Table 3.

Metrics of IHC-decision tree and machine learning algorithms

	Algorithm	Antibody combination	Acc	Sens	Spec	PPV	NPV	LR+	LR−
IHC-decision tree	Nyman	3,5	0.79	0.65	0.95	0.93	0.71	12.47	0.37
	Colomo	1,2,5	0.84	0.77	0.91	0.91	0.79	8.98	0.25
	Hans	1,2,5	0.89	0.95	0.83	0.86	0.94	5.52	0.06
	Hans*	1,5	0.86	0.95	0.76	0.81	0.94	3.94	0.06
	Choi	1,2,3,4,5	0.93	1.00	0.84	0.87	1.00	6.44	0.00
	Choi*	1,3,4,5	0.83	0.79	0.86	0.86	0.79	5.73	0.24
	VY3	1,2,3	0.90	0.97	0.83	0.86	0.96	5.61	0.04
	VY4	1,2,3,4	0.90	0.97	0.83	0.86	0.96	5.61	0.04
Machine learning	PV	1,3,4,5	0.94	0.95	0.93	0.94	0.95	13.8	0.05
	ANN	1,2,3,4,5	0.94	0.95	0.93	0.94	0.95	13.8	0.05
	BS	1,2,3,4,5	0.94	0.95	0.93	0.94	0.95	13.8	0.05
	SVM	1,2,3,4,5	0.94	0.97	0.91	0.92	0.96	11.23	0.04
	SVM	1,2,3,4	0.94	0.97	0.91	0.92	0.96	11.23	0.04

Metrics correspondent to eight IHC-decision tree algorithms and the best five machine learning algorithms are shown, cases of the VY subset were classified. Numeric Tags 1= CD10, 2 = BCL6, 3 = FOXP1, 4 = GCTE1, and 5 = MUM1. IHC-decision tree algorithms could not overcome any of the remarkable metrics obtained for the best five machine learning algorithms

Acc: accuracy; Sens: sensitivity; Spec: specificity; PPV: positive predictive value; NPV: negative predictive values; LR+: likelihood ratio for positive test results; LR−: likelihood ratio for negative test result; PV: Perfecto–Villela; ANN: artificial neural networks; BS: Bayesian simple; SVM: support vector machine