521
Aust. N.Z. J . Surg. 1992,62,521-524
ORIGINAL ARTICLES HEPATIC CAVERNOUS HAEMANGIOMA: A 10 YEAR REVIEW N.
TAIT, A.
J.
RICHARDSON, G . MUGUTI*AND
J . M.
LITTLE
Department of Surgery, Westmead Hospital, Westmead, New South Wales. Australia and *Department of Surgery, Mpilo Hospital, Bulawayo, Zimbabwe Between January 1981 and July 1991, 61 patients with hepatic haemangiomata were examined at Westmead Hospital. There were 14 males (22%) and 47 females (78%). The age range was 26-85 years with a median of 49 years. Forty-one had abdominal symptoms but these could be attributed to a haemangioma in only seven cases. There was at least one subcapsular lesion in 17 (28%). Six of the seven symptomatic lesions were subcapsular and five of these were giant haemangiomata (i.e. more than 4 cm in greatest diameter). One large symptomatic lesion was intrahepatic. No association was observed between hepatic haemangiomata and other hepatic or extrahepatic diseases. Haemangiomata were resected from six patients, four of whom were symptomatic. Symptoms improved in all four but did not resolve completely in any. Follow-up ranged from nil in five patients to 108 months in one. The median follow-up was 12 months after initial diagnosis. Ten patients showed evidence of change in their lesions or symptoms while under observation. Only three had worsening symptoms or suspected change in size of a haemangioma. This study highlights the benign, static nature of most hepatic haemangiomata. When this lesion is suspected, the diagnosis should be confirmed with ultrasound (US) and labelled red blood cell scanning (RBCS). Referral for evaluation by a specialist hepatobiliary surgery unit is necessary when symptoms are intolerable, increasing size is definitely demonstrated or the diagnosis is uncertain and cannot be established without specialized investigations. Bleeding into or from these lesions is rare. Key words: hepatic haemangioma, liver.
Introduction
Methods
From January 1981 to July 1991, 61 patients with Hepatic cavernous haemangioma is the most comhepatic space occupying lesions examined at Westmon primary hepatic t ~ m o u r . ' - In ~ this era of mead Hospital were found to have hepatic cavernnon-invasive, readily available organ imaging it is ous haemangiomata. A retrospective analysis of frequently an incidental finding4 The use of highly their case records was carried out. The following sensitive and specific imaging such as abdominal information was documented: age; sex; symptoms; ultrasound (US) and labelled red blood cell scanning (RBCS) leaves few in diagnostic d o ~ b t . ~ , ~ . ' past and present liver disease or disease that may affect the liver; sex steroid use; mode of presentaAlthough French's statement that: 'The pathologition; method of diagnosis: site, size and number of cal and clinical significance of the cavernous lesions; histology; follow-up and changes in symptumour is very slight" appears accurate, the natural toms; and tumour size or tumour number during history of this lesion is still poorly understood. Fear follow-up. of a more sinister neoplastic process and the threat of spontaneous rupture still troubles doctors and their patients, even where the organ imaging diagResults nosis of haemangioma is certain. This anxiety prompts many referrals to specialist hepatobiliary A G E A N D SEX D I S T R I B U T I O N units. There has been little evidence to support size at There were 61 patients of whom 47 were women. diagnosis, uncertain change in size during followThe median age of all patients was 49 years (range up, minor symptoms or fear of rupture as indica26-85). tions for resection of hepatic haemangioma.'-" SYMPTOMS Hepatic resection is now resorted to infrequently. Nineteen patients had no abdominal symptoms. Correspondence: Dr N. Tait, Department of Surgery, WestForty-one (67%) had abdominal pain or discomfort mead Hospital, Westmead, NSW 2145, Australia. but these symptoms were thought likely to arise from haemangioma in only seven. Four of these Accepted for publication 15 January 1992
522
TAIT ETAL..
