Table 4.

The diagnostic performance in differentiating malignancies from BOT based on various MR-based radiomics models

Model	Group	SEN	SPE	PPV	NPV	ACC	AUC(95% CI)
2d_cor	Training	0.708	0.936	0.919	0.759	0.821	0.90(0.85–0.96)
2d_cor	Testing	0.729	0.851	0.833	0.755	0.789	0.82(0.73–0.90)
3d_cor	Training	0.875	0.717	0.764	0.846	0.798	0.85(0.77–0.93)
3d_cor	Testing	0.936	0.717	0.772	0.917	0.828	0.84(0.76–0.93)
2d_sag	Training	0.776	0.902	0.884	0.807	0.840	0.89(0.83–0.96)
2d_sag	Testing	0.729	0.824	0.795	0.764	0.778	0.79(0.69–0.88)
3d_sag	Training	1.000	1.000	1.000	1.000	1.000	1.0(1.0–1.0)
3d_sag	Testing	1.000	0.980	0.980	1.000	0.990	1.0(1.0–1.0)

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