Skip to main content

Table 1 Clinical characteristics of the training, validation, and test datasets

From: Deep learning predicts cervical lymph node metastasis in clinically node-negative papillary thyroid carcinoma

Characteristics

Training and validation sets

Test set A

Test set B

Number

848

94

89

 Lymph node metastasis

400 (47.2%)

47 (50.0%)

44 (49.4%)

 No lymph node metastasis

448 (52.8%)

47 (50.0%)

45 (50.6%)

Age, mean ± SD, years

42.61 ± 10.42

43.78 ± 9.27

41.53 ± 10.94

 ≤ 45

553 (65.2%)

55 (58.5%)

55 (61.8%)

 > 45

295 (34.8%)

39 (41.5%)

34 (38.2%)

Gender

 Female

675 (79.6%)

78 (83.0%)

72 (80.9%)

 Male

173 (20.4%)

16 (17.0%)

17 (19.1%)

Primary site

 Right lobe

413 (48.7%)

48 (51.1%)

44 (49.4%)

 Left lobe

420 (49.5%)

43 (45.7%)

43 (48.3%)

 Isthmus

15 (1.8%)

3 (3.2%)

2 (2.2%)

Metastatic site

 Central lymph node

271 (67.8%)

30 (63.8%)

28 (64.0%)

 Central and lateral lymph node

89 (22.3%)

12 (24.5%)

11 (24.7%)

 Lateral lymph node

40 (9.9%)

5 (11.7%)

5 (11.2%)

Prophylactic CND

 Tumor size > 4 cm

359 (42.3%)

41 (43.6%)

27 (30.3%)

 Extrathyroidal extension (ETE)

125 (14.7%)

25 (26.6%)

34 (38.2%)

 ETE and Tumor size > 4 cm

364 (43.0%)

28 (29.8%)

28 (31.5%)

TI-RADS category

 4A category

29 (3.4%)

4 (4.3%)

5 (5.6%)

 4B category

175 (20.6%)

22 (23.7%)

22 (24.7%)

 4C category

384 (45.3%)

37 (38.9%)

38 (42.7%)

 5 category

260 (30.7%)

31 (32.9%)

24 (27.0%)

US size, mean ± SD, mm

12.74 ± 6.58

10.93 ± 5.91

11.47 ± 6.35

  1. Qualitative variables are in n (%), and quantitative variables are in mean ± SD