三种临床评分方法对急性肺栓塞预测价值的比较

A comparison of the predictive values of three clinical scoring systems for suspected acute pulmonary embolism based on mnitidetector CT angiography

目的以CT肺血管造影（CTPA）为金标准，评价临床普遍应用的三种国外急性肺栓塞评分方法的预测效能，探讨适用于我国人群的评分方法。方法连续纳入570例（男321例，女249例，年龄18～75岁，平均55岁）行CTPA检查的临床疑似急性肺栓塞的住院或门诊患者。分别采用Wells、Geneva和改良Geneva评分法评价每例患者，并预测其急性肺栓塞发生的可能性。先由2名中年资影像学医师分别独立盲法评价CTPA，评价结果不一致时由1名高年资医师决定。应用受试者工作特征曲线分析评价三种评分方法的预测价值。结果570例中169例患者确诊为急性肺栓塞。三种临床评分方法两两一致性检验结果显示K值为0．269—0．374，P〈0．05；其中Geneva评分和改良Geneva评分的一致度较好。三种评分方法两两存在正相关关系，Geneva评分和改良Geneva评分之间的相关关系较密切。Wells评分、Geneva评分和改良Geneva评分对评估APE的评估的阳性预测值分别为83．8％、53．3％和61．3％，阴性预测值分别为85．0％、80．6％和80．0％。三者的受试者工作特征曲线下面积分别为：Wells评分0．823，Geneva评分0．677，改良Geneva评分0．661，三者比较，除了Geneva评分和改良Geneva评分相比差异无统计学意义（u=0．352，P〉0．05）外，其余两两之间的差异均有统计学意义（u=3．535，4．285，均P〈0．01）。结论三种临床评分方法均可以对急性肺栓塞作出较为准确的预测，但是Wells评分的预测价值最高，比较适合于我国人群。

Objective To explore whether acute pulmonary embolism （APE） can be quantitatively predicated early with 3 clinical scoring systems, with multidetector CT angiography （MDCTA） as the gold standard, and therefore to select a scoring system more suitable for the Chinese. Methods Five hundred and seventy consecutive inpatients with highly suspected APE underwent prospective MDCTA at the time of initial diagnosis. Three clinical predication scoring systems （Wells＇ , Geneva＇ and revised Geneva＇ ） were used to estimate APE in low, moderate and high probability groups. Two radiologists independently reviewed the MDCTA without any clinical information. When consensus could not be reached, a third radiologist with 20-years＇ experience was asked to make the final decision. The threshold value for the prediction of APE by the 3 scoring systems was measured by receiver-operating-characteristics （ ROC ） analysis. Results APE was identified in 169 of the 570 cases. Kappa analysis for the 3 scoring systems revealed a low level of agreement ： 0. 269 - 0. 374, P 〈 0.05. The result of the Geneva score was consistent with that of the revised Geneva score, between them there was an excellent correlation. The positive predictive values of Wells,Geneva, revised Geneva scores for APE were 83.8%, 53.3%, and 61.3% respectively, while the negative predictive values were 85.0%, 80. 6%, and 80. 0%, respectively. ROC analysis showed that the area under curve （AUC） of Wells, Geneva and revised Geneva score for APE was 0. 823 （95% CI：0. 710 -0. 976）, 0. 677 （95 % CI： 0. 646 - 0. 990）, and 0. 661 （95 % CI： 0. 631 - 0. 983 ）, respectively. The Wells score showed the best discriminatory ability as eompared to the other 2 scores. Conclusion The 3 scoring systems can be used for both inpatients and emergency cases, while the Wells Score may be more accurate for Chinese people for predicting APE.