Fig 1.
The workflow of this study.
Table 1.
Baseline characteristics of patients in cohorts.
Fig 2.
Number and ratio of handcrafted features.
A show CT features, B show PET features.
Fig 3.
Radiomic features selected using a LASSO regression model for subgroups.
A-C The coefficients of each feature in the most predictive feature subset. The abscissa is the coefficient, and the ordinate shows the reserved features. The larger the coefficient is, the more predictive effect of the feature is. A shows feature selected in the clinic model, B shows feature selected in the RS model, C shows feature selected in the combined model, D MSE of 10 fold cross validation. E Coefficients of 10 fold cross validation.
Table 2.
Univariate logistic regression analysis of clinical predictors of histology.
Table 3.
Multivariate logistic regression analysis of clinical predictors of histology.
Table 4.
The final selected PET-CT radiomics features for RS model and combined model.
Table 5.
Performance of four machine learning algorithms for differentiating pathological subtypes in the training and internal validation cohort.
Fig 4.
Comparison of receiver operating characteristic (ROC) curves for predicting subtype of pathology.
A shows the ROC curve of LR in the training cohort; B shows the ROC curve of LR in the validation cohort.
Table 6.
DeLong test within different models based on LR classifier for the validation cohort.
Fig 5.
Clinical utility of prediction models.
A shows Nomogram of a clinical radiomics model developed based on a logistic regression model for the training cohort. gender 1:male 2:female. B,C show that Decision curve analysis (DCA) was conducted for the prediction model based on the logistic regression model in the training (B) and validation cohorts (C). D,E show Calibration curves of the nomogram based on the logistic regression model in the training (D) and validation cohorts (E).