Skip to main content

Table 1 Patient characteristics for training and test cohorts

From: Deep learning tools for the cancer clinic: an open-source framework with head and neck contour validation

Characteristic

Dataset, No. (%)

Training (210)

Test (19)

Age at diagnosis, median (range)

61 (40–84)

65 (53–88)

Sex

  

Female

38 (18.1)

5 (26.3)

Male

172 (81.9)

14 (73.7)

Primary tumor site

  

Base of tongue

45 (21.4)

2 (10.5)

Hypopharynx

10 (4.8)

1 (5.3)

Larynx

34 (16.2)

0 (0.0)

Nasopharynx

10 (4.8)

5 (26.3)

Oropharynx

21 (10.0)

4 (21.1)

Supraglottis

6 (2.9)

1 (5.3)

Tonsil

45 (21.4)

5 (26.3)

Other/unknown

39 (18.6)

1 (5.3)

Tumor stage

  

TX

6 (2.9)

0 (0.0)

T0

12 (5.7)

1 (5.3)

T1

34 (16.2)

3 (15.8)

T2

50 (23.8)

5 (26.3)

T3

87 (41.4)

9 (47.4)

T4

10 (4.8)

1 (5.3)

Unavailable

11 (5.2)

0 (0.0)

Node stage

  

N0

46 (21.9)

3 (15.8)

N1

37 (17.6)

2 (10.5)

N2

99 (47.1)

13 (68.4)

N3

17 (8.1)

1 (5.3)

Unavailable

11 (5.2)

0 (0.0)

  1. Tumor staging ranges from TX/T0 (cannot be measured/found) to T4, with larger numbers indicating larger tumors. Node staging ranges from T0 (no cancer in nearby lymph nodes) to N3, with larger numbers representing greater presence of cancer in lymph nodes