Multi-centre phase II trial of Thalidomide in the treatment of unresectable hepatocellular carcinoma

Abstract

Hepatocellular carcinoma (HCC) is a hypervascular tumour, which overexpresses vascular endothelial growth factor. Thalidomide is an antiangiogenic agent with activity in refractory multiple myeloma. The purpose of this multi-centre phase II study was to evaluate the efficacy and toxicity of thalidomide in patients with advanced HCC. From February 2000 to November 2001, 37 patients with histologically proven, bi-dimensionally measurable advanced, unresectable HCC were enrolled. The starting dose of Thalidomide was 100 mg per day and escalated weekly by 100 mg to a maximum dose of 800 mg/day according to individual patient tolerance. Radiological tumour response and treatment related toxicities were prospectively assessed. Thirty-seven patients were evaluable for toxicity and 24 patients were evaluable for response. The median Thalidomide dose was 400 mg/day. There was no complete response (CR). One patient had a radiological partial response (PR) (3%; 95% confidence interval [95% CI], 0% to 8%) and six (16%) patients had stable disease for more than 6 months. Somnolence and fatigue were the most common side effects, occurring in 84% and 73% of patients respectively. Thalidomide monotherapy is tolerable and associated with modest antitumour activity in advanced HCC.

Hepatocellular carcinoma (HCC) carries a significant disease burden worldwide, accounting for more than 400 000 deaths annually [1]. Of the estimated 350 000 new cases of HCC diagnosed per year, one-third occurs in China and another third elsewhere in Asia, where hepatitis B is endemic and carrier rates are high [2]. In Singapore, HCC is the third most common cancer among males, with an age-standardised rate of 18.5 per 100 000 per year. It is also the second most common cause of cancer-related death among males, with an age-standardised rate of 18.0 per 100 000 per year. The disease is strongly associated with hepatitis B related liver cirrhosis [3].

The only curative method of treating HCC to date is complete surgical extirpation. Unfortunately, this can only be achieved in approximately 10% of newly diagnosed HCC patients. This is largely because of the advanced stage of the disease and the accompanying liver cirrhosis at presentation. Even when initial resection is successful, there is still a high rate of early recurrence. Ikeda et al. [4] reported a 3-year recurrence rate of 72% after initial complete resection.

Unresectable HCC patients have a uniformly poor prognosis and present a major therapeutic challenge. If untreated, these patients have a median survival of less than six months [5]. Terminally, these patients often succumb to liver failure or acute complications of portal hypertension. Therefore, effective systemic agents are being sought for palliative treatment for unresectable HCC. In the largest trial of tamoxifen use in unresectable HCC, tamoxifen was not more effective than placebo in prolonging survival [6], [7]. Chemotherapy, either as single agent or in combination, or with interferon-α (IFN-α), achieved only very modest anti-tumour activity [8], [9]. Furthermore, approximately 60% of all HCC patients have underlying liver cirrhosis that potentially reduces the systemic clearance of cytotoxic drugs that are eliminated by the hepatic route. This often results in chemotherapy being poorly tolerated and hazardous.

Histologically, HCC is a very vascular tumour. Hepatocarcinogenesis is associated with the expression of angiogenic mediators such as vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) [10]. TNP-470, a fumagillin analogue with anti-angiogenic properties, has been shown to suppress HCC growth in animal models [11], suggesting that anti-angiogenesis may be an effective anti-tumour strategy in HCC. Antiangiogenic approaches to therapy of HCC have been shown to be feasible in rodent models [12].

Thalidomide is an oral sedative and anti-inflammatory agent introduced in the 1950s [13]. However, its use has been curtailed by its teratogenic effects, most notably phocomelia. Thalidomide inhibits angiogenesis. In animal experiments, thalidomide abrogated basic fibroblast growth factor-induced angiogenesis in the rabbit cornea micropocket assays [14]. Thalidomide also increased tumour hypoxia in the Lewis Lung Carcinoma model, presumably via an anti-angiogenic mechanism. Studies published on the efficacy of thalidomide in advanced hepatocellular carcinoma have reported modest responses to therapy with acceptable toxicity [15–18]

We undertook a multi-centre phase II study to determine the efficacy and toxicity profile of maximally tolerated doses of thalidomide in patients with advanced, unresectable HCC.

