中国人惰性前列腺癌临床筛选标准的比较

A head-to-head comparison of contemporary indolent prostate cancer screen protocols in Chinese

目的:依据临床病理资料对目前国际上使用最广泛的5种术前判断惰性前列腺癌的筛选标准进行探讨,以探索这些标准对国人惰性前列腺癌筛查的有效性。方法:回顾性分析2006年1月至2014年4月,314例于北京大学第一医院行前列腺穿刺活检确诊,并行根治性前列腺切除术的前列腺癌患者的临床病理及手术病理资料。以手术病理Gleason评分≤6、分期T2、癌灶总体积≤0.5 m L作为惰性前列腺癌的病理标准,对比国际上研究较多的5种判断惰性前列腺癌的临床标准,分别为Epstein标准(EC)、Sloan-Kettering纪念肿瘤中心标准(Memorial Sloan-Kettering Cancer Center,MSKCC)、国际前列腺癌研究-主动监测标准(Prostate Cancer Research International:Active Surveillance,PRIAS)、UCSF标准(University of California,San Francisco,UCSF)和UM标准(University of Miami,UM),对灵敏度、特异度以及受试者工作特性曲线下面积(areas under the receiver-operating curve,AUC)等指标进行统计分析,探讨各方法预测惰性前列腺癌的效能差异。结果:临床病理资料中,共有49例(16%)病例符合5种标准中的任意一种筛选标准,不考虑重复计算,其中符合Epstein标准者24例、符合MSKCC者33例、符合PRIAS者28例、符合UCSF者34例、符合UM标准者22例,仅8例(3%)符合全部标准。手术病理中,共24例(8%)符合惰性前列腺癌的病理标准。通过比较临床病理与手术病理资料,UCSF、MSKCC两标准在敏感性和特异性上均有良好的表现,AUC也大于其他3种标准,统计效能较高;Epstein、PRIAS两标准的阳性预测值较高,敏感性、特异性略低,统计效能中等;UM标准的敏感性、阳性预测值及AUC都较低,统计效能显著低于其他4种标准;符合全部5种筛选标准不能提高统计效能。结论:UCSF标准筛选惰性前列腺癌的预测效能高于其他标准,但其准确性仍较低,符合全部5种筛选标准并不能提高预测效能,此结果与国外大样本结果并不完全一致,因此需要进一步探讨更适合中国人的筛选模型。

Objective： To compare the diagnostic accuracy of five internationally used indolent prostate cancer screen protocols in Chinese prostate cancer patients. Methods： Retrospective analysis was made of the consecutive cohort of 314 patients,from Jan. 2006 to Apr. 2014,who had both prostate biopsy and radical prostatectomy in Peking University First Hospital. The Gleason score≤6,p T2,tumor volume≤0. 5 m L,margin negative and lymph nodes negative were defined as indolent prostate cancer. The predictive value of five indolent screen criteria including Epstein,Memorial Sloan-Kettering Cancer Center（ MSKCC）,Prostate Cancer Research International： Active Surveillance（ PRIAS）,University of California,San Francisco（ UCSF）,and University of Miami（ UM） were evaluated in Chinese prostate cancer patients. Measures of diagnostic accuracy and areas under the receiver-operating curve（ AUC） were calculated for each protocol and compared. Results： A total of 16%（ 49 cases） of the patients met the inclusion criteria of at least one protocol,including 24 cases in Epstein,33 cases in MSKCC,28 cases in PRIAS,34 cases in UCSF,and 22 cases in UM. Three percent were eligible for all the studied criteria.UCSF and MSKCC protocols had the highest sensitivity and specificity than the others. The Epstein and PRIAS protocols demonstrated acceptable positive predictive value,but the specificity and sensitivity were inefficient. The UM protocol was performed unsatisfiedly on sensitivity,positive predictive value and AUC. A strict limited protocol which contained all the five protocols could not improve the predictive accuracy. Conclusion： The UCSF protocol had better diagnostic accuracy than the others,but the results were not satisfied. A further investigation on indolent prostate cancer screening in Chinese patients is needed.