The added value of radiomics from dual-energy spectral CT derived iodine-based material decomposition images in predicting histological grade of gastric cancer

Table 2 Diagnostic performance of models in training and testing cohorts

	AUC (95% CI)	ACC	SEN	SPE	PPV	NPV
Model-Clinical
Training	0.674 (0.543–0.804)	0.694	0.795	0.536	0.729	0.625
Testing	0.847 (0.612–0.950)	0.710	0.579	0.917	0.917	0.579
Model-CP
Training	0.802 (0.693–0.911)	0.792	0.818	0.750	0.837	0.724
Testing	0.781 (0.612–0.950)	0.742	0.737	0.750	0.824	0.643
Model-IMD
Training	0.871 (0.793–0.950)	0.792	0.818	0.750	0.837	0.724
Testing	0.759 (0.582–0.936)	0.774	0.790	0.750	0.833	0.692
Model-CP–IMD
Training	0.900 (0.830–0.971)	0.861	0.818	0.927	0.947	0.765
Testing	0.851 (0.711–0.991)	0.839	0.842	0.833	0.889	0.769
Model-Combine
Training	0.910 (0.837–0.983)	0.875	0.955	0.750	0.857	0.913
Testing	0.912 (0.778–1.000)	0.936	0.747	0.917	0.974	0.917

AUC area under the receiver operating curve, 95% CI 95% confidence interval, ACC accuracy, SEN sensitivity, SPE specificity, PPV positive predictive value, NPV negative predictive value

ISSN: 1471-2342