TO EDITOR - BMJ
Transcript of TO EDITOR - BMJ
J Clin Pathol 1990;43:1036-1039
LETTERS TOTHE EDITOR
Primary hepatic vascularleiomyosarcoma of probable portalvein origin
Primary sarcomas of the liver comprise 1-2%of hepatic malignancies and most of these areangiosarcomas. To our knowledge only 15cases of primary hepatic leiomyosarcomahave been reported,' of which only two werevascular leiomyosarcomas.23 We report a caseof vascular leiomyosarcoma of the liver.A 67 year old woman had been complaining
of epigastric and right hypochondrial pain fortwo years. On examination she hadhepatomegaly. An ultrasound scan showed anodular vascular tumour in the left lobe oftheliver and a computed tomogram showed threesatellite lesions in the right lobe. All theselesions were enhanced with contrast mediumin the manner of a haemangioma. Atlaparotomy, a large vascular tumour wasfound in the left lobe of the liver. Smallnodules were noted in the right lobe, and asingle lesion 0-8 cm in diameter was found inthe small bowel mesentery. No lesion,however, was found in the gastrointestinaltract. A palliative left hepatectomy was per-formed. Her recovery was satisfactory.The tumour measured 30 x 15 x 10 cm
and weighed 1880 g. It was well circum-scribed and grey-white in appearance, with abossellated surface. The cut surfaces had awhorled appearance and vascular spaces wereevident at the periphery. The small mesen-teric nodule showed similar features. Sec-tions were stained with haemotoxylin and
eosin, elastic van Gieson, and Masson's tri-chrome. In addition, the unlabelled perox-idase anti-peroxidase method was used todetect desmin and vimentin (antisera fromDako Ltd), and small fragments were pro-cessed for electron microscopical examina-tion.The tumour and mesenteric nodule consis-
ted of interlacing fascicles of eosinophilicspindle cells with elongated, blunt endednuclei, and showed an intricate vascular pat-tem. There was pronounced nuclearpleomorphism with bizarre tumour giantcells, and numerous mitoses, many of whichwere abnormal. The mitotic index was 35/10high power fields. There were areas ofnecrosis and hyalinisation. A medium sizedvein within the tumour showed pleomorphicsmooth muscle cells streaming outwards fromthe vessel walls (figure). Bile ducts wereincorporated deep within the tumour. Thetumour cells stained red with Masson's tri-chrome and showed desmin and vimentinpositivity; electron microscopical examina-tion showed myofilaments and pinocytoticvesicles, confirming a smooth muscle origin.The origin of parenchymal leiomyosar-
comas has been discussed; most are thoughtto originate from connective tissue or bloodvessels. In the liver the ligamentum teres hasbeen suggested as the origin of most of thereported cases of hepatic leiomyosarcoma.4
Vascular leiomyosarcomas are rare. Theyarise predominantly from larger or mediumsized blood vessels, the inferior vena cavaaccounting for over 75% of these cases. Thecharacteristic features of these tumoursinclude proliferating atypical smooth musclecells, some streaming out from the media ofvessels, and a striking number of neoformedblood vessels of variable size and configura-tion intermingled among them.The two cases described in the liver were
both of hepatic vein origin, and both presen-ted with the Budd-Chiari syndrome.23 Our
case had no evidence of the Budd-Chiarisyndrome. In addition, bile ducts werepresent, a feature not previously describedwithin hepatic leiomyosarcomas. A mediumsized vein showed pleomorphic tumour cellsstreaming out from the media. The combina-tion of these three features all suggest originfrom a portal vein.
In a detailed description of vascularleiomyosarcomas' it was concluded that themitotic index was the most important patho-logical feature on which a prognostic evalua-tion for a vascular leiomyosarcoma could bebased. Three of the patients with a mitoticcount of greater than 35/10 high power fieldshad a poor prognosis. Our case had a mitoticindex of 35/10 high power fields, an indica-tion of poor prognosis, and metastases werepresent.
