年轻女性乳腺癌危险因素及风险预测模型构建

Risk Factors of Breast Cancer in Young Women and Construction of Risk Prediction Model

摘要: 目的：根据收住乳腺外科年轻女性(<40岁)患者的临床及影像资料评估其危险因素，并构建风险预测模型以期更早期、更准确地预测年轻乳腺癌的发生。方法：收集2020年1月~2024年12月收住乳腺外科的211例年轻女性患者的临床和影像资料，根据病理结果分为年轻乳腺癌组(n = 85)和年轻非乳腺癌组(n = 126)。采用单因素分析和二元Logistic回归分析，探讨年轻乳腺癌患者的相关危险因素并构建风险预测模型，绘制ROC曲线，并计算该模型的灵敏度和特异性，利用Bootstrap重复抽样法对该预测模型的准确性进行内部验证，判断该预测模型的可靠性。结果：单因素分析显示NLR、LMR、钙化特点、淋巴结状态、边缘、纤维蛋白原、睾酮、促黄体生成素均有统计学意义(P < 0.05)。二元Logistic回归分析显示NLR、LMR、钙化特点、淋巴结状态、边缘、促黄体生成素是年轻乳腺癌的独立危险因素(均P < 0.05)。此风险预测模型AUC为0.991 (P < 0.001)，特异性为0.968，敏感性为0.976。内部验证校正后C-index指数为0.982。结论：此风险预测模型能较精准地预测年轻乳腺癌是否发生。

Abstract: Objective: To evaluate the risk factors of young female patients (under 40 years old) admitted to the breast surgery department based on their clinical and imaging data, and to construct a risk prediction model in order to predict the occurrence of young breast cancer earlier and more accurately. Methods: Clinical and imaging data of 211 young female patients admitted to the breast surgery department from January 2020 to December 2024 were collected. These patients were divided into the young breast cancer group (n = 85) and the young non-breast cancer group (n = 126) according to the pathological results. Univariate analysis and binary Logistic regression analysis were used to explore the related risk factors of young breast cancer patients and construct a risk prediction model. ROC curves were drawn, and the sensitivity and specificity of the model were calculated. The accuracy of this prediction model was internally validated using the Bootstrap repeated sampling method, and the reliability of the prediction model was judged. Results: Univariate analysis showed that NLR, LMR, calcification characteristics, lymph node status, margin, fibrinogen, testosterone, and luteinizing hormone were statistically significant (all P < 0.05). Binary Logistic regression analysis showed that NLR, LMR, calcification characteristics, lymph node status, margin, and luteinizing hormone were independent risk factors for young breast cancer (all P < 0.05). The AUC of this risk prediction model was 0.991 (P < 0.001), with a specificity of 0.968 and a sensitivity of 0.976. The corrected C-index of internal validation was 0.982. Conclusion: This risk prediction model can accurately predict whether young breast cancer will occur.