Journal of the Royal Society of Medicine Volume 83 September 1990 readmitted in hypoglycaemic coma. During this admission she remained persistently hypoglycaemic, requiring regular glucose infusions, despite maimal doses of diazoxide, increased calorie intake and a further course of streptozotocin. Glucagon temporarily- elevated her blood glucose, but a continuous infusion had to be discontinued due to intolerable nausea. In order to control her symptoms, selective hepatic artery embolization using 0.5 mm pieces of collagen was performed. Pre-embolization angiography revealed an extensive tumour circulation which was significantly' reduced after embolization. Pain, fever, nausea and a transient rise in transaminases complicated the first 24 h after the procedure but by 72 h normoglycaemia was restored and all treatment was withdrawn. Fifteen months later her symptoms returned and she also complained of pruritis. On examination she was jaundiced with massive hepatomegaly and ascites. Once again she was refractory to conventional therapy and, in addition, the longacting somatostatin analogue SMS 201-995 failed to improve her blood sugar. Repeat hepatic artery embolization was therefore attempted. Extensive recanalization of the right and left hepatic arteries was observed and re-embolization using Ivalon sponge was performed. Tum6ur was also seen to be compressing the main hepatic vein causing a functional Budd-Chiari syndrome, explaining her jaundice, hepatomegaly and ascites. An attempt to stent the hepatic vein a few days later was unnecessary as the tumour mass had completely disappeared and the vein was clearly patent. Once again a good response was achieved and the patient was allowed home on no treatment where she remains some 6 months later. Discussion It has long been known that endocrine tumour metastases in the liver depend predominantly on the hepatic artery for their blood supply, whilst normal hepatocytes have a dual supply with 75% derived from the portal vein'. This led to the suggestion that hepatic devascularization might be beneficial in the treatment of hepatic metastases2. Hepatic artery ligation was found to be useful in the treatment of metastatic carcinoid tumours3, but later percutaneous hepatic artery embolization was preferred because' of its lower mortality". More recently selective peripheral vessel embolization has been shown to be more effective as this minimizes the otherwise rapid and extensive formation of collaterals, resulting in greater tumour necrosis7. The procedure can be repeated at intervals for maximal effect. The literature

Skin presentation of chronic myelomonocytic leukaemia resembling cutaneous lymphomas

M S Doutrel C Beylot' P Bioulacl A de Mascarel2 Departments of 'Dermatology and 2Pathology, Hopital Haut-L&veque, UniversitM de Bordeaux, 33600 Pessac, France Keywords: skin lesion; chronic myelomonocytic leukaemia

Chronic myelomonocytic leukaemia, according to the widely used FAB classification, is one of the myelodysplastic syndromes'. As with otlier haemopathies, there can be specific skin involvement, but few such cases have been reported. This is an observation of a case in which the skin lesions were initially diagnosed and treated as lymphoma.