M SUNDARESAN,SB KELLY,
IS BENJAMIN,AB AKOSA
Departments of Histopathology andHepatobiliary Surgery,
Royal Postgraduate Medical School,Hammersmith Hospital,
Du Cane Road,London W12 OHS
1 Chen KTK. Hepatic leiomyosarcomas. J SurgOncol 1983;24:325-8.
2 Koberle F, Pfleger R. Lebervenengeschwulstmit dem Symptomenbild einer EndophlebitisObliterans Hepatica. Wein Arch Inn Med1940;34:73-85.
3 MacMahon HE, Ball HG. Leiomyosarcomas ofthe hepatic vein and the Budd-Chiari syn-drome. Gastroenterology 1971;61:239-43.
4 MacSween RNM, Anthony PP, Sheuer PJ.Tumours and tumour like conditions of theliver and biliary tract. Pathology of the liver.London: Churchill Livingstone, 1987:574-645.
5 Varela-Duran J, Oliva H, Rosai J. Vascularleiomyosarcoma. Cancer 1979;44:1684-91.
Medium sized vein in the tumour with pleomorphic tumour cells streaming out of the media(arrows).
Peliosis hepatis after livertransplantation
Peliosis hepatis is characterised by blood-filled spaces in the liver. Causes and associa-tions identified to date include steroid hor-mone treatment, wasting diseases, cancer,human immunodeficiency virus disease andrenal transplantation.'-' We report peliosis ina liver transplant recipient; as far as we areaware, this has not been previously recorded.
Case report'A 25 year old female factory employeedeveloped acute hepatic failure following anoverdose of paracetamol. Emergency ortho-topic liver transplantation was carried outfour days after the overdose. There wasmismatching of blood groups (donor bloodgroup A, recipient group 0). A protocolbaseline liver biopsy specimen taken at thetime of revascularisation of the graft showedsubstantially normal liver. Four days later asecond biopsy was performed because ofrising serum aspartate transaminaseactivities. This showed no evidence ofcellular rejection. Scattered acidophil bodiesand increased numbers of liver cell mitoseswere noted. A further biopsy specimen on day7 showed, in addition to the above changes,portal infiltration, endotheliitis affecting por-tal venules and minor bile duct changes. Thepicture was interpreted as cellular rejection.
1036 on July 4, 2022 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.43.12.1036-b on 1 Decem
ber 1990. Dow
nloaded from
Letters to the Editor
rather than days after beginning oftreatment,although it appeared earlier after treatmentwith anabolic steroids.' We cannot entirelyrule out a possible role of drugs, but all thedrugs given to our patient are commonly usedafter transplantation. The pathogenesis ofpeliosis remains controversial, and venousoutflow obstruction, endothelial damage, andliver cell necrosis have been implicated; thepresence of reticulin strands in the lesionssupports these last two factors.We suggest that in our patient the blood
cysts may have been formed during a processof liver cell loss following endothelialdamage. This is usually seen in portal orhepatic veins rather than sinusoids, but it ispossible that the unusual location of thedamage reflects the mismatching of bloodgroups between graft donor and recipient.
PJ SCHEUERLA SCHACHTER
S MATHUR
AK BURROUGHSK ROLLES
Departments of Histopathology,Medicine and Surgery, Royal Free Hospital
and School of Medicine,Pond Street, London NW3 2QG
1 Bagheri SA, Boyer JL. Peliosis hepatisassociated with androgenic-anabolic steroidtherapy. Ann Intern Med 1987;81:610-18.
2 Degott C, Rueff B, Kveis H, et al. Peliosishepatis in recipients of renal transplants. Gut1978;19:748-53.
3 Haboubi NY, Ali HH, Whitwell HL, Ackrill P.Role of endothelial cell injury in the spectrumof azathioprine-induced liver disease afterrenal transplant. Am J Gastroenterol1988;83:256-60.
4 Snover DC. The pathology of acute rejection.Transplan Proc 1986;18:123-7.
5 Grond J, Gouw ASH, Poppema S, et al. Chronicrejection in liver transplants: a histopathologicanalysis of failed grafts and antecedent serialbiopsies. Transplan Proc 1986;83:128-35.
An irregular cystic area of liver cell loss andhaemorrhage (top) is seen near a terminalhepatic venule (below) (haematoxylin andeosin).
