Multi-classification model incorporating radiomics and clinic-radiological features for predicting invasiveness and differentiation of pulmonary adenocarcinoma nodules

Table 6 The performance of multi-classification using SVM-OVO for predicting histological classification of pulmonary adenocarcinoma nodules

Cohort	Model	Two-classification				Three-classification				Five-classification
		AUC	Accuracy	Sensitivity	Specificity	Accuracy	F1-score	Recall	Precision	Accuracy	F1-score	Recall	Precision
Train	Clinic-radiological	0.932	0.874	0.864	0.885	0.749	0.721	0.713	0.734	0.557	0.539	0.534	0.613
	Radiomics	0.962	0.877	0.848	0.909	0.846	0.830	0.820	0.842	0.700	0.696	0.685	0.722
	Combined	0.985	0.934	0.924	0.945	0.883	0.878	0.870	0.886	0.703	0.693	0.690	0.715
Test	Clinic-radiological	0.905	0.848	0.838	0.859	0.740	0.696	0.697	0.700	0.513	0.506	0.505	0.514
	Radiomics	0.938	0.868	0.825	0.915	0.753	0.712	0.706	0.724	0.553	0.531	0.536	0.559
	Combined	0.942	0.894	0.875	0.915	0.767	0.718	0.710	0.731	0.607	0.594	0.603	0.593

ISSN: 1475-925X