593

contains many reports of the use 6f thi*tphnique since it was first described in 19775A. Most series report a mortality rate of 1%, mainly due to fulminant hepatic failure in patients with severely compromised liver function prior to the procedure. Morbidity from hepatic abscess formation and septicaemia can be reduced by the use of prophylactic antibiotics and that from urate nep1'ropathy can be prevented by prior treatment with allopurinol. The theoretical risk of gallbladder infarction appears to be very rare. Over 70% of reported cases have involved carcinoid tumours, with fewer reports involving non-seeetory tumours, glucagonomas and gastrinomas. To our knowledge there is no previous report of its use in metastatic insulinoma. In summary, we have reported a case of incapacitating hypoglycaemia secondary to metastatic insulinoma refractory to conventional therapy which responded on two occasions to hepatic artery embolization. We feel that this technique should be considered early in other such cases. References 1 Breedin C, Young G. Blood supply of neoplasms in the liver. Am J Pathol 1954;30:969-85 2 Markovitz J. The hepatic artery. Surg Gynecol Obstet 1952; 96:644-6 3 Jugdutt BI, Watanabe M, Turner FW. Hepatic artery ligation in the treatment of the carcinoid syndrome. J Clin Oncol 1983;1:727-40 4 Melia WM, Nunnerley HB, Johnson PJ, Williams R. Use of arterial devascularization and cytotoxic drugs in 30 patients with carcinoid syndrome. Br J Cancer 1982;46:331-9 5 Jaffer' A, Ajani MD, Humberto Carrasco C, et al. Islet cell tumours metastatic to the liver. effective palliation by sequential hepatic artery embolization. Ann Intern Med 1988;108:340-4 6 Allison'DJ, Modlin IM, Jenkins WJ. Treatment of carcinoid metastases by hepatic artery embolization. Lancet 1977; ii:1323-5 7 Pueyo I, Jimenez JR, Hernandez J, et al. Carcinoid syndrome treated by hepatic artery embolization. AJR 1978;131:511-14 8 Maton PN, 'Camilleri M, Griffin G, et al. Role of hepatic artery embolization in the carcinoid syndrome. BMJ 1983;287:932-935. 9 Mitty HA, Richard RP, Warner MD, et al. Control of the carcinoid syndrome with hepatic artery embolization. Radiology 1985;156:623-6 10 Carrasco CH, Chsrnsangarej C, Ajani J, et al. The carcinoid syndrome: effective palliation by sequential hepatic artery embolization. AJR 1986;147:149-54

(Accepted 30 April 1990. Correspondence to Dr D A Heath. Dr Wells is now at the Department of Haematology, Royal Free Hospita4 London)

Case report The patient, A 68-year-old man, was referred to us for skin lesions that had appeared 2 months earlier. His face, chest and back showed about 10 purplish-blue, non-tender nodules varying in diameter from about 1-4 cm (Figure 1). The rest of the clinical examination was normal. Biopsy of a nodule showed under a slightly drawn epidermis, through the entire thickness of the dermis, a layer proliferation infiltrating between the collagen tracts, made up of mononuclear cells. with 6ften large, uneven, notched nuclei which are sometimes in mitosis. The infiltrate remains away from the basal cell (Figure 2). However, it ensheathes the hair follicles without penetrating into the pilary sheaths. Immunohistochemistry showed marking of about 90% of the infiltrate cells with LEU3 (Becton Dickinson) monoclonal antibody revealing CD4 lymphocytes whilst marker analysis was virtually negative with antisera to B lymphocytes (PAN B-DAXO), T lymphocytes.(LEU4, Becton Dickinson) and. cytotoxic-suppressor T lymphocytes (LEU2a-2b, Becton Dickinson). The blood count was normal except for a monocytosis (0.8x 10 1). The bone marrow biopsy only showed as a trilinear hyperplasy. A scan of the thorax and

0141-0768/90/ 090593-02.00/0 © 1990 The Royal Society of Medicine

594

Journal of the Royal Society of Medicine Volume 83 September 1990

FLgure 14 Nodulareruption.....oeba

Figure 1. Nodular eruptiion over back

Figure 2. Dermal infiltrate of leukaemia cells

LEU3 positive, but also reacted positively to anti-HLA Dr (Becton-Dickinson), LEU11 (Becton-Dickinson) and antimonocyte (Bethesda Research Laboratories). The same staining was observed with circulating monocytes. Treatment was started with hydroxyurea (Hydrea) (500mg a day) butthe patient became gradually weaker, splenomegaly increased and a bone marrow showed acute myelomonoblastic leukaemia. The patient died a month later. Discussion There are two points raised by this observation. Firstly, specific skin involvement in chronic myelomonocytic leukaemia would appear to be exceptional. We have found reports of only 10 cases in medical literature of such lesions which can take the form of nodules, a pruriginous rash or a maculo-papular eruption2 . In each case the biopsy showed a non-epidermotropic dermal infiltrate composed of mononuclear cells. Occasionally, as was the case with our patient and the three cases reported by Copplestone5, these lesions are the first symptom of haemopathy. When they occur in the course of the disease, prognosis is variable. They can herald an acute phase of the disease, but in other cases no deterioration of the patient's condition is reported. The second point concerns the use of monoclonal markers. It is certain that it was the positive reaction with antisera to LEU3 which is the usual observation for CD4 lymphocytes that led to the diagnosis of lymphoma. However, this antibody is equally reactive with cells of the monocytemacrophage lineage6. Screening the second biopsy with an expanded panel of antisera enabled us to confirm the origin of the cells composing the infiltrate. This would show how important it is to use immunohistochemistry in haemopathic skin infiltration but on condition that the screening panel of antisera be sufficiently large to allow clear interpretation and to avoid errors due to cross-reactions.

