Prediction of hepatic necroinflammatory activity in patients with chronic hepatitis B by a simple noninvasive model

Table 4 Diagnostic accuracy of the H value in the training and validation cohorts

	Cutoff	G0–1	≥ G2	SEN	SPE	PPV	NPV	LR+	LR−	Population involved (%)	Interpretation
Training cohort (n = 401)	Low cutoff			0.908	0.485	0.720	0.782	1.761	0.191	59.4	Absence of necroinflammatory activity (78% certainty)
	< 0.375	216	22
	> 0.375	84	79
	High cutoff			0.538	0.920	0.908	0.577	6.743	0.502	40.6	Presence of necroinflammatory activity (91% certainty)
	< 0.720	128	110
	> 0.720	13	150
Validation cohort (n = 401)	Low cutoff			0.951	0.376	0.612	0.881	1.523	0.130	50.9	Absence of necroinflammatory activity (88% certainty)
	< 0.375	194	10
	> 0.375	123	74
	High cutoff			0.495	0.954	0.918	0.646	10.83	0.529	49.1	Presence of necroinflammatory activity (92% certainty)
	< 0.720	101	103
	> 0.720	9	188

In the validation cohort, using a cut-off score of H ≤ 0.375, significant necroinflammatory activity was excluded with high accuracy (88.1% NPV and 95.1% sensitivity). Similarly, applying a cut-off score of H > 0.720, significant necroinflammatory activity was correctly identified with high accuracy (91.8% PPV and 95.4% specificity)
Italic values indicate significant positive values in the validation cohort
SEN sensitivity, SPE specificity, PPV positive predictive value, NPV negative predictive value, LR+ positive likelihood ratio, LR− negative likelihood ratio

ISSN: 1479-5876