Primary cardiac Kaposi's sarcoma

Thorax (1976), 31, 595.

Primary cardiac Kaposi's sarcoma

DAVID A. LEVISON and PETER d'A. SEMPLE

Department of Pathology, Ninewells Hospital and Medical School, and Department ofRespiratory Diseases, Kings Cross Hospital, Dundee

Levison, D. A. and Semple, P. d'A. (1976). Thorax, 31, 595-600. Primary cardiac Kaposi'ssarcoma. We report the clinical, laboratory, and necropsy findings in a 14-year-old boywith a primary Kaposi's sarcoma of the heart. Primary cardiac Kaposi's sarcoma andangiosarcoma are compared, and the pathogenesis of Kaposi's sarcoma is discussed.Relevant literature is briefly reviewed.

Primary angiosarcoma of the heart is itself a raretumour, with only 45 previously reported cases(Patt, Halkin, and Jaffe, 1974; MassachusettsGeneral Hospital, 1975). Even rarer as a primaryin the heart is that other vascular tumour-Kaposi's sarcoma, with only five previous reports(Choisser and Ramsey, 1939; Weller, 1940;Aegerter and Peale, 1942; Contreras, 1957; Gel-fand, 1957). These two tumours have undoubtedlybeen confused in the past, and by presenting thiscase we hope to go some way towards clarifyingthe situation. Our patient (aged 14) is youngerthan any previously reported with either primarycardiac angiosarcoma or Kaposi's sarcoma; hisright atrial tumour is histologically a typicalKaposi's sarcoma.

CASE REPORT

A 14-year-old boy was found to have a largeheart in a chest radiograph, taken because ofschool contact with a case of active tuberculosis.A presumptive diagnosis of tuberculous peri-cardial effusion was made and he was admitted tehospital. There was no previous illness of note,and his only symptoms were dry cough, retro-sternal discomfort, anorexia, and occasionalvomiting during the preceding week. On examina-tion he looked well without dyspnoea or cyanosis.The heart rate was 90 per minute with no pulsusparadoxus and systemic blood pressure 114/80mmHg. The area of precordial dullness was in-creased and the apex beat was impalpable. Heartsounds were soft with no murmurs. The liveredge was palpable 3 cm below the costal marginand was slightly tender. The jugular venous pres-sure was not raised.

Investigations on admission: erythrocyte sedi-mentation rate 17 mm/hr, haemoglobin 16-9 g/dl,white cell count 17-5X109/1 (neutrophil leucocy-tosis). Urea and electrolytes normal. SGOT94 IU/ml, HBD 268 IU/ml, bilirubin 12 ,umol/l.Sputum negative for tubercle bacilli. Electro-cardiogram showed sinus tachycardia (110 beatsper minute), low voltage, and ventricular extra-systoles.

Tuberculous pericardial effusion still seemed themost likely diagnosis despite a negative Mantouxtest. Rifampicin, 600 mg, isoniazid, 300 mg, andethambutol, 800 mg, were started two days afteradmission. On the fifth day he became dyspnoeicand mildly cyanosed. Although the jugular venouspressure was still not visibly raised, the liver edgewas now palpable 5 cm below the costal margin,and early cardiac tamponade was suspected. Onelitre of bloodstained pericardial fluid was aspiratedwith little benefit. Its protein content was 50 g/l.No tubercle bacilli were seen in a Ziehl-Neelsenstained film of the aspirate, and cytological ex-amination showed only red cells, histiocytes, andlymphocytes.

