Including mRECIST in the Metroticket 2.0 criteria improves prediction of hepatocellular carcinoma-related death after liver transplant.

BACKGROUND AIMS: In the context of liver transplantation (LT) for hepatocellular carcinoma (HCC), prediction models are used to ensure that the risk of post-LT recurrence is acceptably low. However, the weighting that 'response to neoadjuvant therapies' should have in such models remains unclear. Herein, we aimed to incorporate radiological response into the Metroticket 2.0 model for post-LT prediction of "HCC-related death", to improve its clinical utility. METHODS: Data from 859 transplanted patients (2000-2015) who received neoadjuvant therapies were included. The last radiological assessment before LT was reviewed according to the modified RECIST criteria. Competing-risk analysis was applied. The added value of including radiological response into the Metroticket 2.0 was explored through category-based net reclassification improvement (NRI) analysis. RESULTS: At last radiological assessment prior to LT, complete response (CR) was diagnosed in 41.3 partial response/stable disease (PR/SD) in 24.9PD) in 33.8-year rates of "HCC-related death" were 3.1 9.63.4 PR/SD, or PD, respectively (p textless0.001). Log(10)AFP (p textless0.001) and the sum of number and diameter of the tumour/s (p textless0.05) were determinants of "HCC-related death" for PR/SD and PD patients. To maintain the post-LT 5-year incidence of "HCC-related death" textless30 the Metroticket 2.0 criteria were restricted in some cases of PR/SD and in all cases with PD, correctly reclassifying 9.4related death", at the expense of 3.5net NRI was 5.8. CONCLUSION: Incorporating the modified RECIST criteria into the Metroticket 2.0 framework can improve its predictive ability. The additional information provided can be used to better judge the suitability of candidates for LT following neoadjuvant therapies. LAY SUMMARY: In the context of liver transplantation for patients with hepatocellular carcinoma, prediction models are used to ensure that the risk of recurrence after transplantation is acceptably low. The Metroticket 2.0 model has been proposed as an accurate predictor of "tumour-related death" after liver transplantation. In the present study, we show that its accuracy can be improved by incorporating information relating to the radiological responses of patients to neoadjuvant therapies.