Table 1 Characteristics of study objects.

From: Using deep learning to predict temporomandibular joint disc perforation based on magnetic resonance imaging

 

Non perforated group

Perforated group

p-value

Number of joints

N = 168

N = 131

 

Age (year)***

  

< 0.001

Median [IQR]

27.0 [22.0; 33.0]

32.0 [26.5; 44.5]

 

Gender*

  

0.029

Male

30 (17.9%)

10 (8.4%)

 

Female

138 (82.1%)

120 (91.6%)

 

Shape of the disc***

  

< 0.001

Biconcave

30 (17.9%)

1 (0.8%)

 

Folded

46 (27.4%)

7 (5.3%)

 

Flattened

34 (20.2%)

17 (13.0%)

 

Eyeglass-shaped

40 (23.8%)

35 (26.7%)

 

Amorphous

18 (10.7%)

71 (54.2%)

 

Signal intensity of bone marrow***

  

< 0.001

Normal

152 (90.5%)

87 (66.4%)

 

Low

16 (9.5%)

44 (33.6%)

 

Fluid collection

  

0.178

Grade 0

82 (48.8%)

54 (41.2%)

 

Grade 1

58 (34.5%)

52 (39.7%)

 

Grade 2

24 (14.3%)

18 (13.7%)

 

Grade 3

4 (2.4%)

7 (5.3%)

 

Disc displacement***

  

< 0.001

Normal + ADcR

51 (30.4%)

7 (5.3%)

 

Early ADsR

56 (33.3%)

31 (23.7%)

 

Late ADsR

61 (36.3%)

93 (71.0%)

 

Joint space***

  

< 0.001

Normal

38 (22.6%)

7 (5.3%)

 

Narrowing

125 (74.4%)

91 (69.5%)

 

Bone to bone contact (close)

3 (1.8%)

12 (9.2%)

 

Bone to bone contact (open)

2 (1.2%)

21 (16.0%)

 

Changes of condyle and fossa***

  

< 0.001

2 or less features

149 (88.7%)

72 (55.4%)

 

2 or more features

19 (11.3%)

58 (44.6%)

 
  1. IQR Interquartile range, ADcR anterior displacement with reduction, ADsR anterior displacement without reduction.
  2. *p < 0.05, **p < 0.01, ***p < 0.001.