不同版本Briganti列线图对前列腺癌淋巴结转移预测效能的比较研究

The ability of different Briganti nomograms to predict lymph node metastasis in prostate cancer

目的比较不同版本Briganti列线图对前列腺癌患者淋巴结转移的临床预测效能。方法回顾性分析2012年10月至2021年4月由单术者完成的583例接受根治性前列腺切除术+盆腔淋巴结切除术患者的病例资料。中位年龄67(63,72)岁。中位体质量指数24.39(22.58,26.35)kg/m^(2)。中位前列腺特异性抗原(PSA)22(12,43)ng/ml。临床T分期T_(1)期65例,T_(2)期357例,T_(3)期140例,T_(4)期21例。穿刺病理的国际泌尿病理协会(ISUP)分组1组30例,2组109例,3组104例,4组160例,5组180例。穿刺阳性针数百分比50%(33%,83%)。采用Briganti列线图2006版、2012版和2017版预测患者的淋巴结转移情况。与2006版相比,2012版和2017版新增变量分别为穿刺阳性针数百分比,以及穿刺病理最高级别阳性针数百分比和较低级别阳性针数百分比。本研究583例中,适用于2006版、2012版和2017版列线图的患者分别为560例、513例和357例,以此作为差异验证队列;同时适用于3个版本列线图的患者为357例,以此作为通用验证队列。采用受试者工作操作特征(ROC)曲线的曲线下面积(AUC)、校准曲线图和临床决策曲线分析3个版本列线图的预测效能。结果差异验证队列中,2006版、2012版和2017版列线图的AUC值分别为0.738(95%CI0.690~0.785)0.765(95%CI0.717~0.814)和0.779(95%CI0.724~0.834),差异无统计学意义(P>0.05)。通用验证队列中,2006版、2012版和2017版列线图的AUC值分别为0.744(95%CI0.682~0.805)、0.759(95%CI0.700~0.818)和0.779(95%CI0.724~0.834),差异无统计学意义(P>0.05)。校准曲线图显示,在0~40%预测概率内,2012版和2017版的预测概率与实际风险具有较好的一致性。临床决策曲线分析结果显示,在0~33%的预测阈值内,2012版的临床获益大于另两个版本列线图。结论Briganti列线图适合预测前列腺癌患者盆腔淋巴结转移。2012版和2017版列线图具有较好的预测效能,临床使用时可根据能获得的预测变量选择版本。

Objective To compare the predictive efficacy of different versions of Briganti nomogram in predicting lymph node metastasis in Chinese patients with prostate cancer.Methods From October 2012 to April 2021,583 cases with prostate cancer who underwent radical prostatectomy and pelvic lymphadenectomy by a single surgeon were retrospectively collected.For all 583 patients,the median age was 67(63,72)years old,median BMI was 24.39(22.58,26.35)kg/m^(2),median PSA was 22(12,43)ng/ml.There were 65 cases,357 cases,140 cases and 21 cases with clinical stage T_(1),T_(2),T_(3),and T_(4).There were 30 cases,109 cases,104 cases,160 cases and 180 cases for ISUP 1 group,2 group,3 group,4 group and 5 group.The median percentage of positive biopsy cores was 50%(33%-83%).The validated nomograms were Brigantis 2006,2012 and 2017.Compared with the 2006 edition,the new variables in the 2012 edition and 2017 edition were the percentage of positive biopsy cores,the percentage of the highest grade positive biopsy cores and the percentage of the lower grade positive biopsy cores,respectively.The validation patients for the 2006,2012 and 2017 versions of nomogram were 560,513 and 357,respectively,which were used as the differential validation cohorts.A total of 357 patients were validated for all three versions of nomogram,which was considered as the general validation cohort.The area under the receiver operating characteristic(ROC)curve(AUC),calibration curve and clinical decision curve analysis were used to evaluate the predictive efficacy of the three versions of nomograms.Results In the differential validation cohort,the AUC values of the 2006,2012 and 2017 versions of the nomogram were 0.738(95%CI 0.690-0.785),0.765(95%CI0.717-0.814)and 0.779(95%CI 0.724-0.834),respectively.There was no significant difference in AUC values among versions(P>0.05).In the general validation cohort,the AUC values of the three versions of the nomogram were as follows 0.744(95%CI 0.682-0.805),0.759(95%CI 0.700-0.818)and 0.779(95%CI 0.724-0.834),respectively.There was no significant difference in AUC values among the three versions(P>0.05).The calibration curve showed that the prediction probability of 2012 and 2017 editions was in good agreement with the actual risk within the prediction probability of 0-40%.Analysis of the clinical decision curve showed that the clinical benefit of the 2012 version was greater than that of the other two versions in the prediction threshold of O-33%.Conclusion Briganti nomogram is suitable for predicting pelvic lymph node metastasis in Chinese patients with prostate cancer.The 2012 and 2017 versions of the nomogram have good predictive performance,and the versions can be selected according to the predictive variables that can be provided.