Table 1 Characteristics of dataset stratified by training and test datasets.

From: Foundation model based prediction of lung cancer survival using temporal changes in dual time point CT scans

Characteristic

All patients (n = 102)

Training dataset (n = 86)

Test dataset (n = 16)

p-value

Age (yrs)

Median (IQR)

60.0 (9.3)

60.0 (9.4)

60.5 (8.7)

0.100

Sex

Female

60 (58.8%)

49 (57%)

11 (68.8%)

0.380

Male

42 (41.2%)

37 (43%)

5 (31.2%)

0.380

Smoking status

Median (IQR) pack years smoking

40.0 (27.9)

40.0 (30.8)

40.0 (36.5)

0.468

Tumor characteristics

Clinical overall stage

II (n = 9, 8.8%)

9 (8.8%)

9 (10.5%)

0 (0%)

0.175

A

2 (2.0%)

2 (2.3%)

0 (0%)

 

B

7 (6.9%)

7 (8.1%)

0 (0%)

 

III (n = 93, 91.2%)

93 (91.2%)

77 (89.5%)

16 (100%)

0.175

A

77 (75.5%)

65 (75.6%)

12 (75%)

 

B

13 (12.7%)

9 (10.5%)

4 (25%)

 

IV (n = 3, 2.9%)

3 (2.9%)

3 (3.5%)

0 (0%)

0.448

Median (IQR) clinical tumor size at diagnosis (cm)

4.2 (2.4)

4.4 (3.4)

3.3 (1.5)

0.054

Survival

    

2-year survival

73 (72%)

62 (72%)

11 (68.8%)

0.786

Median (IQR) length of follow-up (yr)

3.64 (4.38)

4.07 (4.4)

2.54 (2.4)

0.26

  1. Variable distributions reported as n (%) unless otherwise specified. p-value calculated by two-sided t-test for continuous variables and chi-squared test for binary/categorical variables.