complained of worsening pain during follow-up and ultimately underwent resection of their haemangiomata for symptom control. Another patient had a symptomatic colorectal hepatic metastasis and 33 had symptoms from other non-hepatic intraabdominal conditions. All the symptomatic patients had pain or discomfort at the time of presentation. No symptomless patient developed pain during the observation period. MODE OF DETECTION
Fifty-five patients (90%) had haemangiomata detected by organ imaging. Imaging procedures included ultrasound (US), labelled red blood cell scanning (RBCS), computerized axial tomographic (CAT) scanning and angiography . Magnetic resonance imaging (MRI) was used in only three cases. Ultrasound and RBCS were the most commonly used procedures. All patients diagnosed by organ imaging had at least two imaging procedures. Six had hepatic haemangiomata detected during laparotomy for unrelated causes. LIVER D I S E A S E S
Only three patients had histories of concurrent or past liver disease. One had a metastasis from colorectal carcinoma. Two had past histories of hepatitis, one of whom had hepatitis A and the other had hepatitis of uncertain nature. Five patients had other illnesses that may affect the liver including: congestive cardiac failure (1); Cushings disease (1); advanced chronic obstructive airways disease (1); and alcoholism without clinically established liver disease (2). None of these patients had symptomatic haemangiomata and none required treatment of hepatic haemangioma. S E X STEROID U S E
Of the 45 women, information on oral contraceptive medication (OCM) was available in only 25. Ten of the 25 had a history of OCM use. Two others had used progesterone preparations for treatment of menstrual difficulties. One male had used an oestrogenic medication for treatment of prostatic carcinoma. All four patients who had haemangiomata resected because of worsening symptoms had at some time used OCM. Two were using OCM when they developed symptoms. Cessation of OCM consumption did not reduce their symptoms or slow symptom progression. One of the two went back on OCM with no apparent effect on symptoms or on the size of her haemangioma. The other two patients had taken OCM for only a few months each. Both had stopped OCM use 10 years before developing symptoms from hepatic haemangioma.
NUMBER OF H A E M A N G I O M A T A
One hundred and eight haemangiomata were diagnosed in the 61 patients. The lesions were solitary in 38 (62%) and multiple in 23 cases. There was at least one subcapsular lesion in 17 (28%) cases. Six of the seven symptomatic patients had haemangiomata reaching the liver capsule. Three of the six had more than one lesion. One symptomatic patient had a 6.5 cm intrahepatic haemangioma with dense adhesions fusing the overlying liver surface to the diaphragm. SIZE
The diameter of the haemangiomata was measured on organ imaging in 55 cases or on the fixed and cut resected specimen in six. Median diameter was 3.5 cm with a range of 1-16cm. Six of the seven symptomatic lesions were giant haemangiomata with a greatest diameter of more than 4cm. The four lesions that were excised because of increasing symptoms measured 16, 8, 6.5 and 5cm respectively. There were 19 asymptomatic haemangiomata measuring more than 4 cm. Four of these were subcapsular. TREATMEN 1
Fifty-five patients received no active treatment. Of the six who had haemangiomata removed, four had resections of symptomatic lesions. Two had small lesions that were incidental findings at laparatomy. These were excised to exclude hepatic spread of unrelated intra-abdominal neoplasms. RESULTS OF T R E A T M E N T
All four patients undergoing elective resection of symptomatic haemangiomata reported subjective improvement. Symptoms did not fully resolve in any, although the benefit gained was considered substantial and worthwhile by all four. In particular, all four required less analgesic medication after the operation and two felt that reduction in pain had led to major improvements in social and family life. There were two minor complications from treatment, one wound haematoma and one wound infection. OBSERVATION A N D FOLLOW-UP
After the initial diagnosis, only seven patients, all with small lesions, two of which were excised at laparatomy, received no hepatic follow-up. Median follow-up for the remaining 54 patients was 12 months (range 1-108 months). During follow-up, 44 patients had no change in their symptoms or in the size or number of their lesions. Ten patients showed some change, including: increased symptoms (4) and possible increased size (2); a second lesion found on follow-up investigations (3); de-
523
HEPATIC HAEMANGIOMA
creased lesion size (1); and disappearance of a lesion (2). All four patients with increased symptoms had worsening right upper quadrant pain. On previous imaging two had shown growth of their lesions but review of the imaging did not support this. In one, the initial measurement had not been of the greatest diameter of the haemangioma. In the other, different measurements had been made on non-comparable equipment. nisroLoG'Y
Typical histologic changes of cavernous haemangioma were seen in all six resected lesions. One was largely sclerotic but areas of sclerosis were also seen in the other five. All four symptomatic lesions that were excised showed small areas of old and new thrombosis. None had acute inflammatory changes, evidence of old or fresh necrosis or of bleeding into the lesions. Three of the four reached the liver capsule. One was intrahepatic but the liver surface over the haemangioma was fused to the diaphragm by dense adhesions. The subcapsular liver, between the capsule and the superficial surface of the haemangioma was histologically normal.
