Management of cavernous hemangioma of the liver

Management of cavernous hemangioma of the liver

Management of Cavernous Hemangioma of the Liver Mika N. Sinanan, MD, Thomas Marchioro, Cavernous hemangioma of the liver was diagnosed in 12 of 60 ...

612KB Sizes 1 Downloads 83 Views

Management of Cavernous Hemangioma of the Liver Mika N. Sinanan,

MD,

Thomas Marchioro,

Cavernous hemangioma of the liver was diagnosed in 12 of 60 patients (20 percent) evaluated for surgery of neoplastie liver disease. All were female, from 29 to 77 years old. Six patients presented with abdominal pain and seven had taken estrogens. Indications for surgery included uncertain diagnosis, symptoms, large lesion greater than or equal to 6 cm, and hypoproliferative anemia. Three right lohectomies, one left lateral segmentectomy, one open biopsy, and one right trisegmentectomy were performed. There were no deaths, one sukphrenic abscess, and one bile leak. The remaining seven patients were observed and at 2 to 6 years postoperatively had followed a heqtign course. Resectional tlrerapy may he considered for superficial large or symptomatic lesions in the appropriate patient, hut most hepatic hemangiomas follow a henign course.

MD, Seattle, Washington

avernous hemangioma of the liver is the most common C benign tumor of the liver [I]. Most of these lesions are small and asymptomatic. Patients are generally referred for surgical evaluation because the hemangioma is large, symptomatic, or because the diagnosis is in question. The natural history of hemangiomas and the indications for resection or ablative therapy have been debated in the literature. What has become evident is that hepatic hemangioma can usually be diagnosed and localized by noninvasive techniques, may be hormonally driven, and most pose little risk to the patient. MATERIAL

AND METHODS

Records detailing the 20-year experience of one surgeon in the management of neoplastic diseases of the liver at one institution were reviewed retrospectively. Patients carrying the diagnosis of hemangioma of the liver were selected for detailed chart review. Follow-up was sought by a questionnaire sent to the patient or family. Telephone confirmation was obtained in most cases. RESULTS

From the Department of Surgery, University of Washington, Seattle, Washington. Requests for reprints should be addressed to Thomas Marchioro, MD, Department of Surgery, RF-25 University of Washington, Seattle, Washington 98195. Presented at the 75th Annual Meeting of the North Pacific Surgical Association, Coeur d’Alene, Idaho, November 4-5, 1988.

Hemangioma was diagnosed in 12 of 60 patients (20 percent) referred for surgical evaluation with radiographic findings of a hepatobiliary neoplasm (Table I). Patients with hemangioma were all women and ranged in age from 29 to 77 years with a mean age of 50 years. The oidest patient undergoing major liver resection was 60 years old. Pain, often of a longstanding nature and vague, diffuse character, was the reason for initial radiographic investigation in 6 of 12 patients. One patient had a palpable abdominal mass and the remaining five patients had hepatic lesions detected serendipitously during investigation for other problems, including symptomatic cholelithiasis and anemia in one patient each. No patients reported fever or sweats, jaundice, or other specific gastrointestinal symptoms, and a history of calculous gallbladder disease was present in only two patients. Eight of 12 patients reported prior or concomitant estrogen use either as replacement after oophorectomy (in 6 patients) or in birth control pills (in 2 patients). Laboratory evaluation including liver function tests was unremarkable. Ultrasound examination was commonly used for screening. Each patient was also evaluated by computed tomography scan (recently by a dynamic bolus technique) and 10 of 12 were evaluated by selective hepatic angiography [Z]. Patients with symptoms and patients under consideration for resection underwent the most extensive preoperative radiologic evaluation to define regional anatomy and concomitant disease. Radionuclide scans in two patients were of little additional benefit. On the basis of preoperative radiographic findings,

THE AMERICAN JOURNAL OF SURGERY

VOLUME 157 MAY 1989

519

TABLE I ttepatobiliary

Metastatic carcinoma Hemangioma Hepatoma Cholangiocarcinoma Gallbladder carcinoma Fibroncdular hyperplasla Miscellaneous*

