前列腺癌的危险因素及列线图预测模型的构建

Risk factors of prostate cancer and construction of nomogram prediction model

目的 基于前列腺癌的危险因素,构建前列腺癌危险因素列线图预测模型,为筛选出前列腺癌高危人群提供预测工具。方法 选取2019年9月-2021年9月在新疆医科大学第一附属医院泌尿外科经病理学诊断的182例前列腺癌患者和218例非前列腺癌患者,分析其一般临床资料筛选出前列腺癌的危险因素,通过危险因素构建列线图预测模型,采用一致性指数(C-index)、校准图和ROC曲线对该模型的一致性和预测准确性进行评估。结果 文化程度、户籍、癌症家族史、前列腺特异性抗原、游离前列腺抗原/结合前列腺抗原(F/T)比值、糖尿病、吸烟史、饮酒史是前列腺癌的影响因素(P<0.05)。基于上述8个危险因素构建列线图预测模型,C-index为0.915(95%CI 0.887~0.943)。校准图显示该模型的预测患病情况和实际患病情况之间具有较高的一致性。ROC曲线下面积为0.944(95%CI 0.924~0.964),显示本模型具有较准确的预测能力。结论 本研究构建的列线图预测模型有较准确的预测效能、临床价值高,对前列腺癌高危人群的筛选具有重要意义,有助于医师对患者做出较为准确的评估,可为前列腺癌患者个体化诊疗提供依据。

Objective Based on the risk factors of prostate cancer, a nomogram prediction model of prostate cancer risk factors was constructed to provide a prediction tool for screening high-risk groups of prostate cancer. Methods 182 patients with prostate cancer and 218 patients with non prostate cancer were diagnosed by urology pathology in the hospital from September 2019 to September 2021 were selected. The general clinical data were analyzed to screen the risk factors of prostate cancer, and the nomogram prediction model was constructed through the risk factors. The consistency and prediction accuracy of the model were evaluated by consistency index(C-index), calibration chart and ROC curve. Results The results showed that education, registered residence, family history of cancer, PSA, F/T ratio, diabetes, smoking history and drinking history were the influencing factors(P<0.05). The nomogram prediction model was constructed based on the above 8 risk factors, and the C-index was 0.915(95%CI 0.887 ~ 0.943). The calibration chart shows that the predicted disease situation of the model was in good agreement with the actual disease situation. The area under ROC curve was 0.944(95%CI 0.924 ~ 0.964), which showed that the model had more accurate prediction ability. Conclusion The nomogram prediction model constructed in this study has more accurate prediction efficiency and high clinical value. It is of great significance for screening of the high-risk groups with prostate cancer, it helps doctors to make a more accurate evaluation of the patients, and it provides a basis for individualized diagnosis and treatment of prostate cancer patients.