Clinical features and outcomes of hepatocellular carcinoma in Caucasian cirrhotic patients on long‐term analogue therapy for hepatitis B

Long‐term oral nucleos(t)ide analogue (NUC) therapy in hepatitis B virus (HBV)‐related compensated cirrhotics prevents clinical decompensation but not hepatocellular carcinoma (HCC) development.


| INTRODUCTION
Hepatitis B virus (HBV) is a major health problem worldwide, with roughly 240 million people with chronic infection. 1 If left untreated, patients with HBV infection are at high risk of progression to cirrhosis, clinical decompensation, hepatocellular carcinoma (HCC) and liver-related death. [2][3][4] International guidelines recommend either a short course of pegylated interferon (Peg-IFN) or the long-term administration of third-generation nucleos(t)ide analogues (NUCs), such as entecavir (ETV) or tenofovir disoproxil fumarate (TDF) or tenofovir alafenamide, to stop the progression of liver disease. [2][3][4] Studies with TDF and ETV have clearly shown reversal of clinical decompensation, 2 prevention of clinical decompensation, as long as HBV is the only cause of liver damage, [5][6][7][8][9] and regression of grade 1 oesophageal varices in most of the patients (83%) with a negligible risk of de novo occurrence of oesophageal varices. 10 On the other hand, chemoprevention of HCC by long-term administration of ETV or TDF is still a matter of debate. In ETV-treated patients with cirrhosis, the annual incidence of HCC ranged from 2% to 4.1% in Asian studies 6,[11][12][13] and is 2.6% in European studies, whereas in European studies of TDF therapy of patients with cirrhosis HCC, risk ranged from 3.7% to 4%, 8,9,13 that is, very similar to the estimates of the course of HBV in untreated patients. 14 Because the early diagnosis of HCC increases indications for curative therapies with clear implications on prognosis, identification and close surveillance for the risk of HCC is of strategic importance, 2,3 and this is also true for ETV or TDF-treated cirrhotic patients. In fact, HCC was the only cause of liver-related death in the 1269 patients without cirrhosis, compared with 503 patients with cirrhosis where 50% died of liver disease (HCC in 75% of cases) and 50% died of nonliver-related causes. 15 The development of HCC is a major event affecting all causes mortality, whereas it stands as the only factor affecting liver-related mortality in HBV patients under NUCs.
Prospective studies with a follow-up long enough to assess the outcome of patients who developed HCC during anti-HBV therapy are lacking.

All consecutive de novo HCCs diagnosed between 2005 and 31
December 2016 in patients with a HBV-related liver disease on NUC therapy at our outpatient clinic, in 2 different hospitals in Milan-Fondazione IRCCS C a Granda Ospedale Maggiore Policlinico and Ospedale San Giuseppe, were enrolled in this retrospective study. Exclusion criteria were (1) HCC detected at baseline or occurring within 6 months after starting NUC therapy, (2) HIV and HDV coinfection, as well as autoimmune hepatitis, while co-factors of disease have not been excluded (i.e. alcohol intake, overweight and diabetes). Data on HCC incidence and predictors in NUC-treated patients from our Centres have been previously published. 8

| Surveillance and treatment
The initial anti-HBV treatment was lamivudine or ETV or TDF as monotherapy. For lamivudine-resistant patients, adefovir 10 mg/d was added from 2003 and switched to TDF 245 mg/d from 2008. [6][7][8] Surveillance for HCC included abdominal ultrasound and AFP every 6 months in cirrhotics, having Child-Pugh Turcotte score A or B, or Child-Pugh Turcotte score C if they could potentially have a liver transplant, according to international recommendations. 17 Six-month surveillance with AFP monitoring was performed also for patients with advanced fibrosis (Ishak score [4][5]. HCC was diagnosed per 2005 American Association for the Study of Liver Diseases criteria until 2010 and thereafter per the updated criteria 18,19 using contrast imaging techniques like contrast-enhanced ultrasound, computedtomography (CT) scan and magnetic resonance imaging (MRI), whenever the "typical" vascular pattern for HCC of wash-in in the arterial phase and wash-out in the portal or equilibrium phase was demonstrated. An ultrasound-guided fine-needle biopsy was performed to restore sensitivity in nodules escaping radiological diagnosis, using a Radiological response was evaluated by expert radiologists: complete response was disappearance of any enhancement of the target lesions, partial response was ≥30% decrease of the sum of diameters of viable target lesions, stable disease was any case that did not qualify for previous definitions, and progressive disease was ≥20% increase in the sum of viable target lesions (including emergence of new lesions). Patients achieving a radiological complete response had a CT or MRI repeated every 3 months for 24 months and subsequently were switched to US surveillance at 6-month intervals. Retreatment of patients who failed to achieve a complete response was decided by multidisciplinary clinical team as follows: RFTA was repeated once in patients with incomplete tumour necrosis; TACE was repeated until complete response was achieved, or progression demonstrated. Sorafenib dose was reduced by 50% in the presence of grade 2 toxicity, whereas it was discontinued whenever tumour progressed or grade 3-4 side-effects did not improve following 7-day discontinuation. 22 A second-line treatment was proposed by multidisciplinary clinical team for patients with tumour progression, or unfit to further cycles of first-line treatment due to technical constraints, like failure of catheterisation of the feeding artery or ethanol diffusion, residual tumour not suitable to be managed with the first-line treatment, or tumour recurrence. The same policy was adopted for third-and forth-line treatments. Patients were followed until death or last available clinical evaluation, up to 31 July 2017. Patients who missed a scheduled visit were contacted by phone call to collect information on clinical status and outcome.

