| Part one Establish Adiagnosis Models of Indeterminate Solid Pulmonary Nodules Smaller Than 2cm Based on Imaging Radiomics PredictionObjective:Preoperative diagnostic models of solid indeterminate pulmonary nodules Smaller than 2 cm were established and validated based on computed tomography(CT)images,clinical data,and radiomic features.Methods:Retrospective analysis of computed tomography images and clinical data of 413 patients with solid malignant pulmonary nodules(n =210)and benign pulmonary nodules(n =203)cm confirmed by histopathology between 2017 and 2019,Wilcoxon rank sum testing,chi-square test analysis of differences between groups,And the clinical model was established by single factor and multiple factor Logistic regression analysis.Radiomics software was used to depict the regions of interest(ROI)of each nodule and to extract 1691 radiomics features,Of these,289 were trained,The verification group was 124 cases.minimum absolute shrinkage and selection operator(LASSO)and random forest algorithm(RF)were used to select the radiomics features and construct the radiomics model.Combined with clinical features,subjective CT signs and radiomics score(RF-Score)to construct radiomics nomogram.the performance of the model was evaluated by evaluating the area(AUC),sensitivity,specificity,accuracy,positive predictive value(PPV)and negative predictive value(NPV)under the receiver operating characteristic curve.Results:Age,sex,pleural depression and RF-Score showed good predictive performance in the diagnosis cm solid uncertain pulmonary nodules less than 2.Within the training group(P=0.001)and the validation group(P=0.016),the AUC of clinical models were 0.766(95% confidence interval[CI]:0.712-0.821]),0.746(95 % CI:0.659-0.832);the AUC of the radiomic model was 0.896(95% CI:0.861-0.932),0.831(95 % CI:0.758-0.905);the AUC of nomograms are 0.920(95% CI:0.890-0.950),0.845(95% CI:0.777-0.913).In the training and validation groups,compared with the AUC values of the clinical model and the radiomics model / nomogram,all have statistical significance.Part two Discussing the Clinical Value of Different Follow-up Intervals for Diagnosis of Indeterminate Solid Pulmonary Nodules Smaller Than 2 cmObjective: By calculating the volume doubling time of solid indeterminate pulmonary nodules smaiier than 2cm,the clinical value of different follow-up intervals in the diagnosis of them was discussed.Methods:Collected between 2017 and 2019,using the same scan line conditions and parameters of 2 or more times the chest CT examination and the pathologic results of 49 patients with SSPN,nodule volume measurement by Oncology analysis software,the automatic calculation volume doubling time,based on the review within 4 w interval,4-12 w,more than 12 w group,in order to calculate different follow-up interval nodules diagnostic sensitivity(SEN),specific(SPE),accuracy(ACC),positive predictive value(PPV),negative predictive value(NPV).Results:Among the 49 nodules,22 cases were malignant and 27 cases were benign.The software calculated the volume doubling time of nodules for two CT examinations:(1362.8±1645.9)d in the benign group and(204.4±101.4)d in the malignant group.The doubling time of malignant nodules was shorter than that of benign nodules,and the difference was statistically significant(P<0.05).SEN at different intervals was 75%,90%,75%;SPE was 87.5%,100%,90%;ACC was 81.3%,94.7%,85.7%;PPV was85.7%,100%,75%;NPV was 77.8%,90%,90%.Conclusion:1.The radiomics model and nomogram established by us can well predict the solid indeterminate pulmonary nodules smaller than 2 cm,and the prediction ability is higher than that of clinical models.2.The clinical value of the 4-12 w follow-up interval for the diagnosis of indeterminate solid pulmonary nodules smaller than 2cm was better than that of the 4 w review interval and greater than 12 w review interval,which was helpful to guide the clinical development of SSPN follow-up strategy. |
PDF Full Text Request