Table 1 Baseline characteristics.

From: Diagnostic ability of artificial intelligence using deep learning analysis of cyst fluid in differentiating malignant from benign pancreatic cystic lesions

Variable

N = 85

Benign

Malignant

p

Patient

n = 62

n = 23

Age (years), mean ± SD

58.2 ± 13.4

57.0 ± 13.7

65.4 ± 9.8

0.003a

Sex (%)

  Male

35 (41.2)

22 (35.5)

13 (56.5)

0.080a

  Female

50 (58.8)

40 (64.5)

10 (43.5)

 

Location of cystic lesion (%)

  Head

29 (34.1)

20 (32.3)

9 (39.1)

0.031a

  Body

25 (29.4)

23 (37.1)

2 (8.7)

 

  Tail

31 (36.5)

19 (30.6)

12 (52.2)

 

Connection of main pancreatic duct and cyst (%)

  Present

34 (40.0)

16 (25.8)

18 (78.3)

<0.001a

  Absent

51 (60.0)

46 (74.2)

5 (21.7)

 

Type of cyst (%)

  Monolocular

16 (18.8)

15 (24.2)

1 (4.3)

0.058b

  Multilocular

69 (81.2)

47 (75.8)

22 (95.7)

 

CEA (ng/mL), median (IQR)

243.6 (17.0–8325.3)

132.3 (4.8–1232.2)

1407.1 (111.5–30300.0)

0.002c

CA19-9 (LU/mL), median (IQR)

4740.0 (411.6–357855.0)

3273.0 (305.4–346927.5)

33180.0 (1018.0–550000.0)

0.306c

CA125 (U/mL), median (IQR)

45.5 (6.5–1300.3)

52.5 (9.8–1435.8)

33.0 (4.0–1070.0)

0.406c

Amylase (U/L), median (IQR)

3101.0 (112.0–25917.5)

2934.0 (116.8–38837.8)

3430.0 (100.0–11892.0)

0.318c

Cyst fluid sampling procedure (%)

  Surgery specimen

59 (69.4)

38 (61.3)

21 (91.3)

0.008a

  EUS-FNA

26 (30.6)

24 (38.7)

2 (8.7)

 

Malignant predictive value by AI, median (range)

0.068 (0.00–0.99)

0.049 (0.00–0.79)

0.928 (0.05–0.99)

<0.001c

Malignant predictive value by AI using only CEA, median (range)

0.101 (0.01–0.97)

0.072 (0.01–0.80)

0.902 (0.07–0.97)

<0.001c

  1. CEA, carcinoembryonic antigen; CA19-9, carbohydrate antigen 19-9; CA125, carbohydrate antigen 125; EUS-FNA, endoscopic ultrasound-guided fine needle aspiration; IQR, interquartile range; SD, standard deviation.
  2. pa Chi-squared test, pb Fisher’s exact test, pc Mann-Whitney U test.