BPG is committed to discovery and dissemination of knowledge
Retrospective Study
©Author(s) (or their employer(s)) 2026.
World J Radiol. Feb 28, 2026; 18(2): 116486
Published online Feb 28, 2026. doi: 10.4329/wjr.v18.i2.116486
Table 1 Clinical features of the patients
Clinical parameterTraining group (n = 80)
Validation group (n = 36)
P value
n
Response group (n = 50)
Nonresponse group (n = 30)
P value
n
Response group (n = 22)
Nonresponse group (n = 14)
P value
Sex0.6090.9560.309
    Male734528311912
    Female752532
Age0.9540.2040.908
    ≤ 60452817211110
    > 6035221315114
Number of tumors0.0050.2210.502
    ≤ 13226620146
    > 14824241688
Maximum tumor diameter0.0030.0050.212
    ≤ 8.6244341023185
    > 8.623616201349
ALBI0.7290.8780.18
    138261220128
    241231816106
    311000
CTP0.2480.9270.257
    A644222281711
    B1688853
Liver cirrhosis0.3830.3420.074
    Yes35301517136
    No4520151998
Portal vein tumor thrombus0.0230.3110.629
    Yes144521141210
    No665922104
NLR0.8620.41
    ≤ 3.37539241516115
    < 3.37541261520119
PLR0.6020.3110.4
    ≤ 114.385372215221210
    > 114.38543281514104
ALT (U/L)0.1290.5920.517
    ≤ 5046321425169
    > 503418161165
ALB (g/L)0.7290.8781
    ≤ 4042271520128
    > 4038231516106
AST (U/L)0.3130.9270.571
    ≤ 4024177853
    > 40563323281711
AFP (ng/mL)0.9480.6290.93
    < 20281810963
    20-4002515101073
    > 4002717101798
Table 2 Univariate and multivariate logistic regression analyses of the clinical features associated with transarterial chemoembolization treatment response
Clinical featureUnivariate logistic regression analysis
Multivariate logistic regression analysis
P value
OR (95%CI)
P value
OR (95%CI)
Sex0.6120.643 (0.117-3.541)
Age0.9540.973 (0.391-2.425)
Number of tumors0.0064.333 (1.512-12.416)0.0018.067 (2.276-28.594)
Maximum tumor diameter0.0034.250 (1.621-11.144)0.0017.637 (2.358-24.733)
Liver cirrhosis0.3841.500 (0.602-3.735)
Portal vein tumor thrombus0.0290.259 (0.077-0.869)
ALBI0.4091.443 (0.604-3.443)
CTP0.2521.909 (0.631-5.778)
NLR0.8620.923 (0.373-2.283)
PLR0.6030.786 (0.317-1.948)
ALT0.1312.032 (0.890-5.103)
ALB0.7291.174 (0.474-2.906)
AST0.3161.693 (0.605-4.735)
AFP0.9171.030 (0.596-1.777)
Table 3 Radiomic features for predicting transarterial chemoembolization treatment response in hepatocellular carcinoma patients by least absolute shrinkage and selection operator regression analysis
Cohort
Feature type
Feature name
FS-T2WIShape featureElongation
GLSZMLow-gray level zone emphasis
GLDMGray level variance
Arterial phaseGLCMInverse variance
First-order featureMean
Portal vein phaseFirst-order featureMean
Table 4 Predictive performance of each model
Cohort
Model
AUC
Sensitivity
Specificity
PPV
NPV
ACC
F1-score
TrainingFS-T2WI0.7710.3670.8800.6470.6980.6870.468
AP0.6680.3000.9400.7500.6910.7000.429
PVP0.7250.2330.9400.7000.6710.6750.350
JR0.8450.6000.9400.8570.7970.8130.706
RC0.9290.8330.9200.8620.9020.8880.847
ValidationFS-T2WI0.7050.4290.7720.5450.6800.6390.480
AP0.6660.3570.8640.6250.6790.6670.455
PVP0.6750.2860.9550.8000.6770.6940.421
JR0.7990.6430.8640.7500.7920.7780.692
RC0.8150.5000.9090.7780.7410.7500.609