Research Article

Mortality Predicted Accuracy for Hepatocellular Carcinoma Patients with Hepatic Resection Using Artificial Neural Network

Table 1

Potential input variables and output variable for prognostic models.

VariablesValue   value

Input variable:
 Demographic
  Age (years)*0: ≦65, 1: >65 (mean = 57.7)0.04
  Gender0: male, 1: female0.37
 Clinical features
  Comorbidity*0: no, 1: yes0.04
  Liver cirrhosis*0: no, 1: yes<0.001
  Chronic hepatitis0: no, 1: HBV, 2: HCV, 3: HBCV0.68, 0.12, 0.48
  α-Fetoprotein (ng/mL)*0: ≦100, 1: >100<0.001
  AST (U/L)*0: ≦80, 1: >80<0.001
  ALT (U/L)0: ≦80, 1: >800.07
  Total bilirubin (mg/dL)*0: ≦1.0, 1: >1.00.01
  Albumin (g/dL)*0: >3.5, 1: ≦3.5<0.001
  BUN (mg/dL)*0: ≦21, 1: >210.01
  Creatinine (mg/dL)0: ≦1.4, 1: >1.40.24
  Platelet (103/μL)*0: >150, 1: ≦1500.02
  Prothrombin time (%)0: ≦80, 1: >800.43
  ICGR15 (%)0: ≦15, 1: >150.30
  ASA classification*0: ASA = 1, 1: ASA = 2, 2: ASA = 30.01, 0.94
  Child-Pugh classification*0: A, 1: B, C<0.001
  TNM Stage*0: I, 1: II, 2: IIIa, IIIb, IIIc, IV<0.001, <0.001
  Tumor number*0: single, 1: multiple<0.001
  Tumor size (cm)*0: ≦5, 1: >5<0.001
  Portal vein invasion*0: no, 1: yes<0.001
  Biliary invasion*0: no, 1: yes0.01
 Surgical process and outcome
  Surgical procedure*0: laparoscopic, 1: open surgery0.02
  Extent of resection0: minor, 1: major0.12
  Resection margin (mm)0: >10, 1: ≦100.08
  Surgical time (minutes)0: ≦180, 1: >1800.75
  Blood loss (mL)0: ≦1000, 1: >10000.29
  Blood transfusion0: no, 1: yes0.55
  Blood transfusion (mL)0: ≦1000, 1: >10000.07
  Postoperative complication*0: no, 1: yes<0.001
 Prognostic
  Recurrence*0: no, 1: yes<0.001
  Preoperative treatment0: no, 1: yes0.08
  Postoperative treatment*0: no, 1: yes<0.001
Output variable:
 Status0: survival, 1: dead

Significant input variables.