Table 3.

Predictive performance of the image features model, the radiomics signature, and the radiomics nomogram.

Model	Training cohort (n = 132)				Validation cohort (n = 123)
	AUC (95% CI)	Sensitivity	Specificity	Precision	AUC (95% CI)	Sensitivity	Specificity	Precision
Image features model	0.82 (0.75–0.89)	86.42%	84.31%	85.61%	0.72 (0.62–0.82)	59.30%	78.38%	65.04%
Radiomics Signature	0.91 (0.87–0.96)	75.31%	82.35%	78.03%	0.86 (0.79–0.93)	82.56%	70.27%	78.86%
radiomics nomogram	0.93 (0.89–0.97)	76.54%	96.08%	84.09%	0.87 (0.81–0.94)	72.26%	83.78%	76.42%