Discussion The reported incidence of he atic haemangioma has been between 2 and 7%."' The benign nature of these lesions has in the past meant that most went undetected. As imaging techniques improve and become relatively cheaper and more available hepatic space occupying lesions will be found in increasing The most common lesion among these will be the hepatic haemangi~ma.~.'Of the haemangiomata diagnosed, less than 1% will present acutely due to spontaneous rupture and free intraperitoneal Even centres that collect major hepatic surgery referrals will see few haemangiomata present with life-threatening complications. Inatsuki et al., reporting from an institution that expected to perform over 100 hepatic resections in 1991, studied 219 patients undergoing resection of benign hepatic lesions. One hundred and fourteen had cavernous haemangiomata. Ten of these 114 had presented with haemoperitoneum, eight occurring after 'unwise needle biopsy or open biopsy'. Therefore, only two of 114 patients who had a cavernous haemangioma resected presented with spontaneous bleeding. Bleeding in the other eight was iatrogenic, caused mainly by needle biopsies, that frequently failed to make the diagnosis. Inatsuki" and Hobbs' advise that biopsy of a cavernous haemangioma should be avoided. For the majority of patients in whom hepatic haemangioma is suspected, a definitive diagnosis can be reached by using US and RBCS which are economical and without the risk or discomfort of
4
'
invasive These provide a highly sensitive and specific diagnostic combination that can be easily repeated on follow-up, though some large lesions may present a picture of mixed echogenicity on US. In these cases RBCS should still be diagnostic, though angio CT, angiography or MRI may be indicated if a second confirmatory test is required. I 3 , I 4 Once the diagnosis is established, very few patients should require liver resection. Surgical intervention should be avoided unless a haemangioma presents with bleeding, incapacitating symptoms, alteration in size or cannot be accurately d i a g n ~ s e d . ' ~ ' ' ~For ' ~ ,the ~ ~ few that do require treatment, excision is the most suitable option if it can be performed effectively and safely with few complications and low mortality. Non-excisional techniques such as radiation, hepatic artery ligation and embolization have less chance of symptom control and are contraindicated if there is any doubt about the diagnosis. In this series, no patient required operation to confirm the diagnosis. Two of the six resected lesions were incidental findings, confirmed during operation for unrelated abdominal malignancy. In these cases the lesions were small and partly sclerotic. They were resected for histopathologic examination to exclude secondary malignancy. The four patients who had haemangiomata resected because of increasing symptoms were all females with large lesions which were subcapsular in three. None had liver disease, signs of bleeding or necrosis or were the cause of diagnostic doubt. Capsular involvement and size cannot be the only determinants of symptom development as 19 of the 25 giant lesions with diameters greater than 4cm were asymptomatic, even though four of these reached the liver capsule. I 7 3 l 8 A clinician dealing with a hepatic space occupying lesion needs to reach a definitive diagnosis. This can be achieved in more than 90% of haemangiomata using US and RBCS and,6 if necessary, CAT scanning,I4 angiography" or MRI.*' Haemoperitoneum due to rupture of a hepatic haemangioma is a rare event most often caused by a biopsy needle.9,",'3 There seems little reason to limit the lifestyle of patients with haemangiomata except when giant haemangiomata have replaced large amounts of liver substance. Even in this group the danger may be more perceived than real. Details of the natural history of hepatic haemangioma remain unclear. Symptom development may be influenced by capsular involvement and possibly by size but other factors are likely to be involved. Our understanding of this lesion would be improved if patients with giant haemangiomata (> 4 cm) or with symptomatic haemangiomata were included in longitudinal studies under the supervision of specialist hepatobiliary units.
