2.78), being overweight(P<0.001, 95%CI: 1.40-2.82) or obese(P<0.001, 95%CI: 2.08-5.57) and the age of diagnosis ≥46.5 years old(P=0.003, 95%CI: 1.20-2.45) were related to fatty liver. A nomogram was constructed and evaluated. The C-index was 0.668(95%CI: 0.628-0.717). The prediction rate of the nomogram was close to the calibration curve. The accuracy and conformity of the model were good. Conclusion In premenopausal women with HR+ breast cancer, the occurrence of fatty liver is associated with the use of SERM drugs, overweight or obesity, and the age of diagnosis is more than 46.5 years old. The nomogram has good predictive performance.