Identification and validation of radiomic features from computed tomography for preoperative classification of neuroblastic tumors in children

Table 2 Performance of the radiomic-clinical classifier for distinguishing the three types of neuroblastic tumor

Classification step	Tumor Type	Training set
		Accuracy	Sensitivity	Specificity	AUC	Accuracy	Sensitivity	Specificity		AUC
Individual binary classifications	GN vs. non-GN	93.5%	90.0%	94.0%	0.969	84.6%	100%	81.8%	0.875
Individual binary classifications	GNB vs. NB	89.6%	83.3%	91.8%	0.931	86.4%	83.3%	87.5%	0.854
		GN	GNB	NB	Overall accuracy	GN	GNB	NB	Overall accuracy
Three-way classification	GN vs. GNB vs. NB	90.0%	66.7%	91.8%	85.7%	100%	50%	87.5%	80.8%

AUC, area under the receiver operating characteristic curve; GN, ganglioneuroma; GNB, ganglioneuroblastoma; NB, neuroblastoma

ISSN: 1471-2431