abdominal-pelvic region showed no deep adenopathies. In the light of these results the disease was diagnosed as a malignant non-epidermotropic skin lymphoma with CD4 lymphocytes. In view of the rapid evolution of the lesions and the depth of the infiltrate, chemotherapy (cyclophosphamide and vincristine) was started in association with general corticotherapy. The skin lesions regressed with the first course of treatment. But at this stage changes occurred in the blood count with a rapid increase in circulating monocytes to 3700 mm3. Serous and urinary lysozyme also increased. Bone marrow aspiration showed a myelodysplastic syndrome. Chronic myelomonocytic leukaemia was diagnosed, and therapy discontinued, particularly as the skin lesions had resolved after three courses of chemotherapy. However, 3 months later, the patient was again suffering from pinkish-brown weal-like lesions on the thorax, coupled this time with generalized weakness and splenomegaly. The blood count showed anaemia (8.5 g/l haemoglobin), thrombocytopenia (70 x 109 1) and neutropenia (1.3x 1O9 1). Monocytosis was still in the region of 4000 mm3. There was slight myelaemia. A new skin biopsy showed identical results. An expanded panel of antisera was then used on this biopsy; the cells were still

(Accepted 17 October 1989)

Correlation between serum oncogene protein expression and the development of neoplastic disease in a worker exposed to carcinogens

Evidence suggests that oncogene activation may be a relatively early step in the carcinogenic process and that the detection of oncogene activation may be a useful marker for identifying individuals at risk for cancer'. The case reported provides evidence in support of this hypothesis.

P W Brandt-Rauf MD DPH' H L Niman PhD2 S J Smith MPH1 'Division of Environmental Sciences, Columbia University School of Public Health, 60 Haven Avenue, New York, NY10032, USA and 2Progenx Inc, San Diego, California, USA Keywords: ras oncogene; p21 protein; asbestos; colonic neoplasia

References 1 Solal-Celigny P, Desaint B, Herrera A, et al. Chronic myelomonocytic leukemia according to FAB classification - analysis of 35 cases. Blood 1984;63:634-8 2 Duguid JKM, Mackie MJ, McVerry BA. Skin infiltration associated with chronic myelomonocytic leukaemia. Br J Haematol 1983;53:257-64 3 Eubanks SW, Patterson JW. Subacute myelomonocytic leukaemia - an unusual skin manifestation. J Am Acad Dermatol 1983;9:581-4 4 Pozo-Roman T, Menarguez-Palanca IJ, Gomez-Pineda A, Gonzalez-Herrada CM, Lazaro-Ochaita P. Specific cutaneous involvement in the course of chronic myelomonocytic leukemia simultaneously with blastic leukemic transformation. JAm Acad Dermatol 1985;12:943-8 5 Coppelstone JA, Oscier DG, Mufti GJ, Hamblin TJ. Monocytic skin infiltration in chronic myelomonocytic leukemia. Clin Lab Haematol 1986;8:115-19 6 Wood GS, Warner NL, Warnke RA. Anti Leu 3/T4 antibodies react with cells of monocyte/macrophage and Langerhans lineage. J Immunol 1983;131:212-16

Case report A previously described cohort of 24 workers with known exposure to carcinogenic materials was referred to ColumbiaPresbyterian Medical Center in New York for evaluation2. These individuals were screened for the presence of proteins in their serum encoded by nine different oncogenes using a recently developed assay2'3. One 57-year-old white male in this cohort was of particular interest because of his long history of workplace exposure to asbestos, pesticides and polychlorinated biphenyls. He had also smoked 20 cigarettes

0141-0768/90/ 090594-02/$02.00/0 © 1990 The Royal Society of Medicine