Skip to main content

Table 3 Diagnostic performance of multi-b diffusion maps based on ME, BE, SE and DKI models

From: Radiomics-based machine learning analysis and characterization of breast lesions with multiparametric diffusion-weighted MR

Maps

AUC (95% CI)

Sensitivity% (95% CI)

Specificity% (95% CI)

PPV% (95% CI)

NPV% (95% CI)

ME

 ADC0–1000

0.83 (0.80–0.87)

88 (82–93)

79 (68–87)

87 (81–92)

80 (72–88)

 ADCall-b

0.84 (0.80–0.88)

88 (82–93)

81 (71–90)

88 (82–94)

80 (72–88)

BE-IVIM

 D

0.85 (0.81–0.89)

88 (81–93)

82 (73–90)

89 (83–93)

80 (72–89)

 D*

0.80 (0.74–0.83)

83 (76–88)

77 (66–86)

85 (77–91)

73 (65–81)

 f

0.82 (0.77–0.86)

85 (79–90)

79 (68–87)

86 (80–92)

77 (68–84)

SE

 DDC

0.81 (0.77–0.85)

85 (78–90)

77 (66–86)

85 (79–91)

76 (68–84)

 α

0.80 (0.77–0.84)

85 (79–89)

76 (67–86)

85 (80–91)

76 (69–82)

DKI

 D

0.83 (0.80–0.88)

87 (80–94)

79 (69–88)

87 (81–93)

80 (71–89)

 K

0.84 (0.80–0.88)

87 (80–92)

81 (72–89)

88 (83–93)

79 (72–86)

  1. AUC area under the receiver operating characteristic curve, CI confidence interval, PPV positive predictive value, NPV negative predictive value