Patients and methods

Patient eligibility

To be eligible for this study, all patients were required to have pathologically confirmed HCC which were deemed unresectable because of the local extent of tumour involvement (including portal vein thrombosis or infiltration), inadequate liver reserve, comorbid medical illnesses or presence of extrahepatic metastasis or technically resectable HCC, but did not wish to undergo hepatic resection. Patients were required to have at least one site of bi-dimensionally measurable disease on CT scan/MRI scan or chest radiograph with a lesion which measured more than 1 cm in diameter, to be above 18 years old and a Karnofsky Performance Status of ≥70%. Adequate major organ functions (defined as an absolute neutrophil count ≥1.5×10⁹/L, platelet count ≥50×10⁹/L, total bilirubin <50 µmol/L, serum albumin >25 g/L, AST and ALT <5x upper limit of normal for the institution, serum creatinine <2x upper limit of normal for the institution) were also necessary. Prior systemic or regional chemotherapy more than four weeks before the time of registration was allowed provided there was clear evidence of disease progression. Exclusion criteria included concurrent treatment with other experimental drugs or anti-cancer therapy, prior or current diagnosis of other malignancies except adequately treated non-melanotic skin cancer or carcinoma in situ of the cervix, a history of significant neurological or psychiatric disorders including dementia, seizures and grade II or higher peripheral neuropathy and patients unwilling or unable to practice contraception. The respective institutional and governmental agencies approved the study and written informed consent was obtained from all patients.

Treatment plan

Treatment was delivered on an outpatient basis. The treatment consisted of two stages. In the first stage, each patient received an evening dose of thalidomide orally starting at 100 mg/day. Weekly toxicity assessments were performed and the dosage was increased by 100 mg per day every week until at least grade II toxicity was encountered. Dose escalation ceased at this stage and the next lower tolerable dose was the stable daily thalidomide dose used for efficacy and tolerability assessments in the second stage. The maximum dose was 800 mg/day. Above 400 mg/day, thalidomide was given in three divided doses/day such that at least 50% of the dose was given in the evening to minimize the sedative effects. At least 12 weeks of therapy was planned unless there was evidence of tumour progression, unacceptable treatment-related toxicities or refusal to continue with the study. Toxicity was classified according to the National Cancer Institute's Common Toxicity Criteria version 2. Dose modifications were performed for patients experiencing toxicity. Patients who experienced grade III toxicity had treatment discontinued until toxicity improved to grade I or less. Thalidomide was then restarted at one dose level below that was associated with grade III toxicity. Patients experiencing grade III toxicity at 100 mg/day or grade IV toxicity at any dose were removed from the study. Patients with grade II toxicity either continued thalidomide at the same dose or had a dose reduction of one level. Patients who discontinued thalidomide were followed up to document outcome/resolution of toxicity. All patients received appropriate supportive care including analgesics, antiemetics, antibiotics, and nutritional supplements. Corticosteroid therapy was permitted if instituted prior to study entry and continued at a stable dose. Because of the sedating effects of thalidomide, patients were advised not to consume alcohol or commence on benzodiazepenes, major tranquilisers, or sedating anti-depressants. Patients stabilized on these medications prior to entry were allowed to continue them at unchanged dose, but were warned about increased sedation.