Within a few hours of aspiration the patient be-came more cyanosed and dyspnoeic. Serum biliru-bin was found to be 71 ,umol/l, SGOT 2640 IU/ml,and HBD 824 IU/ml, suggesting possible liverdamage; accordingly, rifampicin was withdrawn.By this time Q waves had developed in chestleads VI-V4 of the electrocardiogram, suggestingcoronary occlusion as another possible cause ofenzyme rise. Lest further haemopericardium hadbeen induced by needle trauma during pericardialaspiration, a further 300 ml of sanguinous peri-cardial fluid was aspirated with no clinical benefit.The patient's condition rapidly deteriorated and

595

copyright. on N

ovember 18, 2021 by guest. P

rotected byhttp://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.31.5.595 on 1 O

ctober 1976. Dow

nloaded from

http://thorax.bmj.com/

David A. Levison and Peter d'A. Semple

he died 10 hours after the initial pericardialaspiration and five days after admission. Opencardiac massage was attempted. Necropsy wasperformed 18 hours later.

PATHOLOGICAL REPORT The most significant find-ings were confined to the heart. A thin film offibrin extended over the inflamed epicardial sur-face but there was no residual effusion as the peri-cardial sac had been opened for cardiac massage.The anterior surface of the right atrium was palein colour and lobulate in contour. Coronal sectionof the heart showed a lobulate haemorrhagictumour, approximately 5 cm in diameter, fillingmost of the right atrium and firmly attached tothe lateral, anterior, and septal walls (Fig. 1). Aprobe could just be passed from superior venacava to inferior vena cava behind the tumour. Noother tumour deposits were identified in the heartor elsewhere. Liver, spleen, and kidney showedthe appearances of chronic venous congestion.

FIG. 1. Coronal section of heart viewed from behind.The dilated right atrium shows tumour invading andexpanding its lateral wall and the interatrial septumand bulging into the lumen.

Sections of the right atrial tumour show ahighly vascular tumour invading the myocardium.Its vessels vary markedly in size. In some areasthey are large, dilated, and lined by a single layerof uniform plump endothelial cells, while inothers they are mere slit-like spaces with noobvious endothelial lining. Spaces between thevessels are filled by sheets of polyhedral andspindle cells with hyperchromatic nuclei and ill-defined cytoplasm (Figs 2 and 3). Among thesecells are several mitotic figures. There are exten-sive areas of haemorrhage and necrosis withnumerous macrophages, some containinghae4nosiderin, as well as foci of neutrophils andlymphocytes.

Special stains show little mucin and collagenbut a reticulin stain shows appearances similar tothose reported in 1939 by Choisser and Ramseyin their cases of primary cardiac Kaposi's sar-coma. There is the same pattern varying fromfine, lace-like fibres to dense, broad, branchingprocesses. Most of the vascular spaces appear tobe completely surrounded by reticulin.

Electron microscopy (Fig. 4) shows that themajority of the tumour cells have pale nucleiwith oval or irregular outlines, fine chromatin,and prominent nucleoli. Their cytoplasm containsscanty organelles and occasional pieces of mem-brane-bound and sometimes laminated debris.These cells line small vascular spaces, some ofwhich contain red cells. Some of the tumour cellsare separated by dense material, possibly basementmembrane. Also present are numerous neutrophilpolymorphs, cells with shrunken darker nuclei(probably dying), much fibrin, and cell debris.

Sections of the other organs show the appear-ances of chronic venous congestion, the only un-expected finding being the presence of fairlyextensive focal chronic thyroiditis.

DISCUSSION

Kaposi's sarcoma is best known in Africans andusually presents with the syndrome of lympho-edema, multiple cutaneous tumours, and lymph-adenopathy with ultimate visceral involvement.Cardiac lesions in this form of the disease arerecognized, being present in five out of 19 casesnecropsied by Lothe and Murray (1962).Of much greater rarity are those tumours

previously reported as primary Kaposi's sarcomaof the heart in both Negroes and Caucasians(Choisser and Ramsey, 1939; Weller, 1940;Aegerter and Peale, 1942; Contreras, 1957; Gel-fand, 1957). However, there are those (Glancy,

596

copyright. on N



j.com/


ctober 1976. Dow

nloaded from



FIG. 2. Right atrial tumour. The mixed vascular and cellular pattern is shown.Thin-walled vessels stuffed with red cells are best shown to the left of centre (H andEX240).