524
TAIT ETAL.
References I . OSCHNER J. C. & HALVERT B. (1958) Cavernous haemangioma of the liver. Surgery 43, 577-82. 2. 1sii.m K. G. & RANKL. (1975) Benign tumours of the liver. Med. Clin. North Am. 59, 995-1013. 3. ZAFRANI C. S . (1989) Update on vascular tumours of the liver. J . Hepatology 8, 125-30. J. & HOLLANDS M. J. (1990) 4. LITTLEJ . M . , KENNY Hepatic incidentaloma: A modem problem. World J . Surg. 14, 448-5 1. 5 . FRONTD., ROYALH. D., ISRAEL0. et al. (1981) Scintigraphy of hepatic haemangiomas: The value of Tc-99m labelled red blood cells. J . Nucl. Med. 22, 684-7. 6. ENGELM., MEEKSP. S . , SANDLER M. et al. (1983) Differentiation of focal intrahepatic lesions with 99mTc red blood cell imaging. Radiology 46, 777-82. 7. FARLOW D. C., CHAPMAN P. R., GRUENWALD S . M. et al. (1990) Investigation of focal hepatic lesions: Is tomographic red blood cell imaging useful? World J . Surg. 14. 463-7. 8. FRERICHS F. T. (1861) A Clinical Treatise on Disease of the Liver, Vol. 2 . Translation by Murchison C. New Sydenham Society, London. 9. HOBBSK . E. F. (1990) Hepatic haemangiomas. World J . Surg. 14, 468-7 1. 10. BORMMAN P. C., TERBLANCHE J . , BLUMCART R. et a / . (1987) Giant haemangiomas: Diagnostic and therapeutic dilemmas. Surgery 101, 445.
1 1 . INATsUKl s., TODO
12. 13. 14.
15.
16. 17.
18. 19. 20.
s. & STARZL T. (1990) Excisional therapy for benign hepatic lesions. Surg. Gynecol. Obsrer. 171, 240-6. FOSTERJ. H. (1988) Benign liver tumours. In: Surgery of the Liver and Biliary Tract (Ed. L. H. Blumgart), pp. 1 1 15-27. Churchill Livingstone, Edinburgh. BUFFET C., FRITCEH J. & ERIENNE 3. P. (1982) Hemangiomes geants du foi chez l’adult. A propose de 3 cas. Gasfroenterol.Clin. Biol. 6 , 531-8. ITAIY . , OHTOMO K . , ARAHI T. etal. (1983) Computed tomography and sonography of cavernous haemangioma of the liver. Am. J . Roentgenol. 141, 315-20. SCHWARZ S . I. & HUSSEN W. C. (1987) Cavernous haemangioma of the liver: Single institution report of 16 resections. Ann. Surg. 205, 456-65. Editorial. Hepatic haemangioma - A suitable case for treatment? Lancet Oct 1988; 882-4. BAESTA. C., MARCHEL G. J . , WILMSG. I. et al. (1982) A comparative study of angiography and dynamic computed tomography in liver haemangioma. A report of 15 cases. J . Comput. Tomogr. 6 , 245-55. STARZL T. E., KOEPL. J., WEILR. et al. (1980) Excisional treatment of cavernous haemangioma of the liver. Ann. Surg. 192, 25-7. ADAMY. G . , Hueos A. G. & FORTNER J. G. (1970) Giant haemangiomas of the liver. Ann. Surg. 172, 239. SEALR., LAMLR A., ATLANJ . E. el a / . (1985) Nuclear magnetic resonance imaging of liver haemangiomas. J . Nucl. Med. 26, 1 1 17-22.
