treated with doxorubicin for 6 months followed by cyproterone acetate for a ... Cases 2 and 3 had been treated with cyproterone acetate for 2 and 8 months.
826 Journal ofthe Royal Society of Medicine Volume 78 October 1985
Recombinant human gamma-interferon in primary hepatoceliular carcinoma' Alastair Forbes MB MRCP Roger Williams MD FRCP
Philip J Johnson MB MRCP
Liver Unit, King's College School of Medicine and Dentistry, London SE5 9RX
Summary: A phase II trial of recombinant human gamma-interferon in the treatment of unresectable primary hepatocellular carcinoma is reported. The trial was terminated prematurely after recruitment of only 7 patients because of unacceptable toxicity, predominantly renal. There was no evidence of response in any of the 7 patients but some evidence that disease progression was more rapid during the treatment period. Introduction Gamma-interferon (IFNy, immune or Type II interferon), which is produced by activated T cells and by natural killer cells exposed to interleukin 2, is more potent and has a broader spectrum against several cancer cell lines than alpha-interferon (IFNa) (Fleischmann et al. 1984); IFNy has direct anti-tumour action inhibiting cell proliferation and causing cytolysis; it produces about fifty times greater enhancement of killer cell activity than IFNa (Fleischmann et al. 1984). IFNy is an effective agent in the control of animal tumours (Craine et al. 1978); it has been fully characterized (Rinderknecht et al. 1984) and a recombinant preparation formed after insertion of the human gene into E. coli DNA of equivalent potency to natural IFNy is now available (Immuneron: Biogen Medical Research). Phase I studies in patients with malignant disease indicated a low incidence of side effects, all reversible with doses up to 3 x 106 iu m-2 thrice weekly for eight weeks (K Sriskandan et al., in preparation). Natural killer cell activity is known to be depressed in primary hepatocellular carcinoma (HCC) and the defect in vitro is interferon-responsive (Son et al. 1982). A phase II trial of IFNy therapy in 14 patients with HCC was planned; responses were to be assessed by serial serum alphafetoprotein (AFP) levels, measurements of tumour volume and repeated Karnofsky performance scores (Karnofsky & Burchenall 1949). Such a trial has a 95% chance of detecting a 20% response rate. Unfortunately an unacceptable degree of toxicity necessitated premature termination of the trial after only 7 patients had been entered. Methods The 7 patients (5 male), mean age 48.0 years (range 30-64), had histologically proven and unresectable primary hepatocellular carcinoma. All but one female had underlying cirrhosis; 5 patients had positive serum markers for hepatitis B virus infection (HBV) (4 with HBeAg positivity, all DNA polymerase negative) and one had idiopathic haemochromatosis (Case 7). Presentation was with abdominal pain, weight loss, general malaise and hepatomegaly; 2 patients had fevers, one of whom had episodes of hypoglycaemia. The first patient had been treated with doxorubicin for 6 months followed by cyproterone acetate for a further 3 months prior to entry. Cases 2 and 3 had been treated with cyproterone acetate for 2 and 8 months respectively. In each of these, imaging and tumour markers indicated progressive disease (Figures 1 and 2). Case 4, who presented one month prior to entry, had had an unsuccessful attempt at hepatic embolization. Cases 5, 6 and 7 had presented less than two months but more than two weeks prior to enrolment and had received no previous therapy. No patient was on regular medication at the time of entry. In each case liver biochemistry was abnormal (Table 1) and serum AFP was elevated (Figure 1). Secondary deposits were demonstrable on 'Accepted 3 June 1985
0141-0768/85/10082&-04/$02.00/0
i© 1985 The Royal Society of Medicine
Journal ofthe Royal Society ofMedicine Volume 78 October 1985 827 Table 1. Clinical and laboratory data at time of entry to trial Case Age
Sex
Prothrombin time seconds prolonged Albumin Bilirubin
1 2 3 4 5 6 7
M M M M F F M
0 1 0 1 3 4 4
61 33 64 54 30 40 54
40 32 30 36 41 32 28
8 36 9 19 24 45 10
Alkaline phosphatase
AST
HBV
181 355 363 835 406 661 145
70 60 70 43 182 235 42
sAb eAg eAg eAg
eAg
Secondary site
Local Spine Bones Lung
further investigation in 4 patients (Table 1). On entry to the trial, performance scores on the Karnofsky scale ranged from 50 to 80 (Figure 3); on clinical assessment all patients were expected to survive at least two months. An infusion of 2.7 x 106 iu m-2 IFNy was administered over six hours three times weekly to a total of 12 doses (as based on Phase I study results: E Sriskandan et al., in preparation). Full blood counts and serum biochemistry were monitored at least twice weekly; liver size (cm below costal margin in midclavicular line) and serum AFP were measured weekly, and liver ultrasound scans were repeated every four weeks. Tumour volume was estimated from measurements made on ultrasound scans; one-quarter of the product of the 3 linear dimensions (anterior to posterior; right to left; superior to inferior) was used as an approximation to in r3. Toxicity was expressed according to the recommended World Health Organization grades (Miller et al. 1981). The trial had approval from the ethical committee of King's College Hospital and all patients gave written informed consent. Results In none of the 6 patients completing 12 infusions was there objective evidence of a response; the Karnofsky scores fell (Figure 3), pain was a major problem controlled only by opiates in 4
log A FP 6 5 4
31 2 1
-
8
-4
0
weeks
4
8 Figure 1. Serial alphafetoprotein levels
828 Journal ofthe Royal Society ofMedicine Volwne 78 October 1985
