Table 4 Performance of the constructed models in predicting CRLM.

	Model	AUC	Sensitivity	Specificity	PPV	NPV	Accuracy	F1
	Clinical_model	0.755	0.571	0.902	0.846	0.692	0.742	0.682
Training cohort	DWI_model	0.844	0.870	0.683	0.720	0.848	0.774	0.788
	T2_model	0.834	0.831	0.768	0.771	0.829	0.799	0.800
	M_model	0.853	0.766	0.841	0.819	0.793	0.805	0.792
	U_model	0.890	0.818	0.878	0.863	0.837	0.849	0.840
	Clinical_model	0.697	0.556	0.885	0.625	0.852	0.800	0.588
Validation cohort	DWI_model	0.750	0.889	0.692	0.500	0.947	0.743	0.640
Validation cohort	T2_model	0.786	0.667	0.846	0.600	0.880	0.800	0.632
	M_model	0.808	0.889	0.692	0.500	0.947	0.743	0.640
	U_model	0.842	0.889	0.769	0.571	0.952	0.800	0.696

CRLM, colorectal cancer liver metastasis; M model, multisequence radiomic model; U model, union of the multisequence radiomic model and clinical model; AUC, area under the receiver operating characteristic curve; Sen, sensitivity; Spe, specificity; PPV, positive predictive value; NPV, negative predictive value; and ACC, accuracy.

Quick links

Search