Aust. N.Z. J . Surg. 1992,62,521-524
ORIGINAL ARTICLES HEPATIC CAVERNOUS HAEMANGIOMA: A 10 YEAR REVIEW N.
TAIT, A.
J.
RICHARDSON, G . MUGUTI*AND
J . M.
LITTLE
Department of Surgery, Westmead Hospital, Westmead, New South Wales. Australia and *Department of Surgery, Mpilo Hospital, Bulawayo, Zimbabwe Between January 1981 and July 1991, 61 patients with hepatic haemangiomata were examined at Westmead Hospital. There were 14 males (22%) and 47 females (78%). The age range was 26-85 years with a median of 49 years. Forty-one had abdominal symptoms but these could be attributed to a haemangioma in only seven cases. There was at least one subcapsular lesion in 17 (28%). Six of the seven symptomatic lesions were subcapsular and five of these were giant haemangiomata (i.e. more than 4 cm in greatest diameter). One large symptomatic lesion was intrahepatic. No association was observed between hepatic haemangiomata and other hepatic or extrahepatic diseases. Haemangiomata were resected from six patients, four of whom were symptomatic. Symptoms improved in all four but did not resolve completely in any. Follow-up ranged from nil in five patients to 108 months in one. The median follow-up was 12 months after initial diagnosis. Ten patients showed evidence of change in their lesions or symptoms while under observation. Only three had worsening symptoms or suspected change in size of a haemangioma. This study highlights the benign, static nature of most hepatic haemangiomata. When this lesion is suspected, the diagnosis should be confirmed with ultrasound (US) and labelled red blood cell scanning (RBCS). Referral for evaluation by a specialist hepatobiliary surgery unit is necessary when symptoms are intolerable, increasing size is definitely demonstrated or the diagnosis is uncertain and cannot be established without specialized investigations. Bleeding into or from these lesions is rare. Key words: hepatic haemangioma, liver.
Introduction
Methods
From January 1981 to July 1991, 61 patients with Hepatic cavernous haemangioma is the most comhepatic space occupying lesions examined at Westmon primary hepatic t ~ m o u r . ' - In ~ this era of mead Hospital were found to have hepatic cavernnon-invasive, readily available organ imaging it is ous haemangiomata. A retrospective analysis of frequently an incidental finding4 The use of highly their case records was carried out. The following sensitive and specific imaging such as abdominal information was documented: age; sex; symptoms; ultrasound (US) and labelled red blood cell scanning (RBCS) leaves few in diagnostic d o ~ b t . ~ , ~ . ' past and present liver disease or disease that may affect the liver; sex steroid use; mode of presentaAlthough French's statement that: 'The pathologition; method of diagnosis: site, size and number of cal and clinical significance of the cavernous lesions; histology; follow-up and changes in symptumour is very slight" appears accurate, the natural toms; and tumour size or tumour number during history of this lesion is still poorly understood. Fear follow-up. of a more sinister neoplastic process and the threat of spontaneous rupture still troubles doctors and their patients, even where the organ imaging diagResults nosis of haemangioma is certain. This anxiety prompts many referrals to specialist hepatobiliary A G E A N D SEX D I S T R I B U T I O N units. There has been little evidence to support size at There were 61 patients of whom 47 were women. diagnosis, uncertain change in size during followThe median age of all patients was 49 years (range up, minor symptoms or fear of rupture as indica26-85). tions for resection of hepatic haemangioma.'-" SYMPTOMS Hepatic resection is now resorted to infrequently. Nineteen patients had no abdominal symptoms. Correspondence: Dr N. Tait, Department of Surgery, WestForty-one (67%) had abdominal pain or discomfort mead Hospital, Westmead, NSW 2145, Australia. but these symptoms were thought likely to arise from haemangioma in only seven. Four of these Accepted for publication 15 January 1992
522
TAIT ETAL..