Neoplasm8 Considered for Resection

Hepatlc Resection

Laparotomy Without Resection

9 5 4 1 1 1 2

11 1 4 6 1 ... 2

Nonoperative Management 1 6 1

.. .. 4

Total 21 12 9 7 2 1 6

Three undiagnosed liver masses not blopsled, one adenoma resected, one hamartoma resected, two laparotomies with liver biopsy, and one cyst excision. l

Figure 1. Hepatk hemangkma In a symptomatk patlent. Left, enhanced computed tomography appearance of the right hepatk lobe glant hemangkma. Rtght, cotton-wool pattern on selective hepatk angkgram.

particularly a characteristic computed tomographic and Standard vascular isolation and linger fracture techangiographic appearance, the mass was identified as a niques were used for resection. hemangioma in 10 of 12 patients. One open biopsy and There were no deaths and postoperative complications one resection (lateral segmentectomy, left lobe) were per- included atelectasis with fever in two patients, one proformed because of diagnostic uncertainty. longed bile leak, and a subphrenic abscess. Three- to 9Nine of 12 hemangiomas were localized to the right year follow-up (mean 5 years) revealed mild wound dislobe of the liver and one extended from the right lobe into comfort in several patients and one episode of partial the medial segment of the left lobe. All were solitary. small bowel obstruction occurring 1 year after hepatic Resected lesions measured 6 to 12 cm in greatest dimen- surgery. A recurrent hepatic lesion detected 1 year after sion, significantly larger than the size range (2 to 8 cm) of resection had remained stable for 4 years. unresected lesions. The radiographic hepatic anatomy of The seven remaining patients did not undergo reseca resected hemangioma from a patient eventually treated tion for several reasons: the lesion was thought inconsewith a right hepatic lobectomy is shown in Figure 1. quential in two patients, estrogen-containing medication Preoperative diagnosis by needle biopsy was unsuccessful was discovered and stopped in two patients, the lesion in one of the radiographically indeterminate lesions. exceeded feasible anatomic resection limits in one paIndications for surgical therapy included symptoms tient, one patient was a poor operative risk, and one paalone in one patient, confirmation and treatment of a tient refused surgery. Figure 2 shows the stable computed symptomatic lesion suspected of being a hemangioma in tomography image of a lesion in a patient followed after one patient, large, superficiai lesions (6 cm or larger) in withdrawal of estrogen medication. Two- to 6-year foltwo patients, and treatment of a hypoproliferative anemia low-up (mean 3 years) for six patients (one lost to followin one patient. One open biopsy was carried out. Resec- up) in this group has shown little change in clinical status tions included three right hepatic lobectomies, a right or symptoms. Follow-up ultrasound or computed tomogtrisegmentectomy, and a left lateral segmentectomy. raphy examinations in four of seven unresected patients 520

THE AMERICAN

JOURNAL

OF SURGERY

VOLUME

157

MAY 1989

have shown minimal progression in the size of the lesion except in one patient refusing treatment at last follow-up. For the six patients without a tissue diagnosis of hemangioma, characteristic radiographic studies and a benign course have been taken as evidence for hemangiomas [3]. COMMENTS Autopsy series have identified hepatic hemangiomas in up to 7 percent of unselected adults [4]. They are soft blue-purple lesions and are usually solitary [5]. Most are less than 4 cm in diameter and are not detected during life. The typical patient with hemangioma is an otherwise healthy woman in the fourth to sixth decade of life [a. Associated symptoms occurred in half the patients in this series and have been reported in up to 67 percent of patients evaluated for surgical treatment [7]. They are usually nonspecific and of long duration. Although only one patient in this group had a palpable mass, other series indicate this finding in up to half of the surgically treated patients [ 71. Eight of 12 patients had significant exposure to estrogens. The association of female sex, endogenous estrogen in pregnancy, and exogenous estrogens with hepatic hemangioma growth or recurrence [I,5,8] has been recently reviewed by Conter and Longmire [9]. Although they advocated surgical resection for hormonally driven hemangiomas, two patients in the present study were treated by withdrawal of estrogen and observation and followed a benign course. The possibility that estrogen acts as a trophic factor for hemangioma growth suggests that exogenous estrogen withdrawal, and possibly estrogen receptor blockade by specific antagonists, may slow growth of the lesion. This, however, remains to be tested. Other clinical presentations of hemangioma include obstructive jaundice, biliary colic, gastric outlet obstruction, hemobilia, abscess or thrombosis of the hemangioma, and spontaneous rupture [ 10-131. None were present in this seTHE AMERICAN