FIG. 3. A more solid part of the tumour. Abundant spindle cells, many lining capillarysize vessels, separate larger vascular spaces (H and EX300).

597

copyright. on N



j.com/


ctober 1976. Dow

nloaded from



a0,V B

FIG. 4. Electron micrograph of the tumour. Three red blood cells lie in the lumen of a tumour capillary.The tumour cells have pale cytoplasm containing few organelles and fragments of membranous debris; onerumour cell nucleus with prominent nucleolus is shown (X 15 000).

Morales, and Roberts, 1968) who believe that noneof these cases merits distinction from other formsof angiosarcoma 'on the basis of published photo-graphs'. On this basis, there is some difficulty inseparating Kaposi's sarcoma from other forms ofangiosarcoma, but this is due at least partly tosome cases of histologically typical Kaposi's sar-

coma being reported as angiosarcoma (Hewer andKemp, 1936). Perhaps the possibility of Kaposi'ssarcoma was not actively considered because ofthe absence of typical skin lesions. Glancy et al.(1968) even suggest that cardiac involvement byKaposi's sarcoma should not be diagnosed in theabsence of skin lesions.

598

4...

i

copyright. on N



j.com/


ctober 1976. Dow

nloaded from



*But the histological criteria for the diagnosis ofKaposi's sarcoma are well established (Murrayand Lothe, 1962: Evans, 1968) and widely agreed.The tumour consists of a variable mixture ofvascular components and spindle cells. Thevascular component usually includes dilated bloodvessels as in cavernous haemangioma and numer-ous capillary sized vessels. The large vascularspaces are often incompletely lined, but the capil-lary vessels are nearly always well formed. Seldomdo the vessels contain piled up masses of endo-thelial cells so characteristic of angiosarcoma.The other important component, the spindle cells,may dominate the histological picture. Within thespindle-cell tissue are slits and clefts bounded byspindle cells rather than obvious endothelium, theclefts often containing erythrocytes (appearancessimilar to those illustrated in Fig. 3). Foci oflymphocytes, plasma cells, and macrophages-some containing haemosiderin-are often present.

Angiosarcoma, on the other hand, is generallycomposed of anastomosing channels lined byluxuriant and atypical epithelium (Evans, 1968).Occasionally the lining consists of a single row ofround, polygonal or fusiform cells, but often itscells proliferate to form heaped up masses severallayers thick. Freely anastomosing channels maygive rise to a papillary pattern while in the morecellular areas the neoplastic endothelial cells ac-cumulate in masses within closely opposed vesselsto give the impression of a solid tumour lackingvasoformative tendencies. An inflammatoryinfiltrate, as seen in Kaposi's sarcoma, is not acommon feature.The histological findings in the present 14-year-

old boy with a primary cardiac tumour, describedin detail above, are typical of Kaposi's sarcoma.So too are the appearances illustrated by Lotheand Murray (1962) from a case of cardiac in-volvement without skin lesions-previouslyreported without histological illustration by Gel-fand (1957). This latter case is one of 496 cases ofKaposi's sarcoma seen by these authors and assuch is very strong supportive evidence for theexistence of a primary cardiac form of the disease.

Electron microscopy has not apparently beendone previously on a case of primary Kaposi'ssarcoma of the heart, but our observations aresimilar to those of Pepler and Theron (1962) whoexamined classic skin lesions. However, there isno general agreement with their view that thetumour is derived from Schwann cells. Laminatedstructures are present in the cytoplasm of thisboy's tumour cells, but we interpret these as celldebris rather than axis cylinders.

