Evolving thresholds for liver transplantation in hepatocellular carcinoma: A Western experience

Abstract Hepatocellular carcinoma (HCC) is the second leading cause of cancer‐related deaths worldwide. Once considered an experimental treatment with dismal survival rates, liver transplantation for HCC entered a new era with the establishment of the Milan criteria over 20 years ago. In the modern post‐Milan‐criteria era, 5‐year survival outcomes are now upwards of 70% in select patients. Liver transplantation (LT) is now considered the optimal treatment for patients with moderate to severe cirrhosis and HCC, and the rates of transplantation in the United States are continuing to rise. Several expanded selection criteria have been proposed for determining which patients with HCC should be candidates for undergoing LT with similar overall and recurrence‐free survival rates to patients within the Milan criteria. There is also a growing experience with downstaging of patients who fall outside conventional LT criteria at the time of HCC diagnosis with the goal of tumor shrinkage via locoregional therapies to become a candidate for transplantation. The aim of this review article is to characterize the various patient selection criteria for LT, discuss balancing organ stewardship with outcome measures in HCC patients, present evidence on the role of downstaging for large tumors, and explore future directions of LT for HCC.


| INTRODUC TI ON
Hepatocellular carcinoma (HCC) is the sixth most common cancer and the second leading cause of cancer-related deaths worldwide. 1,2 HCC most commonly occurs in a background of chronic liver disease with or without cirrhosis. [3][4][5] Prognosis is not only dependent on tumor burden, but on underlying liver function and patient performance status. 6 Surgical resection is the standard of care for patients without underlying liver disease and with preserved liver function who develop HCC. In select patients, 5-year survival can approach 70% for those without cirrhosis and even 60% for those with Child-Pugh A cirrhosis. However, due to heterogeneity of the patient population and low utilization of HCC screening, only 10%-37% of patients are candidates for surgical resection at initial HCC diagnosis. [7][8][9] Liver transplantation (LT) for HCC has developed dramatically over the past few decades with new criteria and prognostic factors that are continuing to evolve. In 2017, the number of liver transplants performed in the United States was at an all-time high, with 22% of all liver transplants done for underlying HCC, making HCC the most common indication for LT. 10 Liver transplantation is an especially attractive option for the treatment of HCC in cirrhotic patients as it simultaneously addresses the cancerous lesion and the underlying liver dysfunction, which is at risk for developing new HCC lesions. 11 In patients who have undergone surgical resection for HCC, the 5-year recurrence rate is significantly high, with >50% of patients exhibiting locoregional recurrence. 12,13 The most significant risk factors for recurrence after HCC resection are the presence of underlying cirrhosis and active hepatitis. 14 Patients with cirrhosis are thought to frequently have genetic alterations that represent a field defect putting the entire liver parenchyma at risk for development of cancer. 15 Compared with surgical resection, the recurrence rate of HCC after LT is estimated to be around 15%-20%. [16][17][18] Thus, given that LT both removes detectable/undetectable tumors and preneoplastic lesions that are present in the cirrhotic liver, as well as addressing the underlying cirrhotic liver parenchyma, LT is now considered the optimal HCC treatment for patients with advanced (Child-Pugh B or C) cirrhosis. Better preoperative assessment of liver function, improved accuracy of cross-sectional imaging studies, and surgical technical progress are key factors that have led to reduced mortality with LT, and 90day mortality is currently estimated at around 5%. 19 In selected patients, liver transplantation can offer an expected 5-year survival around 70%. 20 Although the benefits of LT for patients with HCC are clear, there is ongoing debate surrounding the selection of patients who would be best served by LT and how patients who initially fall outside of recognized criteria for transplant at the time of diagnosis should be managed. The aim of this review is to characterize the various patient selection criteria for LT, discuss balancing organ stewardship with outcome measures in HCC patients, present evidence on the role of downstaging for large tumors, and explore future directions of LT for HCC.

| HIS TORIC AL PER S PEC TIVE S
The first liver transplant performed for HCC was completed in 1967 by Thomas Starzl. Due to a lack of guidelines regarding organ allocation in a nascent surgical field, decision making regarding which patients with HCC received liver transplantation was at the discretion of the treatment team. Unfortunately, the lack of recognized organ allocation criteria for HCC patients led to poor outcomes and a temporary moratorium in the United States for liver transplant with HCC acts as an indication of these outcomes. Prior to the global acceptance of organ allocation criteria, the two largest series describing liver transplantation for HCC were published by Bismuth et al 21 and Iwatsuki et al. 22 In a report of the European Liver Transplant Registry, Bismuth detailed the results of 32 European centers performing a total of 1218 liver transplants, with over two-thirds of those cases done after 1984. 21 HCC accounted for 18% of LT indications in this report, with the majority of patients with HCC selected for LT because they were not candidates for resection secondary to significant tumor burden or underlying liver dysfunction. The 30-day mortality rate was 30% for all liver transplant patients and 24% for HCC liver transplant patients specifically, which was the lowest rate of any group.
The overall 2-year survival for HCC patients undergoing liver transplantation was 30%.
Iwatsuki et al 22 published an early description of the American HCC liver transplantation experience in 1985. Five-year survival rates were well under 30%. Seventy-two per cent of patients recurred, with 69% of those recurrences occurring <1 year after transplantation. The grafted liver and the lung were most commonly affected by tumor recurrence. Several cases of patients receiving LT, who shortly thereafter were found to have metastatic disease, were described. The dismal long-term survival rates in these early series were largely related to the lack of any a priori criteria for determining which patients with HCC should receive liver transplantation.