Neutrophils and eosinophils were present inmoderate numbers in the portal infiltrate inaddition to lymphoid cells. Serum enzymeactivities remained raised, and a biopsyspecimen on day 11 showed continuingfeatures of rejection.At day 16, a biopsy specimen showed less
rejection but some canalicular cholestasis andsinusoidal dilatation. There was a single smallhaemorrhagic focus in the parenchyma. Byday 28, more prominent, cyst-like haemor-rhagic foci were present, containing remnantsof the reticulin framework (figure). Sin-usoidal endothelium seemed to have beenpartly lost in these areas, but was presentelsewhere in the hepatic acini. Features ofrejection in the portal tracts were now verymild, and no endothelial lesions were iden-tified in portal venules or terminal hepaticvenules (centrilobular veins). Cholestasispersisted. The patient subsequently im-proved, and no further liver biopsies wereperformed.The patient had received no medication in
the months before overdose, nor was shetaking oral contraceptives. Immediately afteradmission she received dopamine. Aftertransplantation she was given methyl pred-nisolone (0-8 mg/kg daily from day 1, chan-ged to prednisolone on day 4), azathioprine(about 1-5 mg/kg daily from day 11), andcyclosporine (10 mg/kg daily from day 13).Mezlocillin and metronidazole were givenduring the first week, and amphotericin,acyclovir, and ranitidine from day 3. Doses ofdextropropoxyphene/paracetamol anddihydrocodeine were given during the firstweek.
Vascular lesions reported after liver trans-plantation include endotheliitis as part ofcellular (acute) rejection,4 and narrowing ofarteries and arterioles as part of chronicrejection.' We have observed venous throm-bosis leading to the Budd-Chiari syndrome inone of our patients (unpublished observa-tion). To this list must now be added peliosishepatis. Azathioprine has been implicatedafter renal transplantation, but in the patientsreported23 the peliosis has developed months
Postoperative necrotising granulomatain the cervix and ovary
Postoperative necrobiotic granulomata in theprostate and bladder have been welldocumented. In the female genital tract theyhave, more recently, been described in thefallopian tube,' cervix,23 and ovary.4 We haverecently seen two such cases. In the first,necrobiotic granulomata were seen in thelower uterine segment scar of a hysterectomyspecimen from a 29 year old woman who hadhad an uncomplicated lower uterine segmentCaesarean section six months previously. Inthe second case, similar granulomata werepresent in the right ovary of a 39 year oldwoman who had undergone a right ovarianwedge biopsy 18 months earlier at the time ofbilateral salpingectomy and left oophorec-tomy for lower abdominal pain. Both patientshad no clinical or laboratory evidenceof tuberculosis, sarcoidosis, rheumatoidarthritis, parasitic or venereal disease.The granulomata in both cases were of
similar appearance (figure). They were smallin size with brightly eosinophilic stainingcentral areas of fibrinoid necrosis aroundwhich were well defined zones of palisaded
histiocytes admixed with inflammatory cells,including plasma cells, lymphocytes, andoccasional eosinophils. A few Langhans' typegiant cells were also present. No birefringentmaterial was identified in either case andstaining with silver methenamine, ZiehlNeelsen, periodic acid Schiffand Gram stainswas negative.The pathogenesis of such lesions has not
been well defined. No cases have so far beendescribed in which there has not been preced-ing diathermy or surgery, or both, at the samesite. Diathermy is particularly associatedwith the subsequent development ofnecrobiotic granulomata. The primary eventmay be that of traumatic tissue damage.Using x ray energy dispersive analysis, Spag-nolo et al showed the presence of sulphurpeaks in the granulomata of three out of fourcases examined, but not in bladder tissueremote from the granulomata.' They con-cluded that sulphur may have been releasedfrom damaged collagen and elastin which arerich in disulphide bands. Such damage wouldcreate tissue "foreign" to the host, thusstimulating an immune reaction of which thehypersensitivity response may at least play apart. That such granulomata are not com-
Granuloma with central area of necrosis (N) surrounded by histiocytes, lymphocytes, andoccasional giant cells (arrow) (haematoxylin and eosin).
1037 on July 4, 2022 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.43.12.1036-b on 1 Decem
ber 1990. Dow
nloaded from