Table 5.

Showing the molecular classification (IDH mutant and IDH wild type) of grade-IV GBM modeled by using Support Vector Machine as the radiomics classifier on MRI-based sequences.

ImageSingle slice v/s Multi slice	MRI sequence	IDHClassification	Radiomics classifier	Diagnostic Metrics		Validation Process
				AUC	Accuracy	10-fold internal cross-validation	Hold Validation
Single slice analysis	T1C	IDH –VE (694)	Linear SVM	0.91	89.8%	YES	NO
		IDH +VE (137)		0.91
	T2W	IDH –VE (689)	Cubic SVM	0.84	86.9%	YES	NO
		IDH +VE (149)		0.84
Multi-slice analysis	T1C	IDH –VE (83)	Linear SVM	0.87	87%	YES	NO
		IDH +VE (17)		0.87
	T2W	IDH –VE (83)	Quadratic SVM	0.80	91%	YES	NO
		IDH +VE (17)		0.80
	T1C + T2W	IDH –VE (83)	Quadratic SVM	0.89	89%	YES	NO
		IDH +VE (17)		0.89
	T1C + T2W	IDH –VE (83)	Cubic SVM	0.81	90%	NO	YES (90:10)
		IDH +VE (17)		0.81

AUC, Area under the curve; SVM, Support Vector Machine.