Preoperative subcategorization based on magnetic resonance imaging in intrahepatic cholangiocarcinoma

Cancer Imaging

Table 5 Diagnostic performance of the significant imaging features and predictive model

Imaging features	Sensitivity	Specificity	Accuracy	Positive LR	Negative LR	AUC	P value ^a	P value ^b	P value ^c
AP hypoenhancement	30.0% (9/30)	95.2% (60/63)	64.5% (69/93)	6.25	0.74	0.63 [0.50,0.76]	–	< 0.001/< 0.001/0.429	< 0.001/0.008/0.011
Targetoid restriction	86.7% (26/30)	49.2% (31/63)	57.0% (57/93)	1.71	0.27	0.68 [0.57,0.79]	< 0.001/< 0.001/0.429	–	0.250/< 0.001/0.099
Intrahepatic duct dilatation	76.7% (23/30)	82.5% (52/63)	83.9% (78/93)	4.38	0.28	0.80 [0.69,0.90]	< 0.001/0.008/0.011	0.250/< 0.001/0.099	–
Predictive model	80.0% (24/30)	93.7% (59/63)	89.2% (83/93)	12.70	0.210	0.91 [0.85,0.98]	< 0.001/> 0.999/< 0.001	0.500/< 0.001/< 0.001	> 0.999/0.016/< 0.001

AP Arterial phase, LR Likely ratio, AUC Area under curve
^a P values were obtained by comparing sensitivity/specificity/AUC between AP hypoenhancement and other characteristics. ^b P values were obtained by comparing sensitivity/specificity/AUC between targetoid restriction and other characteristics. ^c P values were obtained by comparing sensitivity/specificity/AUC between intrahepatic duct dilatation and other characteristics

ISSN: 1470-7330