Construction of a nomogram to predict the malignant transformation of solitary pulmonary nodules based on artificial intelligence parameters of CT and serological indicators,and its validation
ZHU Lie
YAO Benbo
QIAN Yong
Abstract：Objective To investigate the influencing factors,artificial intelligence(AI)parameters of computed tomography(CT)and serological indicators of malignant transformation of solitary pulmonary nodules(SPNs),and to construct a nomogram to predict it and validate its performance.Methods By a retrospective case-control method,clinical data of 178 SPN patients treated from January 2021 to November 2023 were collected.They were divided into malignant transformation group and non-malignant transformation group according to pathological findings.Demographic characteristics,AI parameters of CT and serological indicators of the two groups were recorded.Lasso-Logistic regression was used to screen risk factors for malignant transformation of SPNs.A nomogram was constructed according to the screened variables,and internal and external verification was carried out.In addition,50 SPN patients admitted from December 2023 to February 2024 were selected according to the same criteria for external verification.Results The incidence of malignant transformation of SPNs was 34.83％with pathological results as the gold standard.Lasso-Logistic regression equation showed that the average CT value,proportion of solid SPNs,smoking history,history of respiratory diseases,macrophage-inhibiting cytokine-1(MIC-1)and cluster of differentiation 147(CD 147)were the influencing factors of malignant transformation of SPNs(P＜0.05).Based on the analysis results of Lasso-Logistic regression equation,a nomogram to predict the malignant transformation of SNPs was created.The area under the curve(AUC)of the nomogram was 0.805 and 0.782,respectively.The diagnostic results were basically consistent with the actual results,with significant clinical net benefit in the range of 0.05-0.8 and 0.1-0.8.In the range of 0-0.75 and 0-0.80,patients with high-risk malignant transformation of SPNs could be effectively distinguished.Conclusion The nomogram model based on the average CT value,solid proportion,MIC-1,CD 147 and other factors is helpful to predict the risk of malignant transformation of SPNs,identify high-risk groups,and guide clinical diagnosis and treatment.
Keywords：solitary pulmonary nodulesmalignant transformationinfluencing factorsnomogram modelartificial intelligence parameters of CTserological indicators
Publication Date：2025-01-28
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：5( 33-37 )