Development and validation of a Radiopathomics model based on CT scans and whole slide images for discriminating between Stage I-II and Stage III gastric cancer

Table 2 Performance of models for predicting discrimination between stages I-II and stage III gastric cancer in training and validation cohorts

Model	Task	AUC (95% CI)	Accuracy	Sensitivity	Specificity	PPV	NPV	Precision	Recall	F1 score
LR_Pathomics	Training	0.892 (0.841–0.943)	0.837	0.871	0.803	0.813	0.864	0.813	0.871	0.841
	Validation	0.680 (0.536–0.825)	0.721	0.886	0.500	0.705	0.765	0.705	0.886	0.785
NaiveBayes_Pathomics	Training	0.809 (0.737–0.881)	0.766	0.914	0.620	0.703	0.880	0.703	0.914	0.795
	Validation	0.679 (0.540–0.818)	0.689	0.771	0.600	0.711	0.652	0.711	0.771	0.740
SVM_Pathomics	Training	0.949 (0.911–0.987)	0.894	0.914	0.873	0.877	0.912	0.877	0.914	0.895
	Validation	0.777 (0.652–0.901)	0.787	0.886	0.654	0.775	0.810	0.775	0.886	0.827
LR_Radiomics	Training	0.742 (0.663–0.824)	0.688	0.857	0.521	0.638	0.787	0.638	0.857	0.732
	Validation	0.720 (0.588–0.852)	0.705	0.743	0.654	0.743	0.654	0.743	0.743	0.466
NaiveBayes_Radiomics	Training	0.752 (0.673–0.831)	0.695	0.843	0.549	0.648	0.780	0.648	0.843	0.733
	Validation	0.733 (0.607–0.859)	0.705	0.714	0.692	0.758	0.643	0.758	0.714	0.735
SVM_Radiomics	Training	0.799 (0.726–0.873)	0.745	0.800	0.690	0.718	0.778	0.718	0.800	0.757
	Validation	0.712 (0.579–0.845)	0.705	0.771	0.615	0.730	0.667	0.730	0.771	0.750
LR_Radiopathomics	Training	0.904 (0.858–0.951)	0.830	0.914	0.746	0.780	0.898	0.780	0.914	0.842
	Validation	0.747 (0.617–0.878)	0.770	0.829	0.692	0.784	0.750	0.784	0.829	0.806
NaiveBayes_Radiopathomics	Training	0.861 (0.801–0.921)	0.787	0.857	0.718	0.750	0.836	0.750	0.857	0.800
	Validation	0.748 (0.624–0.872)	0.738	0.714	0.769	0.806	0.667	0.806	0.714	0.758
SVM_Radiopathomics	Training	0.953 (0.920–0.986)	0.901	0.914	0.887	0.889	0.913	0.889	0.914	0.901
	Validation	0.851 (0.745–0.956)	0.787	0.771	0.808	0.844	0.724	0.844	0.771	0.806

LR logistic regression, SVM support vector machine, AUC area under the curve, PPV positive prediction value, NPV negative prediction value

ISSN: 1471-2407