| Definitions
Cirrhosis was diagnosed by histology or on clinical grounds by using an abdominal ultrasound features of blunted, nodular liver edge accompanied by splenomegaly (>13 cm) and <100 9 10 3 /lL platelets. Persistent HBV suppression meant HBV DNA persistently under the lower limit of quantification of the test used throughout the follow-up.
Rescue transplant was offered to all patients failing first-line therapy for HCC or with recurrence after complete response to treatment with the following conditions: extension of the tumour according to the Milan criteria and age less than 65 years up to 2013 and less than 70 years later.

| Study endpoints
The primary endpoints of the study were clinical features of HCC and alpha-fetoprotein patterns. Secondary endpoints were the response to treatments, development of early and late recurrence after therapy for HCC, and survival.

| Statistical analysis
Data were expressed as counts and percentages for qualitative variables and as median and range for continuous discrete variables. Significance of differences in the distribution of quantitative and qualitative variables was assessed with Student's t test, Kruskal-Wallis, Fisher's exact or chi-square tests as appropriate. All P values were 2-tailed and a level of 0.05 was considered statistically significant. The Kaplan-Meier method was used to estimate outcome rates.
The log-rank test was used to compare curves between patient groups. The cumulative incidence of HCC was estimated by competing risks framework, with death and liver transplantation as

| RESULTS
The study population included 76 subjects after the exclusion of one patient with a HCC developed after liver transplantation, one with a tumour detected outside regular surveillance and 2 patients with a tumour detected within the first 6 months of NUC therapy.
The demographical and clinical features of the enrolled patients are depicted in HCC was radiologically diagnosed in 65 (86%) patients and histologically in the remaining 11 (14%): 59 patients (78%) had a single tumour node, and the median tumour size was 20 mm (range: 6-57), AFP ≤7 ng/mL in 64% of patients, within "Milan criteria" in 71 (93%) resulting in BCLC stage 0/A in 70 (92%) ( Table 2). Among the 5 patients with a more advanced HCC stage, no difference in number of co-factors of HCC aggressiveness has been identified compared with those patients with earlier stages (data not shown).

| Treatment algorithm
The first-line therapy was any potentially curative treatment in 59 (78%) patients: 30 (59%) were treated by RFTA, 21 (41%) by surgical resection (including the 6 patients without cirrhosis) and 8 patients were listed for LT as first treatment approach. In these patients listed for LT, 1 patient was staged BCLC D for liver decompensation,   and multifocality (1). Among the 12 BCLC 0/A patients treated by TACE, complete response was obtained in 6 (50%) patients.  Focusing on the treated groups of these Asian studies, the Contrasting to these and our findings, 2 other studies in Asia in patients treated long-term with NUC therapy showed high positive predictive value for HCC development if AFP levels persistently or abruptly elevated more than 12 or even 6 ng/mL. 25,26 All studies, in agreement with our findings, showed a high specificity of AFP increase during NUC therapy as long as the ALT levels were normal.

| Survival
There are a few limitations in this study, especially the lack of an untreated control group. However, after the pivotal trial on lamivudine in 1997 showing the clinical benefits of NUC therapy in patients with advanced HBV, 40 it is unethical to keep cirrhotic patients untreated, at least in high-income countries. On the other hand, it would be hard to compare these data to those of historical cohorts, due to significant progresses achieved in surgical and locoregional treatments in the last decades. A second limitation is the sample size, even though this is to date the largest cohort for the analysis of HCC management and survival in Caucasian NUCtreated patients. Third, the study was conducted in a tertiary referral, academic centre, where a large expertise in the management of these patients accumulated over the last 30 years. Finally, our patients were homogeneous in terms of geographical origin, ethnicity, HBV genotype, and disease severity, characteristics that may limit the applicability of our finding in other regions. At the same time, the study has several strengths. It describes the largest cohort study to date, patients were managed according to criteria agreed upon internationally, the endpoint was as strong as survival could be, and patients were homogenous and managed by a single-centre unit.
In conclusion, this study demonstrates that the vast majority of HCC developing in Caucasian compensated cirrhotic patients long-LOGLIO ET AL.
| 437 term NUC treated are small, Milan-in, BCLC 0/A lesions amenable of potentially curative treatments with substantial survival benefits.
While acknowledging that these data must be confirmed in other independent cohorts, this study sheds new lights on a relevant topic given that HCC is almost the only complication observed in patients with HBV permanently suppressed by NUC.