Development and validation of machine learning models for predicting no. 253 lymph node metastasis in left-sided colorectal cancer using clinical and CT-based radiomic features

Table 3 Evaluation metrics of the models

Model	AUC	ACC	SEN	SPE	PPV	NPV
Test set
Clinical model	0.694	0.654	0.808	0.577	0.488	0.857
CT model	0.663	0.705	0.462	0.827	0.571	0.754
Radiomics model	0.720	0.667	0.731	0.635	0.500	0.825
Combined model	0.663	0.590	0.885	0.442	0.442	0.885
Temporal validation set
Clinical model	0.743	0.644	0.867	0.533	0.481	0.889
CT model	0.629	0.644	0.933	0.500	0.483	0.938
Radiomics model	0.716	0.756	0.667	0.800	0.625	0.828
Combined model	0.800	0.800	0.800	0.800	0.667	0.889

The model determines the optimal decision threshold by maximizing Youden’s Index
AUC = area under the curve; ACC = accuracy; SEN = sensitivity; SPE = specificity; PPV = positive predictive value; NPV = negative predictive value

ISSN: 1470-7330