Trial Information

Adjuvant α-Interferon Treatment After Resection of Hepatocellular Carcinoma in HCV-Related Cirrhosis: a Randomized Trial on Prevention of Cancer Recurrence

Background

Primary liver cancer (hepatocellular carcinoma, HCC) remains a major cancer-related cause of
death with an estimated incidence of 1 million cases per year worldwide and particular
endemic distribution related to chronic hepatitis carriers. Post-necrotic cirrhosis due to
chronic infection by HBV and HCV is the leading background for HCC development with a yearly
rate of 3% and a 5-year probability of survival of 20%.

Liver surgery, resection or transplantation, appears the only chance of curative treatment
of the tumor, but feasibility is still ranging from 15 to 30% in Western countries at time
of diagnosis. Several alternative treatments are available but their potential curative
effects are lacking, due to the absence of controlled trials.

Liver resection is claimed to be feasible in order 15-30% of patients with HCC with
mortality reported in major centers extremely low (5%). Technical and biological devices
developed in the setting of OLT can now support the limited reserve of cirrhotic livers
after resection.

Regulated segmentectomy is advisable with removal of the whole portal territory belonging to
the tumor. Single nodule tumors (< 5 cm) and compensated cirrhosis (Child class A) are
accepted as the best candidates for liver resection that now can be performed with minimal
blood loss, and minimal ischemia damage. Intraoperative ultra sound examination is now
routinely used as the golden standard for staging and detection of previously undiscovered
neoplastic nodules. Hepatic decompensation with ascites development, cholestasis and
variceal bleeding within three months from surgery are the main negative prognostic factors
for patient survival. Five years survival exceed 50% but tumor recurrences due to cirrhosis
persistence could exceed 30% after three years of follow-up. Several attempts to reduce
recurrence through antiangiogenetic and antiproliferative agents have been proposed.

In the present project a prospective randomized trial of Interferon (as antiproliferative
drug) versus control in anti-HCV positive patients will investigate in homogeneous
categories of resected tumor possible variation in the natural history and namely a
reduction of recurrence rate after curative resection.

Purpose and Study Design

The trial was intentionally designed in 1997 as an unresticted collection of patients with
histologically proved HCC and a HCV-related cirrhosis selected for surgical resection of the
tumor in 4 experienced surgical Liver Units in Italy. Eligible patients were stratified on
the basis of concomitant anti-core antibodies against HBV (anti-HBc) into two populations:
“pure HCV” (HCV-RNA:positive, anti-HBc:negative) and “mixed HCV+HBV” (HCV-RNA:positive,
anti-HBc:positive). Within each strata patients were then randomized in a 1:1 ratio to
treatment with α-interferon (IFN) vs. no treatment (control). No restriction criteria were
applied to tumor stage, as far as the resection of HCC was judged as potentially curative at
pre-operative staging, intra-operative ultrasound and post-operative pathology (clearance of
surgical margins).

The study was originally designed for lymphoblastoid α-interferon; then soon after trial
approval by the NCI-Milan Scientific and Ethic Committee (#98-016) the protocol was amended
for allowing also the use of recombinant α-2a (Roferon-A, Hoffmann-La Roche, Nutley NJ) or α
-2b interferon (Intron-A, Schering-Plough, Kenilworth, NJ). The latter was eventually the
preferred form. Patients allocated to treatment received IFN 3 MU/thrice weekly for 48 weeks
starting within 6 weeks from the operation. The severity of adverse events during treatment
were monitored and rated. Therapy was discontinued in case of life-threatening adverse event
or in case of HCC recurrence during treatment. For severe events other than anemia the IFN
dose was reduced by 50%; full dose could be resumed after the event was resolved or
discontinued if the effect persisted. The percentage of the total predicted dose of IFN
actually assumed by each patients was recorded. Patients assuming at least 80% of the total
dose of IFN and treated for at least 80% of the expected duration of therapy were a priori
defined as adherent to therapy.

Recurrences were a priori defined as “early” or “late” whether or not they occurred within 2
years from the surgical removal of the HCC. In fact early recurrence were related to
intra-hepatic metastases of the primary tumor possibly missed at the time of therapy, while
late recurrence might be due to new cancer foci related to the persistence of HCV-related
carcinogenetic factors.

Primary endpoint of this randomized clinical trial (RCT) was recurrence-free survival (RFS)
while the secondary endpoint were disease-specific survival (DSS) and overall survival (OS).
Further secondary endpoints were the assessment of IFN tolerability in post-surgical
patients and the observation of prognostic factors related to early or late recurrence.
Adherent patients were considered for a subgroup analysis focused on patients who
effectively received the treatment designed to prevent HCC recurrence.

Enrolment for the trial started on June 1998. By December 2002 the predicted sample size
of 150 randomized patients was completed (“pure HCV”: 80 and “mixed HCV+HBV”: 70). The
baseline clinical, laboratory and tumor characteristics of the two arms and viral strata
were comparable. After 45 months of median follow-up the three- and five- years survival
rates were 69% and 52.4% respectively for the control group and 77.3% and 63.6% respectively
for the IFN group (P= 0.471) . At the univariate analysis tumor multiplicity (>1 nodule)
and vascular invasion were significantly related to recurrence. After adjustment of the
relative weight of prognostic factors in a series of Cox models and cumulative incidence
curves calculated on pattern of recurrence, a benefit of IFN was observed on late recurrence
occurring in the pure-HCV patients adherent to treatment (hazard rate: 0.3; 95% Confidence
interval:0.09-0.9, P=0.04).

In conclusion, although adjuvant IFN failed to show a generalized effect on prevention of
HCC recurrence after curative resection, it effectively reduced late recurrence due to new
tumor foci in pure-HCV patient adherent to therapy (i.e.: receiving at least 80% of IFN dose
for at least 80% of time).