Skip to main content

Table 3 Model performance for predicting whole-body composition in the external dataset

From: CT analysis of thoracolumbar body composition for estimating whole-body composition

 

Multi-slice averaging in the waist

Single-slice L2 lower endplate

Single-slice L3 upper endplate

Multi-slice averaging in T12–L1

Single-slice T12 lower endplate

Single-slice L1 upper endplate

Before updating models

Muscle

 R2

0.48

0.41

0.45

0.42

0.46

0.39

 RMSE (L)

2.79

2.98

2.89

2.96

2.86

3.04

 MAE (L)

2.51

2.64

2.58

2.68

2.54

2.71

 Calibration-in-the-large (L)

− 0.04 (− 0.71, 0.63)

0.38 (− 0.38, 1.15)

0.13 (− 0.63, 0.88)

− 0.25 (− 1.01, 0.51)

− 0.05 (− 0.87, 0.77)

0.12 (− 0.67, 0.92)

 Calibration slope

0.87 (0.84, 0.90)

0.84 (0.80, 0.88)

0.86 (0.82, 0.90)

0.88 (0.84, 0.91)

0.87 (0.83, 0.91)

0.85 (0.81, 0.90)

Visceral fat

 R2

0.94

0.93

0.93

0.9

0.87

0.89

 RMSE (L)

0.35

0.38

0.38

0.45

0.51

0.47

 MAE (L)

0.26

0.3

0.3

0.33

0.37

0.36

 Calibration-in-the-large (L)

− 0.02 (− 0.11, 0.07)

− 0.16 (− 0.26, − 0.06)

− 0.17 (− 0.27, − 0.06)

− 0.02 (− 0.12, 0.09)

0.01 (− 0.11, 0.13)

− 0.07 (− 0.18, 0.05)

 Calibration slope

0.99 (0.95, 1.02)

1.03 (1.00, 1.07)

1.04 (1.01, 1.08)

0.93 (0.90, 0.97)

0.93 (0.89, 0.97)

0.96 (0.92, 1.00)

Subcutaneous fat

 R2

0.61

0.44

0.44

0.52

0.53

0.47

 RMSE (L)

3.16

3.8

3.8

3.51

3.48

3.71

 MAE (L)

2.7

3.4

3.4

3.02

2.96

3.2

 Calibration-in-the-large (L)

− 0.30 (− 0.89, 0.29)

− 0.71 (− 1.31, − 0.10)

− 0.78 (− 1.40, − 0.16)

0.16 (− 0.39, 0.72)

0.35 (− 0.21, 0.91)

0.06 (− 0.53, 0.66)

 Calibration slope

0.85 (0.81, 0.89)

0.84 (0.80, 0.87)

0.84 (0.81, 0.88)

0.80 (0.77, 0.84)

0.80 (0.76, 0.83)

0.80 (0.77, 0.84)

After updating models

Muscle

 R2

0.91

0.88

0.89

0.89

0.87

0.88

 RMSE (L)

1.15

1.33

1.29

1.28

1.38

1.35

 MAE (L)

0.81

0.88

0.86

0.88

0.93

0.93

 Calibration-in-the-large (L)

− 0.04

0.38

0.13

− 0.25

− 0.05

0.12

(− 0.71, 0.63)

(− 0.38, 1.15)

(− 0.63, 0.88)

(− 1.01, 0.51)

(− 0.87, 0.77)

(− 0.67, 0.92)

 Calibration slope

1.01

0.84

1

0.88

1.01

1

(0.97, 1.04)

(0.80, 0.88)

(0.95, 1.04)

(0.84, 0.91)

(0.96, 1.06)

(0.95, 1.04)

Visceral fat

 R2

0.94

0.92

0.92

0.92

0.88

0.9

 RMSE (L)

0.34

0.4

0.41

0.39

0.48

0.45

 MAE (L)

0.25

0.3

0.31

0.28

0.35

0.32

 Calibration-in-the-large (L)

− 0.02

− 0.16

− 0.17

− 0.02

0.01

− 0.07

(− 0.11, 0.07)

(− 0.26, − 0.06)

(− 0.27, − 0.06)

(− 0.12, 0.09)

(− 0.11, 0.13)

(− 0.18, 0.05)

 Calibration slope

1.02

1.03

1.04

0.93

1.03

0.96

(0.99, 1.05)

(1.00, 1.07)

(1.01, 1.08)

(0.90, 0.97)

(0.99, 1.08)

(0.92, 1.00)

Subcutaneous fat

 R2

0.9

0.89

0.88

0.9

0.9

0.89

 RMSE (L)

1.63

1.71

1.76

1.57

1.6

1.67

 MAE (L)

1.23

1.3

1.35

1.14

1.17

1.24

 Calibration-in-the-large (L)

− 0.3

− 0.71

− 0.78

0.16

0.35

0.06

(− 0.89, 0.29)

(− 1.31, − 0.10)

(− 1.40, − 0.16)

(− 0.39, 0.72)

(− 0.21, 0.91)

(− 0.53, 0.66)

 Calibration slope

1.05

1.1

1.11

1.01

0.99

1.02

(1.00, 1.09)

(1.05, 1.14)

(1.06, 1.15)

(0.97, 1.05)

(0.94, 1.03)

(0.97, 1.06)

  1. RMSE Root-mean-squared error, MAE Mean absolute error
  2. Data in parentheses indicate 95% confidence intervals