Chimeric antigen receptor-natural killer (CAR-NK) therapy represents an emerging new direction in fighting against cancer. In recent years, it has attracted significant attention, largely due to its notable safety advantages over CAR-T cell therapy and its potential for reduced side effects. In this article, we will review the recent preclinical advances and translational challenges in CAR-NK therapy for the treatment of hepatocellular carcinoma (HCC). Several preclinical studies have successfully demonstrated that targeting HCC-associated tumor antigens such as glypican-3 and cluster of differentiation 147 (CD147) could exert a strong anti-tumor efficacy. However, only a few studies have entered the clinical stage, with none having progressed to late-phase trials. The clinical translation of CAR-NK in HCC is mainly hindered by significant challenges, including the immunosuppressive tumor microenvironment, inefficient tumor trafficking, tumor heterogeneity, and poor persistence of infused cells. To overcome these barriers, researchers have been exploring different innovative strategies such as disrupting the transforming growth factor-β signaling, engineering homing chemokine receptors, developing multi-specific CARs, and enhancing persistence with cytokine support (e.g., interleukin-15). Further ongoing research is important to optimize the CAR constructs and identify effective combination approaches to enhance the overall treatment efficacy.

Aim: To identify factors influencing liver-controlled attenuation parameter (CAP, dB/m) and establish its optimal diagnostic thresholds for hepatic steatosis grading in metabolic dysfunction-associated steatotic liver disease (MASLD).

Methods: A total of 758 prospectively enrolled MASLD patients at Zhongshan Hospital, Shanghai, China (September 2022-June 2023) were randomized into a training cohort (n = 619) and a validation cohort (n = 139), stratified by body mass index (BMI: normal-weight < 25 kg/m²; overweight ≥ 25 kg/m²). Demographics, laboratory parameters, CAP, and magnetic resonance imaging (MRI) proton density fat fraction (PDFF) were recorded. Using MRI-derived PDFF (MRI-PDFF) as the reference, receiver operating characteristic (ROC) analysis was performed to establish CAP thresholds for steatosis grades.

Results: In the training cohort, multivariate analysis showed that CAP was correlated with uric acid, glucose, and high-density lipoprotein (HDL) in normal-weight patients, and with alanine aminotransferase (ALT), total cholesterol (TC), and BMI in overweight patients. Overall cohort CAP thresholds were: S1, 283.5 dB/m [area under the curve (AUC) = 0.880]; S2, 311.5 dB/m (AUC = 0.712); S3, 328.5 dB/m (AUC = 0.799). Normal-weight thresholds: S1, 259.5 dB/m (AUC = 0.922); S2-3, 323.5 dB/m (AUC = 0.779). Overweight thresholds: S1, 304.0 dB/m (AUC = 0.829); S2, 311.5 dB/m (AUC = 0.677); S3, 328.5 dB/m (AUC = 0.767). Validation confirmed robust diagnostic performance.

Conclusion: CAP correlates significantly with metabolic factors (glucose, HDL) and shows a trend toward association with uric acid (P = 0.073) in normal-weight MASLD patients, and with liver markers/metabolic factors (BMI, ALT, TC) in overweight patients. CAP demonstrates good diagnostic accuracy for grading the severity of hepatic steatosis in MASLD, particularly for mild to moderate steatosis.

Highlights

1. Study of 758 MASLD patients stratified by BMI to identify factors influencing CAP and its diagnostic value for hepatic steatosis, using MRI-derived PDFF as reference.
2. CAP correlates with glucose, HDL, and uric acid in normal-weight patients, and with ALT, total cholesterol (TC), and BMI in overweight patients.
3. BMI-specific CAP thresholds were established, with the highest diagnostic performance in normal-weight individuals (AUC = 0.922 for S1).
4. Validation confirmed robust diagnostic accuracy (Youden index 0.81-0.83), supporting CAP as a practical non-invasive tool for grading mild-to-moderate steatosis.