# Hepatocellular Carcinoma Transplant Criteria Show Poor Negative Predictive Value: A Systematic Review and Meta-Analysis

## TL;DR

Current criteria for selecting liver transplant candidates for hepatocellular carcinoma are good at including low-risk patients but poor at excluding high-risk ones.

## Contribution

A systematic review and meta-analysis comparing the diagnostic performance of HCC liver transplant criteria.

## Key findings

- All criteria had an adjusted positive predictive value over 80%.
- Most criteria had an adjusted negative predictive value below 50%.
- A ranking system was proposed to address stakeholder concerns in LT eligibility.

## Abstract

Background/Objectives: Multiple criteria are used worldwide to select hepatocellular carcinoma (HCC) patients with a low risk of recurrence for liver transplantation (LT). However, it remains unclear which criteria are best for the LT-involved stakeholders, particularly in accurately identifying patients at high risk of recurrence. This work aimed to identify the most accurate criteria for selecting HCC patients for LT. Methods: In June 2023, a systematic literature search was conducted in PubMed and CENTRAL to identify studies including LT selection criteria of HCC patients. Data was extracted from recurrence-free survival curves using a validated algorithm and subsequently used to calculate measures of diagnostic performance routinely used in clinical trials. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines were applied. Results: Of 815 records screened, only 17 met our study inclusion parameters, embodying 14 LT selection criteria. All LT criteria achieved an adjusted positive predictive value (aPPV) over 80%, indicating the correct selection of at least 80% of low-risk HCC patients. However, the adjusted negative predictive value (aNPV) was below 50% in most cases, indicating that these criteria cannot correctly identify patients with a true high risk of recurrence. This raises major ethical concerns regarding the models’ ability to exclude patients from LT. Since a perfect model is nonexistent, we created a ranking to account for the distinct concerns of all stakeholders in LT eligibility in the context of HCC. Conclusions: These results highlight the urgent need for refined or newly developed criteria with improved specificity and NPV to select more patients amenable to LT who are currently excluded.