基于CT薄层影像特征的肺结节良恶性评估

Evaluation of Benign and Malignant Pulmonary Nodules Based on Thin-layer CT Imaging Features

目的制订肺结节分级评估系统(PNI-GARS),评价其在肺结节良恶性评估中的应用价值。资料与方法回顾性分析180例患者共200个肺结节,采用PNI-GARS进行分级评定,计算PNI-GARS分级的准确度、敏感度、特异度、阴性预测值、阳性预测值。结果Ⅰ、Ⅱ级的阴性预测值分别为100.0%、97.2%,Ⅲa、Ⅲb、Ⅲc、Ⅳ级的阳性预测值分别为53.8%、85.7%、93.3%、100.0%。将Ⅰ、Ⅱ级归为阴性结节,Ⅲa级及以上归为阳性结节,PNI-GARS分级的准确度、敏感度、特异度、总阳性预测值分别为89.5%、99.3%、68.3%、87.2%。将Ⅰ、Ⅱ、Ⅲa级归为阴性结节,Ⅲb～Ⅳ级归为阳性结节,PNI-GARS分级的准确度、敏感度、特异度、总阳性预测值分别为88.5%、89.1%、87.3%、93.8%。排除Ⅲa级,将Ⅰ、Ⅱ级归为阴性结节,Ⅲb～Ⅳ级归为阳性结节,PNI-GARS分级的准确度、敏感度、特异度、总阳性预测值分别为94.8%、99.2%、84.3%、93.8%。结论 PNIGARS能有效地分类CT筛查出的肺结节,在肺结节的良恶性评估中有较高的应用价值。

Purpose To establish a pulmonary node imaging-grading and reporting system (PNI-GARS) to evaluate its application value in assessing benign and malignant pulmonary nodules. Materials and Methods A total of 200 pulmonary nodules of 180 patients were retrospectively analyzed. PNI-GARS was adopted for grading assessment, the accuracy, sensitivity, specificity, negative predictive value and positive predictive value of PNI-GARS grading were calculated. Results The negative predictive values of grade Ⅰ and Ⅱ were 100.0% and 97.2%, respectively;and the positive predictive values of grades Ⅲa,Ⅲb,Ⅲc and Ⅳ were 53.8%, 85.7%, 93.3% and 100.0%, respectively. Grade Ⅰ and Ⅱ were classified as negative nodules, grade Ⅲa and above as positive nodules, with the accuracy, sensitivity, specificity, and total positive predictive value of PNI-GARS grading being 89.5%, 99.3%, 68.3% and 87.2%, respectively. Grade Ⅰ,Ⅱ and Ⅲa were classified as negative nodules, grade Ⅲb to Ⅳ as positive nodules, with the accuracy, sensitivity, specificity and total positive predictive value of PNI-GARS grading being 88.5%, 89.1%, 87.3% and 93.8%, respectively. Excluding grade Ⅲa, grade I and Ⅱ were classified as negative nodules, and grades Ⅲb to Ⅳ as positive nodules, with the accuracy, sensitivity, specificity and total positive predictive value of PNI-GARS grading being 94.8%, 99.2%, 84.3% and 93.8%, respectively. Conclusion PNI-GARS can effectively classify pulmonary nodules screened out by CT, featuring a relatively high application value in evaluation of benign and malignant pulmonary nodules.