Table 1 Patient characteristics of the study population.

From: Automated bone mineral density prediction and fracture risk assessment using plain radiographs via deep learning

 

Hip testing set

Spine testing set

Hologic training

Hologic testing

p***

Hologic training

Hologic testing

p***

Number

5633

5164

 

7307

18,175

 

Female, n (%)

4380 (77.8)

3997 (77.4)

0.66

5815 (79.6)

14,469 (79.6)

0.96

Mean age (sd), years

72.8 (11.0)

72.2 (11.2)

0.004

67.2 (12.5)

67.1 (10.6)

<0.001

Median time (IQR) between DXA and radiographs

31 (8, 81)

29 (8, 77)

0.13

24 (22, 27)

16 (5, 48)

<0.001

Mean BMI (sd), kg/m2*

23.8 (3.9)

23.9 (3.9)

0.08

24.6 (3.8)

24.5 (3.8)

0.06

Mean BMD (sd) g/cm2

0.678 (0.159)

0.689 (0.156)

<0.001

0.740 (0.173)**

0.762 (0.176)**

<0.001

Median T-score (IQR)

−1.6 (−2.4, −0.6)

−1.5 (−2.3, −0.6)

<0.001

−2.3 (−3.3, −1.3)**

−2.2 (−3.1, −1.1)**

<0.001

Osteoporosis, n (%)

1356 (24.1)

1110 (21.5)

<0.001

3291 (45.0)**

7860 (43.3)**

0.009

  1. *BMI was not available in 330/66 patients in the hip training/testing sets and 223 and 623 in the spine training/testing sets.
  2. **Calculated based on vertebrae with the lowest BMD.
  3. ***Categorical variables (gender and osteoporosis) were compared using Chi-square test. Means were compared using student t-test and medians were compared using Wilcoxon rank-sum test. Two-sided p values were reported.
Back to article page