complained of worsening pain during follow-up and ultimately underwent resection of their haemangiomata for symptom control. Another patient had a symptomatic colorectal hepatic metastasis and 33 had symptoms from other non-hepatic intraabdominal conditions. All the symptomatic patients had pain or discomfort at the time of presentation. No symptomless patient developed pain during the observation period. MODE OF DETECTION
Fifty-five patients (90%) had haemangiomata detected by organ imaging. Imaging procedures included ultrasound (US), labelled red blood cell scanning (RBCS), computerized axial tomographic (CAT) scanning and angiography . Magnetic resonance imaging (MRI) was used in only three cases. Ultrasound and RBCS were the most commonly used procedures. All patients diagnosed by organ imaging had at least two imaging procedures. Six had hepatic haemangiomata detected during laparotomy for unrelated causes. LIVER D I S E A S E S
Only three patients had histories of concurrent or past liver disease. One had a metastasis from colorectal carcinoma. Two had past histories of hepatitis, one of whom had hepatitis A and the other had hepatitis of uncertain nature. Five patients had other illnesses that may affect the liver including: congestive cardiac failure (1); Cushings disease (1); advanced chronic obstructive airways disease (1); and alcoholism without clinically established liver disease (2). None of these patients had symptomatic haemangiomata and none required treatment of hepatic haemangioma. S E X STEROID U S E
Of the 45 women, information on oral contraceptive medication (OCM) was available in only 25. Ten of the 25 had a history of OCM use. Two others had used progesterone preparations for treatment of menstrual difficulties. One male had used an oestrogenic medication for treatment of prostatic carcinoma. All four patients who had haemangiomata resected because of worsening symptoms had at some time used OCM. Two were using OCM when they developed symptoms. Cessation of OCM consumption did not reduce their symptoms or slow symptom progression. One of the two went back on OCM with no apparent effect on symptoms or on the size of her haemangioma. The other two patients had taken OCM for only a few months each. Both had stopped OCM use 10 years before developing symptoms from hepatic haemangioma.
NUMBER OF H A E M A N G I O M A T A
One hundred and eight haemangiomata were diagnosed in the 61 patients. The lesions were solitary in 38 (62%) and multiple in 23 cases. There was at least one subcapsular lesion in 17 (28%) cases. Six of the seven symptomatic patients had haemangiomata reaching the liver capsule. Three of the six had more than one lesion. One symptomatic patient had a 6.5 cm intrahepatic haemangioma with dense adhesions fusing the overlying liver surface to the diaphragm. SIZE
The diameter of the haemangiomata was measured on organ imaging in 55 cases or on the fixed and cut resected specimen in six. Median diameter was 3.5 cm with a range of 1-16cm. Six of the seven symptomatic lesions were giant haemangiomata with a greatest diameter of more than 4cm. The four lesions that were excised because of increasing symptoms measured 16, 8, 6.5 and 5cm respectively. There were 19 asymptomatic haemangiomata measuring more than 4 cm. Four of these were subcapsular. TREATMEN 1
Fifty-five patients received no active treatment. Of the six who had haemangiomata removed, four had resections of symptomatic lesions. Two had small lesions that were incidental findings at laparatomy. These were excised to exclude hepatic spread of unrelated intra-abdominal neoplasms. RESULTS OF T R E A T M E N T
All four patients undergoing elective resection of symptomatic haemangiomata reported subjective improvement. Symptoms did not fully resolve in any, although the benefit gained was considered substantial and worthwhile by all four. In particular, all four required less analgesic medication after the operation and two felt that reduction in pain had led to major improvements in social and family life. There were two minor complications from treatment, one wound haematoma and one wound infection. OBSERVATION A N D FOLLOW-UP
After the initial diagnosis, only seven patients, all with small lesions, two of which were excised at laparatomy, received no hepatic follow-up. Median follow-up for the remaining 54 patients was 12 months (range 1-108 months). During follow-up, 44 patients had no change in their symptoms or in the size or number of their lesions. Ten patients showed some change, including: increased symptoms (4) and possible increased size (2); a second lesion found on follow-up investigations (3); de-
523
HEPATIC HAEMANGIOMA
creased lesion size (1); and disappearance of a lesion (2). All four patients with increased symptoms had worsening right upper quadrant pain. On previous imaging two had shown growth of their lesions but review of the imaging did not support this. In one, the initial measurement had not been of the greatest diameter of the haemangioma. In the other, different measurements had been made on non-comparable equipment. nisroLoG'Y
Typical histologic changes of cavernous haemangioma were seen in all six resected lesions. One was largely sclerotic but areas of sclerosis were also seen in the other five. All four symptomatic lesions that were excised showed small areas of old and new thrombosis. None had acute inflammatory changes, evidence of old or fresh necrosis or of bleeding into the lesions. Three of the four reached the liver capsule. One was intrahepatic but the liver surface over the haemangioma was fused to the diaphragm by dense adhesions. The subcapsular liver, between the capsule and the superficial surface of the haemangioma was histologically normal.
