Skip to main content

Table 1 Baseline characteristics of the training and validation cohorts

From: Real-time automatic prediction of treatment response to transcatheter arterial chemoembolization in patients with hepatocellular carcinoma using deep learning based on digital subtraction angiography videos

Variables All patients (n = 605) Training cohort (n = 360) Internal validation cohort (n = 124) External validation cohort (n = 121)
Age (years) 55.0 ± 11.9 55.06 ± 12.2 55.5 ± 11.6 54.5 ± 11.3
Sex (male) 548 (90.6) 330 (91.7) 107 (86.3) 111 (91.7)
HBV + 530 (87.6) 306 (85.0) 107 (86.3) 112 (92.6)
BCLC stage
 A 67 (11.1) 36 (10.0) 16 (12.9) 15 (12.4)
 B 538 (88.9) 324 (90.0) 108 (87.1) 106 (87.6)
Child-Pugh score
 5 437 (72.2) 263 (73.1) 100 (80.6) 74 (61.2)
 6 106 (17.5) 62 (17.2) 15 (12.1) 29 (24.0)
 7 38 (6.3) 21 (5.8) 8 (6.5) 9 (7.4)
 8 17 (2.8) 10 (2.8) 2 (1.6) 7 (5.8)
 9 7 (1.2) 4 (1.1) 1 (0.8) 2 (1.7)
Ascites 43 (7.1) 24 (6.7) 7 (5.6) 12 (9.9)
PT (s) 12.8 ± 5.9 13.1 ± 7.6 12.3 ± 1.3 12.5 ± 1.6
TBIL (μmol/L) 18.6 ± 16.6 19.1 ± 18.6 16.7 ± 8.7 17.5 ± 13.4
ALB (g/L) 40.7 ± 23.9 41.8 ± 30.7 40.2 ± 4.7 38.0 ± 5.3
AST (≥40, IU/L) 351 (58.0) 230 (63.9) 69 (55.6) 52 (43.0)
ALT (≥40, IU/L) 311 (51.4) 198 (55.0) 69 (55.6) 44 (36.4)
CRP (≥1, mg/L) 313 (86.9) 313 (86.9) 108 (87.1) 96 (79.3)
AFP (≥200, ng/ml) 267 (44.1) 172 (47.8) 51 (41.1) 44 (36.4)
Treatment response
 Responders 335 (55.4) 176 (48.9) 69 (55.6) 90 (74.4)
 Non-responders 270 (44.6) 184 (51.1) 55 (44.4) 31 (25.6)
Combined with other treatment (yes) 388 (64.1) 237 (65.8) 81 (65.3) 70 (57.9)
Rounds of TACE (≥2) 523 (86.4) 320 (88.9) 104 (83.9) 99 (81.8)
  1. Qualitative variables are in n (%) and quantitative variables are in mean ± SD, when appropriate. HBV Hepatitis B virus, BCLC Barcelona Clinic Liver Cancer, AFP a-Fetoprotein, PT Prothrombin time, TBIL Total bilirubin, ALB Albumin, AST Aspartate aminotransferase, ALT Alanine aminotransferase, CRP C-reactive protein