ries. Such complications, although rare, provide the strongest indication for intervention. Identification of hemangioma by laboratory examination has been unrewarding. Multimodality radiologic evaluation remains the principal technique for diagnosis and localization and has been recently reviewed [3,14]. Most patients in this series were screened initially with ultrasonography. Hypoechoic, hyperechoic, or a mixed appearance of the lesion are all common findings on sonographic examination of hemangiomas [ 151. Definitive diagnosis was achieved in 10 of 12 patients with a combination of computed tomography scan and angiography. Sequential computed tomography images slowly opacifying to isodensity represent a hemangioma with a sensitivity and specificity of over 90 percent for lesions over 2 cm in diameter [la]. Lesions less than 2 cm are more accurately detected by ultrasonography. Selective hepatic angiography, though invasive, is reported to be as accurate as dynamic bolus computed tomography (with the same 2 cm minimum size limitation), permits specific arterial dye injection for angiocomputed tomography, and provides a map of the hepatic vascular anatomy necessary for resection [ 171. Alternative radiographic techniques which were not utilized in this study but which have been reported include the highly specific technetium-99m red cell scintigraphy and Tz-weighted image magnetic resonance imaging, both potentially useful for lesions less than 2 cm [la]. Radiographically guided needle biopsy, attempted in one patient in this series, is also being used for definition of indeterminate lesions [ 191. A reasonable diagnostic plan for a small, asymptomatic liver lesion suspected of being a hemangioma might include dynamic bolus computed tomography scan together with scintigraphy, magnetic resonance imaging, or guided needle biopsy for confirmation. Larger lesions presenting with symptoms or as a mass should be evaluated with dynamic (arterial or veJOURNAL

OF SURGERY

VOLUME

157

MAY 1989

521

SINANAN AND MARCHIORO

nous) bolus computed tomography and angiography since this group is more likely to require resection. Although most hemangiomas are small, asymptomatic, and deserve only regular follow-up, some lesions are clearly of greater consequence. Schwartz and Hussar [7] have reported hepatic resection for indications of uncertain diagnosis, increased size in association with rapid growth, significant symptoms, consumptive coagulopathy (Kasabach-Merritt syndrome), or mass effect. Bornman et al [20] have proposed similar guidelines for resection. Trastek et al [21] retrospectively reviewed a group of 49 patients with giant hemangioma of the liver, 13 of which were treated with resectional therapy. Unresected patients observed for 5 years had no increased long-term morbidity or mortality. Trastek et al found only 21 cases of spontaneous rupture in a literature review of complications attributed to hemangioma. Given the benign natural history of most hepatic hemangiomas, a synthesis of our experience and that of others cited above would suggest that surgical resection be reserved for (1) anatomically favorable lesions of indeterminate diagnosis, (2) superficial large lesions, particularly if increasing in size, and (3) lesions complicated by infection, rupture, coagulation abnormalities, or mass effect. External beam irradiation or occlusion of arterial inflow by surgical or angiographic means have all been reported as alternative treatments for unresectable hemangiomas [ 22,231. These techniques should, however, be reserved for use in the unusual circumstance where the anticipated consequences of expectant management outweigh the risk of intervention. The results of long-term follow-up in this study showed recurrence of hemangioma in only one resected patient and stable lesions in six of seven unresected patients. Adson [II] and others have noted a similar stable course for hepatic hemangiomas. This serves to emphasize that for the small stable lesion without severe systemic effects, withdrawal of supplemental estrogen treatment and close follow-up seems to be safe and reasonable. Should the lesion progress in an adverse manner, however, surgical resection is preferable to nonsurgical intervention, which should be reserved for the high-risk patient or nonresectable lesion. REFERENCES 1. Edmondson HA. Tumors of the liver and intrahepatic bile ducts. In: Atlas of tumor pathology. Sect. VII, fascicle 25. Washington, DC: Armed Forces Institute of Pathology, 1958. 2. Frenny PC, Marks WM. Hepatic hemangioma: dynamic bolus CT. AJR 1986; 147: 711-9.

