Table 1 Patient demographics and nodal US features.

From: Diagnosis of thyroid micronodules on ultrasound using a deep convolutional neural network

Characteristics

Malignant nodules

Benign nodules

Malignancy rate (%)d

P-value

No. of patients

317

47

  

 Age (years)a

46.0 ± 12.0

45.9 ± 13.0

 

0.97

 Sexb

   

0.34

  Female

251 (79.2%)

40 (85.1%)

  

  Male

66 (20.8%)

7 (14.9%)

  

No. of nodules

323

47

  

 Nodule size (mm)c

5.3 ± 1.5

5.8 ± 2.2

 

0.14

 KSThR TIRADSc

   

0.10

  3

4 (1.2%)

3 (6.4%)

57.1 (18.7)

 

  4

37 (11.5%)

9 (19.2%)

80.4 (5.8)

 

  5

282 (87.3%)

35 (74.5%)

89 (1.8)

 

 CNN TIRADSc

   

< 0.001

  2

1 (0.3%)

2 (4.3%)

33.3 (27.2)

 

  3

2 (0.6%)

3 (6.4%)

40 (21.9)

 

  4

36 (11.1%)

13 (27.7%)

73.5 (6.3)

 

  5

284 (87.9%)

29 (61.7%)

90.7 (1.6)

 
  1. All data except age and malignancy rate are numbers of patients or nodules, with percentages in parentheses.
  2. KSThR Korean Society of Thyroid Radiology, TIRADS Thyroid Imaging Reporting and Data System, CNN convolutional neural network.
  3. aPatient-level comparison using the Student’s t test for continuous variables. bPatient-level comparison using Pearson’s χ2-test for categorical variables. cNodule-level comparison using logistic regression with the generalized estimating equation method. dStandard errors are in parentheses.
Back to article page