Discussion The reported incidence of he atic haemangioma has been between 2 and 7%."' The benign nature of these lesions has in the past meant that most went undetected. As imaging techniques improve and become relatively cheaper and more available hepatic space occupying lesions will be found in increasing The most common lesion among these will be the hepatic haemangi~ma.~.'Of the haemangiomata diagnosed, less than 1% will present acutely due to spontaneous rupture and free intraperitoneal Even centres that collect major hepatic surgery referrals will see few haemangiomata present with life-threatening complications. Inatsuki et al., reporting from an institution that expected to perform over 100 hepatic resections in 1991, studied 219 patients undergoing resection of benign hepatic lesions. One hundred and fourteen had cavernous haemangiomata. Ten of these 114 had presented with haemoperitoneum, eight occurring after 'unwise needle biopsy or open biopsy'. Therefore, only two of 114 patients who had a cavernous haemangioma resected presented with spontaneous bleeding. Bleeding in the other eight was iatrogenic, caused mainly by needle biopsies, that frequently failed to make the diagnosis. Inatsuki" and Hobbs' advise that biopsy of a cavernous haemangioma should be avoided. For the majority of patients in whom hepatic haemangioma is suspected, a definitive diagnosis can be reached by using US and RBCS which are economical and without the risk or discomfort of
4
'
invasive These provide a highly sensitive and specific diagnostic combination that can be easily repeated on follow-up, though some large lesions may present a picture of mixed echogenicity on US. In these cases RBCS should still be diagnostic, though angio CT, angiography or MRI may be indicated if a second confirmatory test is required. I 3 , I 4 Once the diagnosis is established, very few patients should require liver resection. Surgical intervention should be avoided unless a haemangioma presents with bleeding, incapacitating symptoms, alteration in size or cannot be accurately d i a g n ~ s e d . ' ~ ' ' ~For ' ~ ,the ~ ~ few that do require treatment, excision is the most suitable option if it can be performed effectively and safely with few complications and low mortality. Non-excisional techniques such as radiation, hepatic artery ligation and embolization have less chance of symptom control and are contraindicated if there is any doubt about the diagnosis. In this series, no patient required operation to confirm the diagnosis. Two of the six resected lesions were incidental findings, confirmed during operation for unrelated abdominal malignancy. In these cases the lesions were small and partly sclerotic. They were resected for histopathologic examination to exclude secondary malignancy. The four patients who had haemangiomata resected because of increasing symptoms were all females with large lesions which were subcapsular in three. None had liver disease, signs of bleeding or necrosis or were the cause of diagnostic doubt. Capsular involvement and size cannot be the only determinants of symptom development as 19 of the 25 giant lesions with diameters greater than 4cm were asymptomatic, even though four of these reached the liver capsule. I 7 3 l 8 A clinician dealing with a hepatic space occupying lesion needs to reach a definitive diagnosis. This can be achieved in more than 90% of haemangiomata using US and RBCS and,6 if necessary, CAT scanning,I4 angiography" or MRI.*' Haemoperitoneum due to rupture of a hepatic haemangioma is a rare event most often caused by a biopsy needle.9,",'3 There seems little reason to limit the lifestyle of patients with haemangiomata except when giant haemangiomata have replaced large amounts of liver substance. Even in this group the danger may be more perceived than real. Details of the natural history of hepatic haemangioma remain unclear. Symptom development may be influenced by capsular involvement and possibly by size but other factors are likely to be involved. Our understanding of this lesion would be improved if patients with giant haemangiomata (> 4 cm) or with symptomatic haemangiomata were included in longitudinal studies under the supervision of specialist hepatobiliary units.
