Clinical Report: SELECT Scoring System for HCC Risk in Viral Hepatitis Patients

Overview

The SELECT scoring system was developed and validated to predict hepatocellular carcinoma (HCC) risk in patients with viral hepatitis using ultrasound (US) features combined with clinical and laboratory data. This model demonstrated improved discrimination for identifying high-risk patients eligible for intensive surveillance with alternative imaging modalities compared to existing risk scores.

Background

Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality worldwide, primarily due to late-stage diagnosis. Surveillance using ultrasound (US) is standard but has limited sensitivity for early HCC detection. Alternative imaging modalities offer better detection but are costly and less accessible, necessitating selective use in high-risk patients. Existing HCC risk models rarely incorporate imaging features, prompting the development of the SELECT model integrating US findings with clinical and laboratory parameters to enhance risk stratification in viral hepatitis patients.

Data Highlights

The study retrospectively analyzed patients with chronic hepatitis B or C aged 40-75 years undergoing biannual US surveillance across three institutions. Key ultrasound features assessed included coarse or cirrhotic echotexture, hepatic steatosis, cirrhotic nodules, ascites, and splenomegaly. The SELECT model was developed using data from institution 1 (n=development and internal validation cohorts) and externally validated with data from institutions 2 and 3. Model performance was evaluated by Uno C-index, calibration plots, and comparison with five established HCC risk scores. Inter-observer agreement for US features was assessed using Fleiss kappa statistics.

Key Findings

• The SELECT model incorporated significant ultrasound features alongside clinical and serologic parameters to predict 5-year HCC risk.
• It demonstrated superior discrimination (Uno C-index) compared to existing risk scoring systems such as aMAP, THRI, ADRESS-HCC, Velazquez score, and mPAGE-B.
• Two optimal cut-offs stratified patients into low-, intermediate-, and high-risk groups, with the high-risk group having an annual HCC incidence ≥ 2.5%, suitable for intensive surveillance.
• Inter-observer agreement for key US features included in the model ranged from moderate to substantial, supporting reproducibility.
• The model enables selective use of costly alternative imaging modalities by identifying patients most likely to benefit from intensified surveillance.

Clinical Implications

The SELECT scoring system provides clinicians with a validated tool to stratify HCC risk in viral hepatitis patients by integrating ultrasound imaging features with clinical and laboratory data. This facilitates targeted surveillance strategies, optimizing resource utilization by reserving intensive imaging modalities for patients at highest risk. Implementation may improve early HCC detection rates and patient outcomes while maintaining cost-effectiveness.

Conclusion

The SELECT model enhances HCC risk prediction by incorporating ultrasound features, outperforming existing scoring systems and enabling more precise identification of viral hepatitis patients who would benefit from intensified surveillance. This approach supports personalized surveillance strategies to improve early HCC detection.

References

1. Global Cancer Statistics 2020 -- Primary Liver Cancer Epidemiology
2. Surveillance Recommendations for HCC -- Early Detection Importance
3. Limitations of Ultrasound in Early HCC Detection
4. Alternative Imaging Modalities for HCC Surveillance
5. Cost-effectiveness of Selective Intensive Surveillance
6. Existing HCC Risk Scoring Systems -- aMAP, THRI, ADRESS-HCC, Velazquez, mPAGE-B
7. Inter-observer Agreement Metrics -- Fleiss Kappa Interpretation