SHA.LIN、S.T.O.N.E评分系统和Guy＇s分级法预测经皮肾镜取石术结石清除率准确性的比较研究

Evaluation and comparison of SHA.LIN, S.T.O.N.E nephrolithometry scoring system and Guy＇s stone score in predicting the accuracy of the percutaneous nephrolithotomy（PCNL） outcomes

目的 比较SHA.LIN、S.T.O.N.E肾结石评分系统和Guy＇s分级法预测经皮肾镜取石术（PCNL）后结石清除率及术后并发症的准确性.方法 回顾性分析2011年6月至2015年6月收治的123例行PCNL的肾结石患者的临床资料,男71例,女52例.年龄24 ～78岁,平均54岁.体质指数16.4 ～33.2 kg/m2,平均24.2 kg/m2.术前血肌酐38 ～ 331 μmol/L,平均82 μmol/L.术前尿常规检查WBC阳性96例（78.0％）,尿细菌培养阳性33例（26.8％）.美国麻醉师协会（American Society of Anesthesiology,ASA）评分：1分40例（32.5％）,2分70例（56.9％）,3分13例（10.6％）.左肾结石69例（56.1％）,右肾结石54例（43.9％）.目标穿刺肾盏：上盏47例（38.2％）,中盏76例（61.8％）.由同一名医师利用术前泌尿系CT＋三维重建,分别根据SHA.LIN、S.T.O.N.E肾结石评分系统和Guy＇s分级法对患者术前结石情况进行评分、分级,分析SHA.LIN、S.T.O.N.E评分和Guy＇s分级与结石清除情况、手术时间、估计失血量、住院天数、术后并发症等临床数据的关系.绘制受试者工作特征曲线（ROC）并计算曲线下面积（AUC）检测SHA.LIN、S.T.O.N.E评分系统和Guy＇s分级法预测结石清除率的敏感性和特异性.结果 本组123例的手术时间50～160min,平均（87.4±25.8） min.估计失血量40 ～ 850 ml,平均（166.4±186.9） ml.总住院时间为7～30 d,平均（13.5±3.9）d;术后住院时间为4 ～25 d,平均（9.2±3.4）d.结石清除率为76.4％ （94/123）,33例（26.8％）出现术后并发症.结石清除组和残留组的SHA.LIN评分分别为8.2分和11.8分,S.T.O.N.E.评分分别为7.3分和8.6分,Guy＇s分级分别为2.4分和3.0分,差异均有统计学意义（P＜0.05）.SHA.LIN评分、S.T.O.N.E.评分和Guy＇s分级与术后结石清除状态及术后并发症的发生情况均相关,差异均有统计学意义（P＜0.05）.SHA.LIN评分与手术时间（P＜0.05）、估计失血量（P＜0.05）相关,S.T.O.N.E.评分和Guy＇s分级与手术时间（P＜0.05）相关.SHA.LIN、S.T.O.N.E评分系统和Guy＇s分级法ROC曲线的AUC分别为0.924（95％ CI0.874～0.975）、0.795 （95％ CI0.698 ～0.891）和0.701 （95％ CI0.592 ～0.811）.S.T.O.N.E评分与Guy＇s分级的AUC相比差异无统计学意义（P=0.124）.SHA.LIN评分分别与S.T.O.N.E评分和Guy＇s分级法的AUC相比差异均有统计学意义（P＜0.05）.结论 SHA.LIN、S.T.O.N.E评分系统和Guy＇s分级法均能较准确预测PCNL的术后结石清除状态及术后并发症的发生,SHA.LIN评分与S.T.O.N.E评分及Guy＇s分级相比预测能力更准确.

Objective To compared the accuracy of the SHA.LIN,S.T.O.N.E.nephrolithometry scoring system and Guy＇s stone score in the same cohort and to determine which was most predictive of percutaneous nephrolithotomy （PCNL） outcomes.Methods A retrospective analysis was conducted of clinical data of 123 patients with nephrolithiasis undergoing PCNL from June 2011 and June 2015 in two hospitals.The SHA.LIN nephrolithometry score,the S.T.O.N.E.score and the Guy＇s stone score based on preoperative computerized tomography images were calculated.A single observer reviewed all images and assigned scores.Logistic and linear regression analysis were performed and Receiver operating characteristic （ROC） curves were drawn to determine the most predictive scoring system.Results The stone free rate was 76.4% （94/123） and postoperative complications occurred in 33 （26.8%） cases.In stone-free patients vs.those with residual stones the mean SHA.LIN nephrolithometry score was 8.2 vs.11.8,the mean S.T.O.N.E.score was 7.3 vs.8.6 and the mean Guy＇s stone score was 2.4 vs.3.0 （each P =0.000）.Logistic regression revealed that SHA.LIN nephrolithometry score,S.T.O.N.E.score and the Guy＇s stone score were significantly associated with stone-free status （each P =0.000） and postoperative complications （P 〈0.05）.The SHA.LIN nephrolithometry score was associated with operation time and estimated blood loss （P =0.000 and 0.011）.S.T.O.N.E.score and Guy＇s stone score were associated with operation time （P =0.000 and 0.006,respectively）.All scoring systems did not predict length of stay.There was significant difference in the areas under the curves（AUC） for the SHA.LIN vs.S.T.O.N.E.scoring system [0.924 （95 % CI 0.874-0.975） vs.0.795 （95 % CI 0.698-0.891）,P =0.001] and for the SHA.LIN nephrolithometry score vs Guy＇s stone score [0.924 （95% CI 0.874-0.975） vs.0.701 （95% CI 0.592-0.811）,P 〈 0.000].Conclusion All scoring systems have comparable accuracies and the predictive accuracy of the SHA.LIN scoring system is more precise than the S.T.O.N.E.score and the Guy＇s score in predicting post-PCNL stonefree status and postoperative complications.