524
TAIT ETAL.
References I . OSCHNER J. C. & HALVERT B. (1958) Cavernous haemangioma of the liver. Surgery 43, 577-82. 2. 1sii.m K. G. & RANKL. (1975) Benign tumours of the liver. Med. Clin. North Am. 59, 995-1013. 3. ZAFRANI C. S . (1989) Update on vascular tumours of the liver. J . Hepatology 8, 125-30. J. & HOLLANDS M. J. (1990) 4. LITTLEJ . M . , KENNY Hepatic incidentaloma: A modem problem. World J . Surg. 14, 448-5 1. 5 . FRONTD., ROYALH. D., ISRAEL0. et al. (1981) Scintigraphy of hepatic haemangiomas: The value of Tc-99m labelled red blood cells. J . Nucl. Med. 22, 684-7. 6. ENGELM., MEEKSP. S . , SANDLER M. et al. (1983) Differentiation of focal intrahepatic lesions with 99mTc red blood cell imaging. Radiology 46, 777-82. 7. FARLOW D. C., CHAPMAN P. R., GRUENWALD S . M. et al. (1990) Investigation of focal hepatic lesions: Is tomographic red blood cell imaging useful? World J . Surg. 14. 463-7. 8. FRERICHS F. T. (1861) A Clinical Treatise on Disease of the Liver, Vol. 2 . Translation by Murchison C. New Sydenham Society, London. 9. HOBBSK . E. F. (1990) Hepatic haemangiomas. World J . Surg. 14, 468-7 1. 10. BORMMAN P. C., TERBLANCHE J . , BLUMCART R. et a / . (1987) Giant haemangiomas: Diagnostic and therapeutic dilemmas. Surgery 101, 445.
1 1 . INATsUKl s., TODO
12. 13. 14.
15.
16. 17.
18. 19. 20.
s. & STARZL T. (1990) Excisional therapy for benign hepatic lesions. Surg. Gynecol. Obsrer. 171, 240-6. FOSTERJ. H. (1988) Benign liver tumours. In: Surgery of the Liver and Biliary Tract (Ed. L. H. Blumgart), pp. 1 1 15-27. Churchill Livingstone, Edinburgh. BUFFET C., FRITCEH J. & ERIENNE 3. P. (1982) Hemangiomes geants du foi chez l’adult. A propose de 3 cas. Gasfroenterol.Clin. Biol. 6 , 531-8. ITAIY . , OHTOMO K . , ARAHI T. etal. (1983) Computed tomography and sonography of cavernous haemangioma of the liver. Am. J . Roentgenol. 141, 315-20. SCHWARZ S . I. & HUSSEN W. C. (1987) Cavernous haemangioma of the liver: Single institution report of 16 resections. Ann. Surg. 205, 456-65. Editorial. Hepatic haemangioma - A suitable case for treatment? Lancet Oct 1988; 882-4. BAESTA. C., MARCHEL G. J . , WILMSG. I. et al. (1982) A comparative study of angiography and dynamic computed tomography in liver haemangioma. A report of 15 cases. J . Comput. Tomogr. 6 , 245-55. STARZL T. E., KOEPL. J., WEILR. et al. (1980) Excisional treatment of cavernous haemangioma of the liver. Ann. Surg. 192, 25-7. ADAMY. G . , Hueos A. G. & FORTNER J. G. (1970) Giant haemangiomas of the liver. Ann. Surg. 172, 239. SEALR., LAMLR A., ATLANJ . E. el a / . (1985) Nuclear magnetic resonance imaging of liver haemangiomas. J . Nucl. Med. 26, 1 1 17-22.
