AI and Cancer Research ── Liver ── 2026-10-17
Liver cancer: treatment, trials and AI
- Abstract
Background and purpose Transcatheter arterial chemoembolization (TACE) is the primary treatment for intermediate-stage hepatocellular carcinoma (HCC). However, significant heterogeneity in patient response to TACE necessitates non-invasive and precise methods for evaluating treatment efficacy. This study aims to explore the predictive ability of combined intratumoral habitat imaging and peritumoral radiomics features for TACE response. Materials and methods We retrospectively enrolled HCC patients who underwent TACE treatment at four centers between January 2015 and December 2023. All patients underwent contrast-enhanced MRI. Patients from centers A and B formed the training cohort, while those from centers C and D comprised external validation cohorts. Using three-phase enhanced MRI (arterial, portal venous, and delayed phase), K-Means clustering divided intratumoral regions into three habitat subregions (Habitat 1, 2, and 3), and peritumoral features were extracted. Clinical variables were also collected. Key radiomics features were selected by LASSO regression, and clinical indicators were identified through univariable and multivariable logistic regression. The XGBoost algorithm constructed intratumoral, peritumoral, clinical, and combined models. Performance was evaluated using ROC curves, calibration curves, decision curve analysis (DCA), and DeLong tests.
Results: A total of 174 HCC patients who underwent initial TACE treatment were enrolled, including 104 in training and 70 in validation cohorts. Fifteen key intratumoral features from the habitat subregions and 24 radiomics features from the peritumoral region were identified, primarily derived from wavelet transforms and texture analysis. The wavelet-LLL_glszm_ZoneVariance_ PVP feature from Cluster 2 contributed significantly to the model and had a positive impact (highest SHAP value). Alpha-fetoprotein (AFP) (P = 0.015) and Tumor size (P = 0.016) remained significantly associated with TACE treatment response in multivariable analysis. The combined model based on clinical, intratumoral habitat and peritumoral radiomics showed superior predictive performance, with AUCs of 0.950 (95% CI: 0.89-1.00), 0.850 (0.79-0.91) and 0.830 (0.77-0.89) in training and validation cohorts, outperforming single models (Clinic: AUC = 0.700, Tumor_Habitat: AUC = 0.880, Peritumor_Radiomics: AUC = 0.870, P<0.05). Calibration and DCA demonstrated good agreement between predicted and observed outcomes and favorable net benefit across the evaluated threshold probabilities.
Conclusion: The combined model integrating AFP, Tumor size, intratumoral habitat imaging, and peritumoral radiomics could accurately predict initial TACE response in HCC patients. Cluster 2-derived features showed relatively greater contributions among the intratumoral habitat subregions. This non-invasive prediction tool may provide complementary information for pretreatment risk stratification. Future prospective multicenter studies are needed to verify the clinical applicability of the model.
Journal IF-equivalent: 2.7 (OpenAlex 2-year mean citedness, value as of 2026-10-02, retrieved 2026-10-03; not the official Clarivate IF)Reference: Wang Z, Yin N, Su R, Zhou M, Sun Z, Li J, et al. Intratumoral habitat combined with peritumoral radiomics to improve the predictive performance of initial TACE efficacy: a multicenter retrospective study. Front Oncol. 2026 Oct 14;16:1772206. doi:10.3389/fonc.2026.1772206.Checked: Abstract only - Abstract
Background: This study developed and validated an machine-learning-based radiomics model integrating tumor habitat radiomics, peritumoral radiomics, and clinical features to predict progression-free survival (PFS) following repeat transarterial chemoembolization (TACE) in patients with hepatocellular carcinoma (HCC).
Methods: This retrospective multicenter study included 270 patients with HCC. Patients from Center 1 were assigned to training (n = 94) and internal validation (n = 41) cohorts, while Centers 2 (n = 82) and 3 (n = 53) served as external validation cohorts. Tumor habitats and 5-mm peritumoral regions were generated from contrast-enhanced CT images. Radiomics features were selected using univariate Cox and LASSO-Cox regression. A fusion model and HPC risk score were constructed by integrating selected radiomics signatures and clinical variables. Performance was assessed using time-dependent ROC, C-index, calibration, decision curve analysis, and Kaplan–Meier analyses.
Results: The optimal habitat cluster number was K = 2. Eleven habitat features, three peritumoral features, and alpha-fetoprotein were retained in the fusion model. The model achieved AUCs of 0.923, 0.921, and 0.897 for 3-, 6-, and 9-month progression-free survival in the training cohort, and 0.932, 0.863, and 0.916 in the internal validation cohort, respectively; all external AUCs exceeded 0.82. The fusion model showed higher C-index values than the clinical model across cohorts. High-risk patients had significantly shorter PFS than low-risk patients (all P < 0.001). Calibration and decision curve analyses demonstrated good agreement and clinical utility.
Conclusion: The CT-based fusion model and HPC risk score effectively predicted progression risk in HCC.
Journal IF-equivalent: 5.2 (OpenAlex 2-year mean citedness, value as of 2026-10-15, retrieved 2026-10-16; not the official Clarivate IF)Reference: Wang F, Hu Y, Shi ZX, He XY, Sui YH, Wen FK, et al. CT habitat and peritumoral radiomics for predicting progression after repeat TACE in hepatocellular carcinoma: a multicenter study. Front Pharmacol. 2026 Oct 14;17:1942002. doi:10.3389/fphar.2026.1942002.Checked: Abstract only