【摘要】 目的 基于人工神经网络挖掘技术，分析与前列腺癌患者术后早期尿控恢复相关的危险因素并建立预后模型。方法 回顾性选取2018年4月至2022年4月收治于我院的124例前列腺癌患者，按照2∶1的比例随机分为建模组(83例)和验证组(41例),建模组患者用于筛选变量并建立预测模型，根据患者术后是否出现尿失禁将建模组患者分为尿失禁组和无尿失禁组，比较两组患者一般资料，分析影响患者术后尿失禁的危险因素，并构建人工神经网络模型；使用建模组数据集绘制受试工作者特征(ROC)曲线和校准曲线评价模型的区分度和准确度，用验证组数据集对模型进行外部评价。结果 结果显示保留膀胱颈、保留前列腺尿道、保留神经血管束、后方筋膜重建、术后规律提肛训练是患者术后尿失禁的保护因素(P<0.05),建模组和验证组ROC曲线和校准曲线均显示模型区分度和准确度良好，临床决策曲线显示模型临床应用价值较高。结论 前列腺癌患者保留膀胱颈、保留前列腺尿道、保留神经血管束、后方筋膜重建并进行术后规律提肛训练，可以有效提高尿控恢复预后。更多还原

【Abstract】 Objective To analyze the risk factors associated with early postoperative urinary control recovery in prostate cancer patients and establish a prognostic model based on artificial neural network mining technology. Methods Clinical data of 124 patients with prostate cancer who were admitted to our hospital from April 2018 to April 2022 were retrospectively analyzed in the study. The patients were randomly divided into the modeling group(83 cases) and the validation group(41 cases) according to the ratio of 2∶1. Patients in the modeling group were used to screen variables and establish predictive models. Patients in the modeling group was further divided into urinary incontinence group and non-urinary incontinence group according to the occurrence of urinary incontinence after surgery. The general data of the two groups were compared. The risk factors of postoperative urinary incontinence were analyzed by multivariate logistic regression analysis, and then artificial neural network model was constructed. The receiver operating characteristic(ROC) curve and calibration curve were drawn to evaluate the discrimination and accuracy of the model, and the validation group dataset was used to evaluate the model externally. Results Multivariate logistic regression analysis showed that bladder neck preservation, prostatic urethra preservation, neurovascular bundle preservation, posterior fascia reconstruction, and regular anal levator training were the protective factors of postoperative urinary incontinence(P< 0.05). The ROC curve and calibration curve of the modeling group and the validation group revealed good discrimination and accuracy of the model. The clinical decision curve indicated that the model had high value in clinical application. Conclusion In patients with prostate cancer, bladder neck preservation, prostatic urethra preservation, neurovascular bundle preservation, posterior fascia reconstruction and regular anal levator training can effectively improve the prognosis of urinary control recovery.更多还原