预测PI-RADs v2评分≤3分患者前列腺穿刺术后非临床有意义结果的列线图

Nomogram predicting non-significant outcomes after prostate biopsy in patients with PI-RADs v2 ≤3

目的构建能够预测PI-RADs v2评分≤3分患者前列腺穿刺术后非临床有意义结果的列线图,并内部验证其诊断效能。方法采用回顾性研究的方法,数据选自2012年1月至2018年7月于首都医科大学附属北京友谊医院行前列腺穿刺患者构成的数据库。统计分析相关临床指标,采用Logistic回归确定穿刺结果为非临床有意义的预测因素,并构建列线图。列线图的诊断性能由校准曲线及决策曲线来评估。结果共491例患者纳入研究。前列腺癌(PCa)的检出率为28. 5%(140/491),临床有意义PCa(CSPCa)的检出率为14. 5%(71/491),临床无意义PCa(CIPCa)的检出率为14. 1%(69/491),良性结果为71. 5%(351/491)。非临床有意义结果(non-CSPCa,包括CIPCa和Benign)的检出率为85. 5%(420/491)。基于年龄、前列腺特异性抗原密度(PSAD)、PI-RADs v2等指标构建列线图。校准曲线证实,列线图的预测一致度较好,若预测值在53%~72%之间,其具备高估风险的可能性。决策曲线显示,当阈值> 70%时,用该列线图决定是否穿刺可获得明显的临床获益。结论基于年龄、PI-RADs v2、PSAD等指标构建的列线图对于前列腺穿刺术后非临床有意义结果有着良好的预测价值。此列线图应用于临床决策可使临床获益。

Objective To develop and validate a nomogram to predict the probability of non-significant outcomes after prostate biopsy in men with PI-RADs v2≤3. Methods It’s a retrospective study. The database was constructed of consecutive men who received transrectal ultrasound-guided 24-core biopsy as well as pre-biopsy multi-parametric magnetic resonance imaging( mp-MRI) in our institution between January 2012 to July 2018. Logistic regression analysis was used to determine the predicting factors and establish the nomogram. The diagnostic efficacy of the nomogram was assessed by calibration curves and decision curves. Results 491 patients composed the database of this study. Detection rates of prostate cancer,clinically significant prostate cancer were 28. 5%( 140/491) and 14. 5%( 71/491). Detection rates of clinically insignificant prostate cancer and benign outcomes were 14. 1%( 69/491) and 71. 5%( 351/491). Detection rate of clinically insignificant outcomes( non-CSPCa) was 85. 5%( 420/491). The nomogram was made based on age,prostate specific antigen density( PSAD) and PI-RADs v2. Internally validated calibration curves showed that the model might overestimated the risk of absence of CSPCa when the threshold was between 53%-72%. Decision curves showed that a better clinical net benefit was met when the threshold was 70% for non-CSPCa. Conclusion The nomogram composed of age,PSAD and PI-RADs v2 had good predicting efficacy for non-significant outcomes. Its application in clinical decision will result in a net benefit.