| THE DE VELOPMENT OF MODERN -DAY CRITERIA
The development of the milan criteria (MC), introduced by and recurrence-free survival rate was 83%, a dramatic improvement from pre-1996 series. This study highlighted that appropriate preoperative selection of patients was critical to improving survival after transplantation. In the 35 patients (73%) whose pathological examinations confirmed that they had tumors which met the predetermined Milan criteria, 4-year overall survival was 85% with recurrence-free survival of 92%. However, in the 13 patients (27%) who were assigned an inaccurate stage prior to LT and had tumors that exceeded those limits, 4-year overall survival was only 50% and recurrence-free survival was 59%.
The Milan criteria were quickly adopted worldwide, as well as built into a prioritization tool in the United Network of Organ Sharing (UNOS). However, as experience with LT for HCC continued to grow, critics voiced concerns that the Milan criteria may be too restrictive, excluding some patients from undergoing LT who would benefit from transplantation. Critics challenged these criteria, saying they were too strict because they excluded specific subgroups with meaningful, albeit slightly lower, opportunities to benefit from LT, as demonstrated by mostly single-institution studies. [24][25][26][27][28] It was also pointed out that cross-sectional imaging techniques had improved to enable detection of very small (<1 cm) lesions which were undetectable when Milan Criteria were first written. Tumor under-staging was observed in 20% of patients, and the 5-year recurrence-free probability for this group was 60% compared to 97% of those who were accurately staged as falling within the criteria. They estimated that the UCSF criteria offered the potential benefit of LT to an additional 5%-20% of patients with HCC who would have otherwise been excluded under the more restrictive Milan criteria, with comparable overall survival and recurrence-free survival rates.
With studies now showing that patients with larger tumors and larger total tumor size were achieving 5-year survivals similar to patients meeting the Milan criteria, there was an increased interest in exploring the survival of patients with tumors exceeding previously defined limits. Despite the pretransplant biopsy, 8% of patients exceeding MC had a poorly differentiated tumor at explant pathology. The inclusion of mandatory biopsy in the selection criteria raises some concern due to large variations in the accuracy of tumor grading ranging from 27.5%-84.8% depending on lesion size and heterogeneity 34 and the possibility of needle-tract seeding (mean incidence 3%, up to 8%) 35 (Table 1) also recently gained interest. In a meta-analysis, DCP was a useful predictive factor indicating a 5-fold increased risk for recurrence after transplantation. 37 These markers, such as AFP and DCP, may represent a way to further refine selection criteria for LT based on tumor biology.  3-, and 4-year survival rates for the Milan cohort was 89%, 76%, and 72% vs 91%, 68%, and 51% for the UCSF cohort ( Figure 4). Although survival was numerically lower for patients outside Milan criteria but within UCSF criteria vs patients within MC, the difference was not statistically significant (P = .21).

| CURRENT UNOS ALLO C ATI ON P OLI CIE S , WAITLIS T TIME S , AND ORG AN S TE WARDS HIP
While there is conflicting data for expanded criteria outside of UCSF criteria, overall these patients appear to have higher recurrence and lower overall survival. The potential benefit from LT to patients exceeding expanded criteria may be outweighed by potential harms to others on the waitlist. Further higher-quality, large-scale studies are needed before expanded criteria can be confidently adopted as standard selection protocol.

| THE ROLE OF DOWN S TAG ING
Another area of active debate is whether patients with HCC tu- Intention-to-treat survival outcomes at 1-and 5-years, respectively, were 86% and 56% in the downstaging group vs 85% and 63% in the T2 group. The 5-year overall survival was 78% when only those patients who were successfully downstaged were included ( Figure 5).
The respective 1-and 5-year recurrence-free probabilities were 95% and 91% in the downstaged group compared to 96% and 88% in the T2 group. None of these differences were statistically significant. A systematic review and pooled analysis of downstaging found the success rate of downstaging to be >40%. 45 Most studies with post-LT survival data reported 1-year overall survival rates >90% (ranging from 87%-100%). However, there was substantial variability in reported long-term survival outcomes, with some studies Mazzaferro 46 published an article in 2016 outlining an envisioned future for a multistep process from HCC diagnosis to LT in patients who are both within and beyond criteria. The proposed system utilizes tumor response to bridging or downstaging treatments as the main drivers for patient selection and allocation priority ( Figure 6). He proposes that all patients with cirrhosis who have treatable HCC by non-transplantation options should be treated upfront, regardless of whether LT is a future therapeutic option. A minimal observation period after the conclusion of a given treatment should be mandatory, as this will allow the tumor to declare its underlying tumor biology and therefore add an additional factor into the selection process. All possible information on tumor biology (AFP lab trends, tissue biopsies) should be obtained and discussed in a multi-disciplinary setting. Finally, the minimum accepted expected survival for patients undergoing LT under these conditions should be set at 60% at 5-years. As discussed previously in this review article, the 5-year >60% overall survival threshold has been determined to acceptably balance LT utility and harms to non-HCC patients on the waitlist.

| CON CLUS ION
Liver transplantation plays an important role in managing patients with HCC, providing the best opportunity for achieving long-term recurrence-free survival. Although the Milan criteria continue to be the most widely utilized transplant criteria worldwide, there is ongoing debate regarding whether the Milan criteria are too restrictive and if other expanded criteria might offer the optimal patient selection. There is also an evolving paradigm shift from "math-based" to "biology-based" selection criteria. Although some patients with smaller (UNOS T1 lesions) or larger (expanded criteria with or without downstaging) HCC lesions might benefit from LT, the benefits to these patients outside the commonly accepted selection criteria need to be balanced against the harms to other patients waiting for liver transplants given the limited availability of donors.

D I SCLOS U R E
Funding: No sources of funding were received.
Conflict of Interest: Authors declare no conflict of interest for this article.
Author Contribution: MRJ and ACY contributed equally to the preparation of this manuscript.