AI and Cancer Research ── Liver ── 2026-10-03
Liver cancer: treatment, trials and AI
- Abstract
Purpose: To develop HepaKineticNet, a deep learning framework that models inter-phase kinetic dynamics on multi-phase contrast-enhanced CT (CECT) to generate intratumoral heterogeneity (ITH) maps, and to assess in an independent external cohort whether a clinical–radiomics fusion model improves overall survival (OS) prediction in BCLC stage B–C hepatocellular carcinoma (HCC) treated with locoregional therapy plus first-line TKI and anti–PD-1 agents.
Patients and Methods: In this bi-institutional retrospective study, HepaKineticNet was developed using 359 patients from one institution (231 training, 128 validation) and evaluated in an independent external test cohort of 95 patients from a second institution. HepaKineticNet integrates a Spatial Kinetic Differential Attention module into an nnU-Net backbone, separately encoding arterial wash-in and delayed wash-out kinetics, with multiscale coefficient-of-variation self-supervision. Ten radiomics features (hierarchical filtering) and three clinical variables selected by LASSO-penalised Cox regression (10-fold cross-validation) were used to build clinical, radiomics, and fusion Cox models.
Results: A total of 277 deaths occurred, 128 of them in the development cohort. Baseline characteristics were balanced (all P> 0.05). In the external test cohort, the fusion model achieved a C-index of 0.713, exceeding the clinical model (0.651; P=0.012). Time-dependent AUCs at 1, 2, and 3 years were 0.760 (95%CI 0.651– 0.864), 0.838 (95%CI 0.736– 0.930), and 0.909 (95%CI 0.686– 1.000), whereas those of the clinical model were 0.676, 0.742, and 0.772. Kaplan–Meier analysis showed median OS of 9.5 vs 19.0 months for high- vs low-risk strata (HR 3.02, 95%CI 1.86– 4.91; P< 0.001). Discrimination was consistent across ten prespecified subgroups.
Conclusion: HepaKineticNet-derived ITH maps encode kinetic and spatial information that modestly improves individualized OS prediction and risk stratification in this combination-therapy HCC population, and may provide a noninvasive decision-support adjunct pending prospective validation. Keywords: radiomics, prognosis, prognostic model, risk stratification, convolutional neural network, immunotherapy, tyrosine kinase inhibitor, locoregional therapy, survival analysis
Journal IF-equivalent: 2.9 (OpenAlex 2-year mean citedness, value as of 2026-10-02, retrieved 2026-10-03; not the official Clarivate IF)Reference: Li J, Bu W, Hanming X, Zhang Q, Song J, Chen W. HepaKineticNet: Deep Learning of CT-Derived Intratumoral Heterogeneity for Overall Survival Prediction in BCLC Stage B-C Hepatocellular Carcinoma Treated with Locoregional Therapy Plus First-Line TKI and Anti–PD-1 Therapy. JHC. 2026 Oct;Volume 13:1-15. doi:10.2147/jhc.s636207.Checked: Abstract only - Abstract
Despite the widespread use of transarterial chemoembolization (TACE) for unresectable hepatocellular carcinoma (HCC), clinical outcomes vary substantially, and pretreatment tools for identifying patients at high risk of early progression remain limited. This study aimed to develop and externally validate an integrated CT-clinical model for predicting time to progression (TTP) after TACE. This retrospective multicenter study included 508 patients from three cohorts. The public WAW-TACE dataset ( \(n = 226\) ) was used for model development and was divided into a training cohort ( \(n = 158\) ) and an internal validation cohort ( \(n = 68\) ). Two institutional cohorts, cohort A ( \(n = 252\) ) and cohort B ( \(n = 30\) ), were used for independent external validation. Eligible patients had treatment-naive unresectable HCC treated with conventional TACE as the initial therapy, preserved liver function, no extrahepatic metastasis, no major vascular invasion, and available pretreatment contrast-enhanced CT. Clinical variables, radiomics features, and 2·5D deep learning features derived from arterial-phase CT were integrated to construct a time-to-progression prediction network (TTP-Net). TTP-Net achieved the highest observed performance among the compared models, achieving C-indices of 0.719, 0.723, 0.713, and 0.707 in the training cohort, internal validation cohort, external cohort A, and external cohort B, respectively. The corresponding 12-month AUCs were 0.872, 0.862, 0.816, and 0.832. Kaplan-Meier analysis showed significant separation between high-risk and low-risk groups in all four cohorts. Twelve-month decision-curve analysis further supported its potential clinical utility. Integration of pretreatment arterial-phase CT-derived imaging features and clinical variables enabled identification of patients with HCC at high risk of early progression after the first TACE session, supporting individualized surveillance planning and timely treatment reassessment after TACE.
Journal IF-equivalent: 5.3 (OpenAlex 2-year mean citedness, value as of 2026-10-02, retrieved 2026-10-03; not the official Clarivate IF)Reference: Wang L, Xia C, Peng Z, Zhao X, Wang Y, Wu Z, et al. Pretreatment arterial-phase CT-based prediction of time to progression after transarterial chemoembolization in hepatocellular carcinoma: a multicenter study. Cancer Imaging. 2026 Sep 30 [Epub ahead of print]. doi:10.1186/s40644-026-01136-3.Checked: Abstract only