Differentiating Small (≤1 cm) Focal Liver Lesions as Metastases or Cysts by means of Computed Tomography: A Case-Study to Illustrate a Fuzzy Logic-Based Method to Assess the Impact of Diagnostic Confidence on Radiological Diagnosis

Table 5

Diagnostic accuracy of reader R1 and reader R2 in assessing malignancy of small focal liver lesions using multidetector row Computed Tomography (metastases versus cysts). Crisp and fuzzy accuracies are reported, together with 95% C.I.s (in brackets) and the divergence 𝛿(𝐹,𝐶) (see the text).

	R1			R2
	Crisp	Fuzzy	𝛿(𝐹,𝐶) (%)	Crisp	Fuzzy	𝛿(𝐹,𝐶) (%)
Sensitivity (%)	90.0	90.0		93.3	94.0
	(77.4–96.3)	(77.4–96.3)	0	(81.6–98.1)	(82.5–98.4)	−0.71
	(27/30)	(27/30)		(28/30)	(28.2/30)
Specificity (%)	100	99.0		95.0	88.0
	(91.1–100)	(89.5–100)	+1.01	(83.8–98.9)	(75.0–95.0)	+7.95
	(20/20)	(19.8/20)		(19/20)	(17.6/20)
PPV; (%)	100	99.3		96.6	92.2
	(91.1–100)	(89.9–100)	+0.74	(85.9–99.5)	(80.1–97.5)	+4.77
	(27/27)	(27/27.2)		(28/29)	(28.2/30.6)
NPV; (%)	87.0	86.8		90.5	90.7
	(73.3–94.4)	(73.6–94.3)	+0.13	(78.0–96.5)	(78.3–96.7)	−0.27
	(20/23)	(19.8/22.8)		(19/21)	(17.6/19.4)
Accuracy (%)	94.0	93.6		94.0	91.6
	(82.5–98.4)	(81.9–98.2)	+0.43	(82.5–98.4)	(79.4–97.2)	+2.62
	(47/50)	(46.8/50)		(47/50)	(45.8/50)