基于胸部CT影像特征構(gòu)建孤立性肺結(jié)節(jié)診斷模型

打開文本圖片集

Diagnosis model isolated pulmonary nodules constructed based on the features chest CT images Yan Xiaoyu， Wang Peipei，Ji Xiang. ， ，999 ， ，

【Abstract】ObjectiveTo establish a diagnostic model based on chest CT image features patients with isolated pulmonary nodules.MethodsOne hundred and twentypatientswith solitary pulmonary nodules were selected from March 2022 to March 2024 were divided into benign nodules group scriptstyle（n=65 ）and malignant nodules group （ n =55）according to pathological findings.Chest CT imaging was performed in allpatients.CT imaging features benign and malignant nodules were compared，including lesion diameter，solid nodules，pure ground glass density nodules，partial solid nodules，blood flow （BF）value，mean transit time （MTT），blood volume（BV）and surface permeability（PS）.Multivariate regressionwas usedtoanalyze thefactors influencing the characteristic parameters chest CT images pulmonary nodule malignancyand to develop arisk prediction model based on the nomogram.The receiver operating characteristic（ROC）curve was used to analyze the diagnostic eficiency the nomogram model for malignant pulmonary nodules.ResultsThe proportion lesion diameter ?10 mm and partial solid nodules in malignant nodules was higher than that in the benign nodules group （all P<0.05 ）.BF，BVand PS in the malignant nodules group were higher than those in the benign nodules group （all P<0.05 ）. analysis showed that lesion diameter ?10 mm，partial solid nodules，BFvalue，BVand PS were the risk factors for malignant pulmonary nodules （ P <0.05）.Based on lesion diameter，partial solid nodule，BF value，BV and PS chest CT image featureparameters，the risk nomogram for the diagnosis benign and malignant pulmonary nodules was established. The total score was 220，and the prediction probabilitywas 0.919.Moreover，theROC analysis nomogram model predictedtheareaunder thecurve （AUC）to beO.9O3，indicatingarelatively high prediction ability the prediction model. Conclusion Lesion diameter ?10mm ，proportion partial solid nodules，BF value，BV and PS in patients with malignant nodules were higher than those in the benign nodules group，and the above indexes were all risk factorsfor malignantpulmonary nodules.Thetotalscore the nomogram risk prediction model basedonthecharacteristic parameters chest CTimages was 220，the prediction probabilitywas O.919，and theprediction ability was high.

[Key Words】Solitary pulmonary nodule; Chest CT; features;Proportional hazards model; Diagnostic vaue部結(jié)節(jié)。（剩余6015字）