volume 0nr, 1
score 100
800
80
600
60
40
40
200
20 _0
-4
0 weeks
4
Figure 2. Serial estimated tumour volumes
8
-8
- weeks
Figure 3. Serial Karnofsky performance scores (O=dead)
patients and the analgesic requirement rose in all patients during the study. AFP levels remained high or rose further (Figure 1). Qualitative assessment of ultrasound scans and the estimated tumour volumes indicated tumour progression in all patients (Figure 2). All patients had pyrexia (WHO Grade 3) and rigors during every infusion, with only a little improvement during the course of treatment; rigors were frequently associated with transient hypertension (diastolic pressures up to 130 mmHg) and the pyrexia rarely settled before the next dose. There was a gradual fall in haemoglobin level from a mean (for the 6 patients) of 11.8 g dl-P' to 9.9 g dl 1 at 4 weeks (WHO Grades 1 or 2) but there were no significant changes in platelet or white cell counts except for lymphocytosis (60% of a total count of 20.2 x 10 I1-) in Case 6. In this patient there was progressive renal impairment and proteinuria from day 5 onwards. Serum creatinine rose from 49 to 159 1umol I1- at two weeks and 24-hour urinary protein loss rose to 1.6 g. There was no improvement with reduction of IFNy dose (to 1.3 x 106 iu m 2) and frequency (to twice weekly); treatment was discontinued after 7 doses on the assumption that IFN, was responsible, but there was no improvement and the patient died from hepatic and renal failure a few days later (WHO Grade 4). Significant proteinuria (WHO Grade 3) and progressive hypoalbuminaemia from 30 to 20 g P` but with only mild impairment of renal function (WHO Grade 1) occurred in Case 3, and Case 1 had proteinuria (Grade 3) but with little drop in albumin or renal dysfunction. Nausea and vomiting related to infusions was a problem in all patients and required therapy in 5 (WHO Grade 3). Two patients had severe hyperaesthesia in the feet on the infusion days but without neurological signs (WHO Grade 2). All but one required continued admission throughout the four-week treatment period because of profound malaise and impairment of conscious level (WHO Grade 1-2). Karnofsky scores fell more rapidly after entry to the trial than in the eight weeks prior to entry (Figure 3), and the rise in AFP was logarithmic in those patients whose initial levels were not already at the upper limit of the assay (Figure 1). In the 4 patients in whom a value for estimated tumour volume was available for four weeks prior to entry, the rate of increase was greater following commencement of IFNy. Patients died at 3.5, 4.5, 7, 8, 10, 11 and 16 weeks from entry to the trial, with mean and median survivals of 8.6 and 8 weeks respectively. All criteria pointed to a deterioration which was more rapid while receiving IFNy than during prior evaluation or treatment. There was no apparent difference in either disease
Journal ofthe Royal Society ofMedicine Volume 78 October 1985 829
severity on entry to the trial or its outcome between those patients who had previously received chemotherapy, and who thus might be considered to have resistant tumours, and those patients being treated for the first time. Discussion The trial was interrupted prematurely after assessment of 7 patients, because the level of toxicity was considered unacceptable. The haematological and neurological abnormalities were minor and similar to those recorded for IFNa (Gatenby & Basten 1984, Oldham 1983, Sikora & Smedley 1983). However the marked proteinuria (which is unusual with IFNa: Oldham 1983) in 3 patients was a major contributory cause of hypoalbuminaemia, and renal failure in one of these patients was the principal cause of death. Proteinuria is not usually a problem in HCC (Okuda 1976) and the high incidence in this series is attributed to IFNY. Renal failure occurs in decompensated liver disease (Wilkinson 1981), but both followed development of proteinuria and preceded hepatic decompensation in the patient described here. The dosage schedule employed was indicated by the Phase I data (E Sriskandan et al., in preparation), but although a number of the patients studied had liver metastases, there was not the very high prevalence of chronic interstitial liver disease of the present study. Such patients are apparently more sensitive to nephrotoxicity from IFNy. A study of 7 patients showing no response has 950/o confidence limits between 0 and 40% which could thus lead to erroneous rejection of an active agent. The side effects described are probably dose-related but our data suggest that primary disease progression was hastened by IFNy and further study of IFNy in HCC using lower doses did not therefore appear justified.
Acknowledgment: We thank Dr L Gerlis of Biogen Medical Research for supplies of gammainterferon. References Crane J L jr, Glasgow L A, Kern E R & Youngner J S (1978) Journal of the National Cancer Institute 61, 871-874 Fleischmann W Rjr, Kimpel G R, Tyring S K & Baron S (1984) Transplantation Proceedings 16,516-523 Gatenby P A & Basten A (1984) Medical Journal ofAustralia 140, 473-474 Karnofsky D A & Burchenall J H (1949) In: Evaluation of Chemotherapeutic Agents. Ed. M C Macleod, Columbia University Press, New York; pp 199-205 Miller A B, Hoogstraten B, Staquet M & Winkler A (1981) Cancer 47, 207-214 Okuda K (1976) In: Hepatocellular Carcinoma. Ed. K Okuda and R L Peters. Wiley, New York: pp 387-436 Oldham R R (1983) Science 219, 902 Rinderknecht E, O'Connor B H & Rodrigues H (1984) Journal ofBiological Chemistry 259,6790-6797 Sikora K & Smedley H (1983) British Medical Journal 286, 739-740 Son K, Kew M & Rabson A R (1982) Cancer 50, 2820-2825 Wilkinson S P (1981) Journal of Clinical Pathology 34, 1241-1244
