Correlation Study Between Imaging Radiomics Features And Molecular Markers Of Benign And Malignant Pulmonary Nodules

Objective Part Ⅰ To discuss the radiomic features of benign and malignant pulmonary nodules and to analyze the differentiating value between benign and malignant pulmonary nodules.Part Ⅱ To investigate the correlation between radiomics features in benign and malignant pulmonary nodules and the mutation of epidermal growth factor(EGFR)gene and the expression of Ki-67 protein.Methods Part Ⅰ Thirty patients with benign pulmonary nodules and thirty with malignant pulmonary nodules which has pathological results were selected.The pulmonary nodules were segmented by ITK-SNAP software,and the features were extracted by using AK software.Part Ⅱ 120 patients having pulmonary nodules were collected,according to the mutation of EGFR gene locus and the expression of Ki-67 protein,the patients were divided into two groups(positive group and negative group)We analyzed the correlation between mutation of EGFR gene locus and expression of ki-67 protein.And we performed the correlation analvsis between radiomics features and the results of EGfR and Ki-67.Results Part Ⅰ There were 396 features obtained after calculation from each group,and there were significant differences in the run-length matrix(RLM)and the gray-level co-occurrence matrix(GLCM)between the benign and malignant groups.Part Ⅱ 1.There was no significant difference in EGFR gene mutation between granulomatous lesions and adenocarcinoma in situ(P>0.05).There was significant difference between granulomatous lesions in minimally invasive adenocarcinoma and invasive adenocarcinoma(P<0.05),but no significant difference in age and sex ratio(P>0.05).2.There was significant difference in the expression of Ki-67 protein between granulomatous lesions and lung adenocarcinoma(P<0.05),but no significant difference with age and sex ratio(P>0.05).3.In terms of radiomics features,there was a correlation between EGFR gene mutation and "Inertia AllDirection offset4" of the gray-level co-occurrence matrix(GLCM)(P<0.05),and the expression of Ki-67 was correlated with"ShortRunl Emphasis_AllDirection_offsetl" of the run-level matrix(RLM)(P<0.05),"InverseDifferenceMoment AllDirection offset4","Inertia_AlIDirection_offset4" of the gray-level co-occurrence matrix(GLCM)(P<0.05).4.There was a correlation between EGFR gene mutation and expression of Ki-67 protein(P<0.05).Conclusion Part Ⅰ Radiomics may be used to differentiate benign and malignant pulmonary nodules and to provide objective basis for differentiating.Part Ⅱ The radiomics features of lung may be used to guide the diagnosis of gene mutation,proliferative activity and prognosis of lunu tumors.