Response evaluation

Pretreatment evaluation included a complete history and physical examination, documentation of performance status and a 12-lead electrocardiogram. Laboratory evaluation included a complete blood count, urinalysis, serum chemistries, liver function tests, partial thromboplastin time (PTT), prothrombin time (PT), baseline α-feto protein (AFP) and serum βHCG pregnancy test for female patients with childbearing potential. Radiographic evaluation included a chest radiograph or computed tomography of the chest, and a computed tomography or magnetic resonance imaging of the abdomen. A complete blood count, chemistry panel and PT/PTT were repeated weekly for the first eight weeks and four-weekly thereafter. Repeat radiographic studies were performed every eight weeks after the start of the study for tumour assessments according to the World Health Organisation criteria for tumour response [19].

Statistical methods

The primary end point of the study was radiological response. Simon's two-stage design [20] was used to calculate the sample size for this phase II trial, using two levels of response rate, P⁰ (5%) and P¹ (20%). A P⁰ is defined as a true response probability that is not of interest and a P¹ is a true response probability that is of definite interest. Accordingly, 37 patients were required for this study. Twelve patients were accrued for the first stage, and then twenty-five more were accrued once one or more responses were observed during the first stage. Standard descriptive and analytic statistics were used to report the patients’ characteristics and toxicities.

Results

Patient characteristics

Between February 2000 and November 2001, a total of 37 patients were enrolled in the study. Nineteen patients were recruited from National University Hospital, Singapore and 18 patients from Royal Prince Alfred Hospital, Sydney. While all the patients were assessable for toxicity, only 24 were assessable for both response and toxicity. Patient characteristics are summarised in Table I.

Table I.  Patient Characteristics

Characteristics	No. of patients	%
Total
Evaluable for toxicity	37	100
Evaluable for response	24	65
Gender
Male	33	89
Female	4	11
Age (years)
Mean	61.6 years	–
Standard Deviation	10.2 years
Range	37–80 years	–
Ethnicity
Asian	28	75.7
Caucasian	8	21.6
Others	1	2.7
Baseline AFP
Mean	31826.6	–
Standard Deviation	87136.7	–
Range	2.9 – 429184	–
Risk Factor for HCC
Hepatitis B virus infection	13	35.1
Hepatitis C virus infection	5	13.5
Hepatitis B and C infection	4	10.8
Unknown	15	40.5
Prior Treatment
Yes	27	73
No	10	27
Child Grading
A	24	64.9
B	7	18.9
Unknown	6	16.2

Multifocal lesions, multilobar disease or vascular involvement were the commonest causes rendering tumours unresectable. Others with more localised disease were not considered surgical candidates because of the large size of the tumours or their underlying liver cirrhosis.

Dose adjustments

All patients were started on 100 mg/day of thalidomide and escalated to a maximal dose of 800 mg/day. Only eight of 37 patients (21.6%) reached the 800 mg/day dose level. Three patients (8.1%) reached 700 mg/day, two (5.4%) 600 mg/day, three (8.1%) 500 mg/day, six (16.2%) 400 mg/day, six (16.2%) 300 mg/day, seven (18.9%) 200 mg/day and two (5.4%) 100 mg/day. The median thalidomide dose tolerated was 400 mg/day.

Response

Anti-tumour response could be radiologically assessed in 24 of 37 patients. None of the assessable patients had a CR. One patient who received a steady dose of 800 mg/day had a radiologic PR (3%, 95% CI, 0% to 8%). In addition, 6/37 (16%) patients had SD lasting more than 6 months. The estimated overall median survival was 166 days (95% CI, 192 days to 322 days). The median survival for patients who had SD or a PR was 322 days (95% CI, 175 days to 514 days). The median duration of response was 279 days (95% CI, 117 days to 466 days).

Thirteen patients were withdrawn from the study early due to various complications and could not be evaluated for response. Of these, three patients suffered grade III rash and pruritus at 100 mg/day of thalidomide (see below), one patient had grade IV vomiting complicated by acute renal failure while another had grade III hypertension. One patient had grade III hyponatraemia unrelated to thalidomide. The treatment course was complicated by subarachnoid haemorrhage in one patient and cardiac failure in another. Four patients opted to withdraw from the study after experiencing grade II toxicities (three patients had somnolence/lethargy and one had rash) and one patient died of intrahepatic haemorrhage from coagulopathy resulting from underlying hepatic dysfunction soon after starting treatment.

