Artificial intelligence and its 3D characteristics in the diagnosis of pulmonary ground glass nodules
ZHANG Daiyi
YANG Xuejiao
SHI Fei
WU Qiong
LI Zhenyun
PU Jianhong
Abstract:Objective To investigate the value of 3D reconstruction of artificial intelligence(AI)and its described character-istics of pulmonary nodules in the diagnosis of benign and malignant pulmonary ground glass nodules.Methods A total of 187 cases of pulmonary ground glass nodule(GGN)with pulmonary nodules detected by low-dose spiral CT(LDCT)and clearly diag-nosed by surgery and pathology were selected in our hospital.AI three-dimensional reconstruction was performed on their pulmo-nary nodules,and pathological results were used as the gold standard to analyze the relationship between 3D characteristics of AI and benign and malignant lung GGN.The diagnostic efficacy of AI and LDCT in malignant GGN were compared.Results Com-pared with the benign GGN group,the average CT value of the malignant GGN group was lower,and the proportion of microvas-cular crossing,lobulating and burring was larger,while,the diameter of the malignant GGN group was smaller and the propor-tion of pleural thickening was less,with statistical significance(P<0.05).Multivariate logistic regression analysis showed that burr(OR=12.754,95%CI:4.137~39.314,P<0.001)and microvascular travel(OR=3.447,95%CI:1.245~9.541,P=0.017)were independent risk factors for predicting malignant GGN.The Yoden index of AI for malignant GGN diagnosis was 0.472,higher than that of LDCT(0.031),and the sensitivity,positive predictive value,and negative predictive value of AI diagnosis were 92.41%,87.58%and 67.65%,respectively,while,the sensitivity,positive predictive value and negative predictive value of LDCT diagnosis were 31.72%,79.31%and 23.26%,respectively.LDCT was significantly lower than AI diagnosis,the differ-ence was statistically significant(P<0.05).But there was no statistical difference between the two methods(P>0.05).Conclu-sion AI is better than LDCT in the diagnosis of malignant lung GGN,and can be used as a powerful assistant for clinicians to screen lung nodules,which has certain clinical significance.
Keywords:Artificial intelligenceGround glass noduleThree-dimensional featureTomographyX-ray computed
Publication Date:2024-12-28
Online Publishing Date:2025-08-15(First online date of this platform, not the publication date of the document)
Pages:4( 61-64 )
