在磁共振阴性的患者中经直肠超声引导下前列腺穿刺活检术诊断前列腺癌的预测模型

Development and Validation of a Predictive Model for Determining Prostate Cancer in Men with Negative Magnetic Resonance Imaging after Transrectal Ultrasound-guided Prostate Biopsy

目的在磁共振(MRI)阴性的患者中,经直肠超声引导下前列腺穿刺活检术诊断前列腺癌的预测因素并不清楚。我们旨在建立和验证一个模型,以预测在MRI阴性患者中,经直肠超声引导下前列腺穿刺活检术诊断前列腺癌的概率。方法在2014年1月1日至2017年12月31日期间,728例MRI阴性,且在我中心接受了经直肠超声引导下穿刺活检术的患者作为建模集。记录患者的临床病理资料。采用Lasso回归进行数据降维和特征选择,然后采用多因素Logistic回归建立预测模型。将该模型在2018年1月1日至2020年6月30日的334名患者中进行验证。在区分度、一致性和临床适用性方面,对预测模型的性能进行了评估。结果该模型的预测因子包括年龄、身高体重指数(BMI)、前列腺穿刺活检史、下尿路症状(LUTS)和前列腺特异性抗原密度(PSAD),其中PSAD是最强的预测因子。该模型区分度为AUROC=0.844 (0.788, 0.901, P <0.001),一致性检验为unreliability test P=0.442;Hosmer-Lemeshow P=0.865。决策曲线分析(DCA)也表明该模型在临床上是有用的。结论本研究提供了一种良好的预测模型,可帮助对患者进行前列腺穿刺活检术前的危险分层,这有助于指导MRI阴性患者诊断策略的制定,例如在低风险患者中,避免进行不必要的活检和潜在的过度治疗。对于风险较高的患者可以采取多次穿刺,甚至饱和穿刺,以避免漏诊。

Objective The interpretation of negative magnetic resonance imaging(MRI) screening results for prostate cancer(PCa) is debatable and poses a clinical dilemma for urologists. No nomograms have been developed to predict PCa in such populations. In this study, we aimed to develop and validate a model for predicting the probability of PCa in men with negative MRI results after transrectal ultrasound-guided systematic prostate biopsy. Methods The development cohort consisted of 728 patients with negative MRI results who underwent subsequent prostate biopsy at our center between 1 January 2014 and 31 December 2017. The patients’ clinicopathologic data were recorded. The Lasso regression was used for data dimension reduction and feature selection, then multi-variable logistic regression analysis was used to develop the prediction model. The model was validated in an independent cohort of 334 consecutive patients from 1 January 2018 and 30 June 2020. The performance of the predictive model was assessed with respect to discrimination, calibration, and decision curve analysis. Results The predictors incorporated in this model included age, BMI, history of previous negative prostate biopsy, prostate specific antigen density(PSAD), and lower urinary tract symptoms, with PSAD being the strongest predictor.The model showed good discrimination with an area under the receiver operating characteristic curve of 0.844(95%confidence interval, 0.788-0.901) and good calibration(unreliability test, P=0.442, Hosmer-Lemeshow P=0.865). Decision curve analysis demonstrated that the model was clinically useful. Conclusion This study presents a good nomogram that can aid pre-biopsy risk stratification for the detection of PCa, and that may help inform biopsy decisions in patients with negative MRI results. Surveillance, rather than biopsy, may be appropriate for patients at lower risk, who could therefore avoid an unnecessary biopsy and potential overtreatment. For patients at high risk, repeated biopsy or saturation biopsy may be appropriate for avoiding missed diagnosis.