Clinical Scorecard: Radiomic Feature Analysis via MRI for Assessing Severity in Advanced Liver Disease

At a Glance

Key Highlights

• Cirrhosis is a leading cause of global mortality with multiple etiologies and is classified into compensated and decompensated stages.
• MELD score, incorporating serum creatinine, bilirubin, INR, and sodium, is used to estimate 90-day mortality and prioritize liver transplantation.
• MRI-based radiomic analysis offers a noninvasive method to extract quantitative imaging biomarkers that may improve severity assessment beyond traditional clinical scores.

Guideline-Based Recommendations

Diagnosis

• Use MRI with standardized acquisition protocols focusing on T1-weighted images for radiomic feature extraction in cirrhosis patients.
• Employ MELD scoring system including serum creatinine, bilirubin, INR, and sodium to stratify disease severity and transplant priority.

Management

• Consider liver transplantation as the only curative treatment for advanced cirrhosis, especially in patients with MELD scores ≥15.
• Use radiomic signatures as adjunctive tools to supplement clinical and laboratory data for patient assessment and treatment planning.

Monitoring & Follow-up

• Regular imaging surveillance with MRI for patients undergoing HCC screening and cirrhosis severity assessment.
• Monitor MELD score changes to evaluate disease progression and transplant eligibility.

Risks

• Limitations in MRI radiomic feature reproducibility due to variability in scanner models, software versions, and acquisition protocols.
• Potential confounding by MRI manufacturer differences necessitates standardized imaging protocols.

Patient & Prescribing Data

Patients with advanced liver cirrhosis undergoing imaging and clinical evaluation for disease severity and transplant candidacy

Radiomic features derived from MRI may provide objective biomarkers to enhance risk stratification and guide transplant prioritization alongside MELD scores.

Clinical Best Practices

• Standardize MRI acquisition protocols including scanner model, software version, and magnetic field strength to ensure radiomic feature reproducibility.
• Integrate radiomic feature analysis with established clinical scoring systems like MELD to improve prognostic accuracy.
• Focus radiomic feature extraction on liver and spleen regions from T1-weighted MRI to capture relevant morphological and textural changes.
• Recognize the current limitation of radiomic analysis to single-center data until multi-center reproducibility is established.

References

• Global mortality from cirrhosis
• MELD scoring system and organ allocation
• Radiomics in liver fibrosis and cirrhosis
• MRI radiomic challenges and reproducibility
• Deep learning for liver fibrosis staging