Table 3.

LIDC multicentre diagnostic performance

Method	Evaluation	Centre A	Centre B	Centre C	Centre D
RFLM (Train)	AUC	0.865	0.842	0.868	0.831
	Sensitivity	0.732 (52/71)	0.761 (67/88)	0.781 (64/82)	0.776 (52/67)
	Specificity	0.907 (166/183)	0.833 (145/174)	0.806 (179/222)	0.778 (140/180)
	Accuracy	0.858 (218/254)	0.809 (212/262)	0.799 (243/304)	0.777 (192/247)
	PPV	0.754 (52/69)	0.698 (67/96)	0.598 (64/107)	0.565 (52/92)
	NPV	0.897 (166/185)	0.874 (145/166)	0.909 (179/197)	0.903 (140/155)
RFLM (Test)	AUC	0.816	0.811	0.852	0.824
	Sensitivity	0.681 (32/47)	0.650 (26/40)	0.807 (25/31)	0.778 (28/36)
	Specificity	0.803 (98/122)	0.812 (121/149)	0.763 (71/93)	0.684 (108/158)
	Accuracy	0.796 (130/169)	0.778 (147/189)	0.774 (96/124)	0.701 (136/194)
	PPV	0.571 (32/56)	0.482 (26/54)	0.532 (25/47)	0.359 (28/78)
	NPV	0.867 (98/113)	0.896 (121/135)	0.922 (71/77)	0.931 (108/116)

AUC area under the curve, PPV positive predictive value, NPV negative predictive value.