Studies on the pathogenesis of Kaposi's sar-coma have produced conflicting results (Lancet,Editorial, 1967). Histochemical, cultural, andelectron microscope studies have not finally re-solved the nature and origin of the spindle cells.Schwann cells have been proposed by someauthors but excluded by others. A 'reticulo-endothelial' origin now seems to be most popular.Of considerable interest is the apparent associa-tion, in temperate climates at least, of Kaposi'ssarcoma with a lymphomatous, leukaemic orother primary neoplasm (Lancet, Editorial, 1967).One possibility raised by this association is thata defect exists in the normal 'immune surveillancemechanism'. The only hint of any immunologicaldisorder in our case is the presence of extensivefocal chronic thyroiditis of the kind usuallyattributed to an autoimmune pathogenesis.Another possible clue to the aetiology of Kaposi'ssarcoma is its peculiar geographic distribution incertain parts of Africa. This, particularly in viewof the recent developments in the study ofBurkitt's lymphoma, raises the possibility of anenvironmental agent; though a virus aetiology haslong been suggested, evidence is lacking.

In conclusion, this case illustrates severalfeatures common to primary cardiac angio-sarcoma and Kaposi's sarcoma such as the youngage group, site of origin in the right atrium, pre-sentation with pericardial effusion and cardiacfailure, and the rapidly fatal outcome. It under-lines the fact that Kaposi's sarcoma occasionallyoccurs in the heart in the absence of typical skinlesions.

We should like to thank Dr. R. N. Johnston for per-mission to report this case, Dr. E. Olsen for reviewingthe histology and encouraging us to publish, and Mr.R. S. Fawkes for the photographs.

REFERENCES

Aegerter, E. E. and Peale, A. R. (1942). Kaposi'ssarcoma: a critical survey. A rchives of Patho-logy, 34, 413.

Choisser, R. M. and Ramsey, E. M. (1939). Angio-reticuloendothelioma (Kaposi's disease) of theheart. American Journal of Pathology, 15, 155.

Contreras, R. (1957). Angiosarcoma de Kaposi,primario del coraz6n: revision de la literaturay descripci6n del sexto caso. Archivos de Institutode Cardiologia de Mexico, 27, 463.

Evans, R. W. (1968). Histological Appearances ofTumours, 2nd edition. Livingstone, Edinburghand London.

Gelfand, M. (1957). Kaposi's haemangiosarcoma ofthe heart. British Heart Journal, 19, 290.

599

copyright. on N



j.com/


ctober 1976. Dow

nloaded from



Glancy, D. L., Morales, J. B., Jr., and Roberts, W. C.(1968). Angiosarcoma of the heart. AmericanJournal of Cardiology, 21, 413.

Hewer, T. F. and Kemp, R. P. (1936). Malignanthaemangio-endothelioma of the heart: report ofa case. Journal of Pathology and Bacteriology,43, 511.

Lancet (1967). Mysterious sarcoma. (Editorial).Lancet, 2, 1290.

Lothe, F. and Murray, J. F. (1962). Kaposi's sar-coma: autopsy findings in the African. ActaUnio Internationalis Contra Cancrum, 18, 429.

Massachusetts General Hospital Case record No. 35(1975). New England Journal of Medicine, 293,494.

Murray, J. F. and Lothe, F. (1962). The histopatho-logy of Kaposi's sarcoma. Acta Unio Inter-nationalis Contra Cancrum, 18, 413.

Patt, Y. Z., Halkin, H., and Jaffe, R. (1974). Primarycardiac angio-sarcoma. Israel Journal of MedicalSciences, 10, 525.

Pepler, W. J. and Theron, J. J. (1962). An electron-microscope study of Kaposi's haemangiosarcoma.Journal of Pathology and Bacteriology, 83, 521.

Weller, G. L., Jr. (1940). The clinical aspects ofcardiac involvement (right auricular tumor) inidiopathic hemorrhagic sarcoma (Kaposi'sdisease). Annals of Internal Medicine, 14, 314.

Requests for reprints to: Dr. D. A. Levison, Depart-ment of Pathology, Ninewells Hospital and MedicalSchool, Dundee, Scotland.

600

copyright. on N



j.com/


ctober 1976. Dow

nloaded from


Documents

Primary cardiac Kaposi's sarcoma