Skip to main content

Table 2 Diagnostic performance of four models in the evaluation of normal renal function

From: MRI texture-based machine learning models for the evaluation of renal function on different segmentations: a proof-of-concept study

 

Training cohort

Validation cohort

Models

AUC

Sensitivity (%)

Specificity (%)

Accuracy (%)

AUC

Sensitivity (%)

Specificity (%)

Accuracy (%)

All-K

0.877 (0.870–0.884)

74.0 (73.5–74.5)

90.0 (89.4–90.6)

83.3 (82.8–83.9)

0.866 (0.863–0.870)

76.2 (75.0–78.2)

74.2 (71.3–74.5)

75.0 (74.5–75.5)

LC-K

0.938 (0.935–0.940)

84.0 (82.9–85.1)

91.4 (90.1–92.7)

88.3 (87.5–89.2)

0.802 (0.800–0.807)

79.2 (75.6–76.8)

83.9 (82.3–85.5)

80.8 (80.2–81.4)

TLCO-K

0.922 (0.912–0.928)

82.0 (81.4–82.6)

84.3 (82.4–86.2)

83.3 (82.3–83.9)

0.800 (0.785–0.816)

57.1 (55.7–58.5)

80.6 (79.9–81.3)

71.2 (70.6–71.7)

PIZZA-K

0.922 (0.914–0.929)

72.0 (71.5–72.5)

90.0 (89.9–90.0)

82.5 (81.7–83.3)

0.800 (0.784–0.958)

85.7 (85.1–86.2)

51.6 (51.1–52.1)

65.4 (64.9–65.9)

  1. Data in parentheses are 95% confidence interval