3. Falappa P, Preziosi P, Cotroneo AR, Danza FM. Hepatic hemangiomas on 33 patients. Evaluation by ultrasound, nuclear medicine, computed tomography, and angiography. Diagn Imaging Clin Med 1983; 52: 245-54. 4. Feldman M. Hemangioma of the liver. Am J Clin Path01 1958; 29: 160-2. 5. Takagi H. Diagnosis and management of cavernous hemangio ma of the liver. Semin Surg Oncol 1985; 1: 12-22. 6. Starzl TE, Koep LJ, Weil R III, et al. Excisional treatment of cavernous hemangioma of the liver. Ann Surg 1980; 192: 25-7. 7. Schwartz SI, Hussar WC. Cavernous hemangioma of the liver. A single institution report of 16 resections. Ann Surg 1987; 205: 456-65. 8. Morley JE, Meyers JB, Sack FS, Kalk F, Epstein EE, Lannon J. Enlargement of cavernous hemangioma associated with exogenous administration of estrogens. S Afr Med J 1974; 48: 695-7. 9. Conter RL, Longmire WP Jr. Recurrent hepatic hemangiomas. Possible association with estrogen therapy. Ann Surg 1988; 207: 115-9. 10. Grieco MB, Miscall BG. Giant hemangiomas of the liver. Surg Gynecol Obstet 1978; 147: 783-7. 11. Adson MA. Mass lesions of the liver. Mayo Clin Proc 1986; 61: 362-8. 12. Berliner L, el Ferzli G, Gianvito L, et al. Giant cavernous hemangioma of the liver complicated by abscess and thrombosis. Am J Gastroenterol 1983; 78: 835-40. 13. Sewell JH, Weiss K. Spontaneous rupture of hemangioma of the liver. Arch Surg 1961; 83: 729-34. 14. Brant WE, Floyd JL, Jackson DE, Gilliland JD. The radiologic evaluation of hepatic cavernous hemangioma. JAMA 1987; 257: 2471-4. 15. Bruneton JN, Drouillard J, Fenart D, Roux P, Nicolau A. Ultrasonography of hepatic cavernous haemangiomas. Br J Radio1 1983; 56: 791-5. 16. Ashida C, Fishman EK, Zerhouni EA, Herlong FH, Siegelman SS. Computed tomography of hepatic cavernous hemangioma. J Comput Assist Tomogr 1987; 11: 455-60. 17. Abrams RM, Beranbaum ER, Santos JS, Lipson J. Angiographic features of cavernous hemangioma of the liver. Radiology 1969; 92: 308-12. 18. Brown RKJ, Games A, King W, et al. Hepatic hemangiomas: evaluation by magnetic resonance imaging and technetium-99m red blood cell scintigraphy. J Nucl Med 1987; 28: 1683-7. 19. Solbiati L, Livraghi T, De Pra L, Ierace T, Masciadri N, Ravetto C. Fine-needle biopsy of hepatic hemangioma with sonographic guidance. AJR 1985; 144: 471-4. 20. Bomman PC, Terblanche J, Blumgart RL, Harries Jones EP, Kalvaria I. Giant hepatic hemangiomas: diagnostic and therapeutic dilemmas. Surgery 1987; 101: 445-9. 21. Trastek VF, van Heerden JA, Sheedy PF II, Adson MA. Cavernous hemangiomas of the liver: resect or observe? Am J Surg 1983; 145: 49-53. 22. Issa P. Cavernous haemangioma of the liver: the role of radiotherapy. Br J Radio1 1968; 41: 26-32. 23. Burrows PE, Rosenberg HC, Chaung HS. Diffuse hepatic hemangiomas: percutaneous transcatheter embolization with detachable silicone balloons. Radiology 1985; 156: 86-8.

End of the Society Papers

522

THE AMERICAN JOURNAL OF SURGERY

VOLUME 157 MAY 1989