Table 1 Characteristics of the study population.

From: Deep learning-based prediction of post-pancreaticoduodenectomy pancreatic fistula

Variable

N = 1333

Age (years)

63.4 ± 10.6

Sex (male)

797 (59.8)

Body mass index (kg/m2)

23.7 ± 3.3

Underlying disease (hypertension/diabetes mellitus/cerebrovascular accident)

399 (29.9)/272 (20.4)/30 (2.3)

Laboratory data

 White blood cell (103/μL)

6220 ± 1964

 Hemoglobin (g/dL)

12.3 ± 4.9

 Total bilirubin (µmol/L)

1.3 ± 2.2

 Albumin (g/L)

3.6 ± 6.2

 Glucose (mg/dL)

138.2 ± 64.1

 Creatinine (mg/dL)

0.7 ± 0.4

Operative indication

 Pancreatic ductal adenocarcinoma

531 (39.8)

 Other malignancies

544 (40.8)

 Low-grade malignancies

193 (14.5)

 Benign disease

65 (4.8)

Neoadjuvant chemotherapy before surgery

88 (6.6)

Pancreatic duct size (mm)

3.8 ± 2.0

Pancreatic texture (soft/hard/firm/unknown)

844 (63.3)/207 (15.5)/267 (20.0)/15 (1.1)

Pylorus preservation

862 (64.7)

Concurrent vessel resection

228 (17.1)

Operative time (min)

330.9 ± 88.2

Intraoperative transfusion

239 (17.9)

Postoperative pancreatic fistula (biochemical leakage/grade B/grade C)

463 (34.7)/83 (6.2)/3 (0.2)

Other postoperative complications

149 (11.2)

  1. Unless otherwise indicated, data presented are the means, with standard deviation in parentheses.