Modeling Cancer-specific Prognosis in Liver Transplantation for Hepatocellular Carcinoma (HCC)

Rationale: Liver transplantation (LT) is one of the curative treatment options for patients with HCC associated to chronic liver disease (cirrhosis). All current international guidelines recommend LT for HCC only within pre-defined criteria.

The use of restrictive criteria to select patients affected by HCC for LT was originally proposed with the Milan criteria. These criteria were based on pathologic assessment of number and size of the HCC nodules on the explanted liver. Subsequently, many authors proposed the expansion of such restrictive criteria (e.g. UCSF, Tumour Volume, Up-To-Seven etc.). All these attempts, based on different combinations of morphologic parameters, have been defined on the pathologic staging of the tumor made on the removed liver, namely after LT, once decision on treatment and treatment itself could not be changed.

Over time, various attempts have also been made to correlate postoperative pathologic findings of HCC to preoperative clinical staging in order to anticipate decision-making. Should a reliable conversion algorithm correlating pre-operative staging with post-transplant prognosis be available a selection of LT candidates based on routine radiologic HCC findings would optimise survival patient survival and organ resource allocation.

Although post-LT pathology / pre-LT radiology correlation have improved over time, significant biases still affect clinical assessment of HCC stage and no reliable protocols has entered clinical practice yet. In addition, robust evidence indicates that other biological markers of aggressiveness (such as α-Fetoprotein levels and clinical response to bridge therapies) have to be added when evaluating pre-operative variables.

Another important issue to be considered when dealing with LT for HCC is the striking progress in treating chronic viral infections (HCV and HBV) that has been achieved over the last few years. Due to the possibility of treating recurrence of viral infections in the transplanted liver, these conditions do not significantly affect mortality following LT as they did in the past.

In the current scenario, the decreasing role of non-oncologic factors in survival of patients treated by LT for HCC enhances the need to define new prognostic models oriented towards cancer-specific survival. These models should include, besides conventional morphological parameters, also the response to bridge therapies delivered in the pre-transplant setting as well as other biologic markers commonly used in clinical practise such as α-Fetoprotein levels.

Although many studies have been conducted, prognostic calculators of cancer-specific survival for HCC patients undergoing an evaluation for LT are not yet available. Such models should be able to provide survival estimates based on pre-treatment oncologic variables.

A tailored assessment of cancer-specific prognosis is strongly advocated also by regulatory authorities, to optimize the allocation/distribution criteria of the limited source of available organs. Therefore, these criteria could be useful in the daily practise of transplant Centres to define different priority levels within the waiting list for LT, both for patients with and without HCC.

Endpoints and clinical relevance: The main goal of the study is the definition of a cancer-specific prognostic model based on pre-operative features (radiologic staging and α-Fetoprotein levels) of a wide population of patients who underwent LT for HCC.

Considering the competitive risk of cancer-specific mortality and death due to other causes, the investigators aim to redefine the Up-To-Seven criteria, as they were developed on the base of pathologic staging.

Primary endpoints: besides conventional outcome endpoints, the study aims to develop and validate a prognostic calculator of recurrence rate and cancer-specific survival for HCC patients undergoing evaluation for LT.

The prognostic algorithm will be based on the competitive risk analysis of cancer-related recurrence and survival vs. non-cancer related outcome.

Clinical relevance: the development of an on-line available prognostic calculator based on pre-operative oncologic factors would provide precise estimates of survival. This could help in the comparison of patients with different disease stage and comorbidities, by defining progressive priority levels to be applied to the waiting list for LT.

The final goal is to provide reliable survival estimates for patients at different disease stage and therefore with different priority.

This perspective may have great clinical impact as would allow the definition of different priority levels for LT in HCC, both to provide a prognosis-oriented treatment for each patient and to optimise the global outcome of the population of patients eligible for LT.

Observational period: Retrospective analysis of a population of patients who underwent LT for HCC from January 2000 to December 2015.

Study population size: The expected number of patients from Italian centres is more than 1000 cases.

If the training set will allow the development of a cancer-specific prognostic model, at least 300 other cases will have to be added as a further set of external validation, preferably belonging to an Eastern series (see above).