尘肺病诊断读片差异性分析

目的]通过读片统计尘肺病诊断结果,并对其差异性进行分析,为提高诊断的准确性提供依据。方法]邀请尘肺病诊断专家分A、B两组对163张已确诊为尘肺病病人的胸片进行双盲读片,对两组诊断结果不一致的胸片,从A、B组各抽取2名医生,并另邀请2名国家级专家进行最终集体读片,记录读片结果,进行数据处理与统计学分析。结果]A、B两组专家对小阴影形态大小的判断一致性最低(Kappa值=0.422,P<0.01);小阴影密集度和小阴影聚集判断一致性稍好(Kappa值=0.518、0.469,P<0.01);大阴影判断一致性最好(Kappa值=0.716,P<0.01)。在胸片的6个分区中,两肺中区是一致性最低的区域(Kappa值=0.362、0.411,P<0.01),两肺上区一致性最高(Kappa值=0.567、0.525,P<0.01);两组专家对期别的判断和把握较好,两组间符合率为75.46%,Kappa值为0.708(P<0.01)。对结果不一致的胸片,A、B两组医生与专家组最终集体读片的诊断分期结果符合率分别为53.65%、41.46%,Kappa值分别为0.393、0.221(P<0.01),A组专家与专家组集体读片结果更为接近。结论]尘肺病诊断读片客观上存在一定差异,小阴影总体密集度和各肺区的分布是关键判断依据,正确判断局部的细节是做出精确诊断的基础,资历深的专家在总体的把握上比较稳健,能根据丰富的读片经验,综合各种小阴影的性质给出正确的尘肺病分期结论。

Deviation analysis of reading chest X-ray film for pneumoconiosis diagnosis

Objective The diagnostic results of pneumoconiosis made by groups of experts based on reading chest X-ray films was compared and the major difference was analyzed, to improve the accuracy of the diagnosis of pneumoconiosis. Methods Two groups (A and B) of experts were invited to read 163 chest X-ray films independently. If the diagnostic result was different for the same chest X-ray films, the expert group (Composed of 2 senior experts and 2 experts selected from group A and group B respectively), were invited to read these films again. Then the reading results were recorded and analyzed. Results The comparison study showed that the consistence of reading results, from lower to higher, were for evaluation of the size of small shadow, the density, the aggregation and big shadow (Kappa=0.422, 0.518, 0.469, 0.716, respectively). It seemed that the experts had lowest consistence in identification of change in middle areas (Kappa=0.362, 0.411, respectively)and highest consistence in identification of change in upper areas (Kappa=0.567, 0.525, respectively) in 6 divisions of chest X-ray films. Two group of experts had the total consistent rate of 75.46% with the Kappa of 0.708 (P < 0.01). Compared with the results made by expert group, consistent rate of group A was 53.65% with Kappa of 0.393 (P < 0.01), and group B was 41.46% with Kappa of 0.221 (P < 0.01). The reading results of group A were closer to expert group. Conclusion There were always the deviation in diagnosis of pneumoconiosis based on reading chest X-ray films. The critical non-consistence was the evaluation of the density of small shadows and its distribution area. Generally, the senior experts can made the best diagnosis based on their rich experience on evaluation of various small shadows.