CT Radiomics to Predict Regenerative Biomarker Expression after Liver Venous Embolization: Proof-of-Concept in a...

Area of Science:

• Regenerative medicine
• Hepatology
• Medical imaging

Background:

• Liver regeneration is crucial for recovery after injury or resection.
• Understanding regenerative responses in healthy versus cirrhotic livers is vital for clinical management.
• Portal vein embolization (PVE) is a technique used to induce liver hypertrophy.

Purpose of the Study:

• To evaluate liver regeneration markers after venous embolization in healthy and cirrhotic swine.
• To assess the predictive capability of baseline computed tomography (CT) radiomics for these regenerative markers.

Main Methods:

• Twenty-two swine (11 healthy, 11 cirrhotic) underwent PVE, combined PVE and hepatic vein embolization (PVE + HVE), or sham procedures.
• Liver remnant tissue was analyzed for angiogenesis (VEGF), proliferation (Ki-67), and integrity (cytokeratins).
• CT radiomics were used to build logistic regression and random forest models to predict biomarker expression.

Main Results:

• PVE + HVE significantly increased expression of VEGF, Ki-67, angiopoietin-1, and caspase 3 compared to sham controls in both liver conditions.
• Angiogenesis was significantly higher after PVE + HVE than PVE in both healthy and cirrhotic livers.
• Ki-67 demonstrated the highest predictive performance with a random forest AUC of 0.84 and logistic regression AUC of 0.81.

Conclusions:

• Liver venous embolization elicits distinct regenerative responses in healthy and cirrhotic swine.
• Baseline CT radiomics show moderate ability to predict key regenerative biomarkers following liver venous embolization.