Related Experiment Video
Updated: Sep 18, 2025
Author Spotlight: Investigating Immune Cell Dynamics in the Tumor Microenvironment — Challenges and Innovations in Cancer Prognosis
Published on: April 12, 2024
Gadoxetic Acid-Enhanced Magnetic Resonance Imaging Features Can Predict Immune-Excluded Phenotype of Hepatocellular
Eisuke Ueshima1, Keitaro Sofue1, Takahiro Kodama2
1Department of Radiology, Kobe University Graduate School of Medicine, Kobe, Japan.
Introduction:
Immunotherapy is the first-line treatment for intermediate-advanced stage hepatocellular carcinoma (HCC), although its outcomes vary. This study aimed to identify imaging biomarkers of immunotherapy susceptibility linked to gadoxetic acid-enhanced magnetic resonance imaging (EOB-MRI) and immune phenotypes, particularly immune-excluded phenotypes, with a tumor immune barrier.
Methods:
We performed immunohistochemical staining with a CD8+ antibody, and samples were classified into immune-inflamed, -intermediate, -excluded, and -ignored phenotypes. We assessed EOB-MRI findings obtained from 104 patients who underwent hepatectomy for HCC and evaluated the relationship between MRI findings and immune phenotype. Spatial transcriptome analysis of tumor tissues in each immune phenotype was performed to characterize the MRI findings. For validation, we analyzed the treatment effect on 60 nodules in another cohort of 27 patients who received combined immunotherapy using anti-programmed death-ligand 1 and anti-vascular endothelial growth factor (VEGF) antibodies.
Results:
HCCs with rim arterial phase hyperenhancement (APHE) (odds ratio [OR] 17.3, p = 0.009), peritumoral enhancement in the arterial phase (OR: 8.6, p < 0.004), and intermediate intensity on the hepatobiliary phase (HBP) measured with a visual 3-point scale (OR: 28.2, p = 0.002) were associated with immune-excluded phenotype, where tumors tended to be larger and of the single nodular type with extranodular growth and confluent multinodular rather than the simple nodular type. Spatial transcriptome analysis revealed a spatial relationship among cytotoxic T lymphocytes, VEGF signals, and cancer-associated fibroblasts at the tumor-invasive margins in this phenotype. From the validation study, nodules with any one of these three imaging findings had a significantly prolonged time to-nodular progression (p = 0.007, median not reached vs. 226 days).
Conclusion:
HCCs with rim APHE, peritumoral enhancement in arterial phase, and intermediate intensity on HBP with visual 3-point scale could be non-invasive biomarkers to predict the immune-excluded phenotype with the tumor immune barrier. These HCCs were most likely to respond to combined immunotherapy.