Regimen-related toxicity

Drug related toxicity is summarised in Table II. Toxicity was mainly grade I or grade II; grade III/IV toxicity was uncommon. No treatment related deaths occurred. Haematologic toxicity was mild with no grade III/IV neutropenia or thrombocytopenia observed. Somnolence was the most common toxicity (84%), although mild in severity. Fatigue, a frequent complication of thalidomide therapy, was observed in 27 patients (72%), with only one patient (3%) experiencing grade IV fatigue. Skin rash was seen in 24 patients (63%) and 16 (37%) suffered pruritis. Grade I/II neurologic toxicity occurred in 17 patients (45%). Constipation (61%) and oedema (47%) were also commonly observed.

Table II.  Toxicity analysis.

	No of Patients (%) (N = 37)
Toxicity	Grade I – II	Grade III – IV
Haematologic
Neutropenia	15 (41%)	0
Thrombocytopenia	2 (15%)	0
Non-Haematologic
Fatigue	26 (70%)	1 (3%)
Somnolence	31 (84%)	0
Rash	19 (51%)	5 (14%)
Pruritis	11 (30%)	3 (8%)
Neuropathy	17 (46%)	0
Constipation	20 (54%)	3 (8%)
Oedema	18 (49%)	0
Vomiting	2 (5%)	1 (3%)

Discussion

Hepatocellular carcinoma is known to be resistant to most chemotherapeutic interventions, and antiangiogenic approaches to treatment are attractive. In our study using thalidomide, a relatively weak antiangiogenic agent, a low response rate was observed, although 16% of patients experienced stable disease for more than the reported median duration of survival of this disease which, in the context of this chemoresistant disease may be clinically meaningful. The lack of cumulative toxicity would permit continuous dosing of this agent as a cytostatic agent. Indeed, antiangiogenic strategies are largely expected to be cytostatic, as demonstrated by the low response rate of single agent bevacizumab in colorectal cancer [21]. On the other hand, a combination of bevacizumab with cytotoxic chemotherapy has shown exciting potential in improving treatment outcome [22], possibly due to the alteration of tumour vasculature [23], and on this basis, it would be of interest to combine thalidomide with chemotherapy in HCC. It has been shown that low dose cyclophosphamide given in a metronomic fashion has antiagiogenic effect [24]. This low dose cytotoxic therapy is not expected to be toxic generally, and it may be feasible to explore the combination of this treatment with thalidomide in HCC.

Our finding of very modest antitumour activity of thalidomide in advanced HCC is consistent with other published studies [15–18]. While the proposition that the antiangiogenic potential of thalidomide would exploit the vascular nature of HCC thus leading to substantial antitumour activity was not borne out in this study, we had observed a radiological partial response in a patient. Further, anecdotal reports of durable clinical response of refractory HCC to thalidomide can be found in the literature [25]. The reasons behind the exciting responses in certain tumours and the lack of appreciable effect on the majority of cases are not clear at this point. Tumour genetic heterogeneity may play a role. Thalidomide was fairly well tolerated in most patients in our study and toxicities were generally of grade I or II. Not surprisingly, fatigue and somnolence featured as the commonest adverse effects as thalidomide was originally conceived as a sedative in the 1950's.

In conclusion, thalidomide has very modest single agent activity in HCC, but the non-cumulative toxicity could permit continuous dosing. Clearly, other more active agents for HCC are urgently needed, but exploration of combinations of thalidomide with chemotherapy, including metronomic schedules, appears warranted in this dismal disease. Thalidomide does not appear to have substantial anti-tumour activity against HCC save for a few exceptions. While we currently do not have any information on patient or tumour factors predicting for response, thalidomide cannot be recommended for widespread clinical use outside of a clinical trial.

This study was supported by Celgene Corporation, Warren, NJ and the Biomedical Research Council Grant No. 04/1/22/19/374.