Abstract Liver hemangiomas are the most common benign neoplasms of the liver and are often diagnosed incidentally on abdominal imaging. Most liver hemangiomas are small (<4 cm) indolent lesions that do not require further intervention or surveillance in the absence of symptoms. The management of giant liver hemangiomas (>4 cm), however, remains controversial. The natural progression of giant hemangiomas is not well defi ned and the risk of life-threatening complications, namely rupture and bleeding, is not clearly established. The development of symptoms such as abdominal pain, mass effects or life-threatening events such as rupture, bleeding or consumptive coagulopathy is an indication for surgical resection or intervention. Asymptomatic giant hemangiomas present a management dilemma as the risk of life-threatening complications is unknown. Several single centers studies have demonstrated the safety of expectant management of even very large asymptomatic hemangiomas. There is no evidence to support using size alone as an indication for resection. Although morbidity and mortality of liver resection for hemangioma has improved, the risks of surgical intervention need to be balanced against the rare risk of adverse events during expectant observation. Keywords Hemangioma • Liver resection • Observation • Complications • Liver • Giant hemangioma • Cavernous hemangioma • Surgery • Resection Introduction Liver hemangiomas are the most common benign neoplasms of the liver and are often diagnosed incidentally on abdominal imaging. Most liver hemangiomas are small (<4 cm) indolent lesions that do not require resection or routine follow-up in the absence of symptoms or complications. The management of giant liver J. Seal (*) Multi-Organ Transplant Institute , Ochsner Medical Center , 1514 Jefferson Highway , New Orleans , LA 70121 , USA e-mail: [email protected] 16 hemangiomas (>4 cm), however, remains controversial [ 1 – 3 ]. In contrast to smaller lesions, giant hemangioma s can become quite large (>40 cm), symptomatic and even cause life-threatening complications. The most common symptoms of giant hemangiomas are abdominal pain and mass effects of the tumor such as biliary obstruction [ 4 , 5 ], Budd Chiari syndrome [ 6 ], vena cava compression [ 7 ], and gastric outlet obstruction [ 8 ]. Although spontaneous rupture is very rare, traumatic rupture or intra-tumoral bleeding often necessitates emergency resection, trans-arterial embolization or both [ 9 ]. In some cases when hepatic resection is not anatomically feasible, both deceased and living donor liver transplant ation have been reported for treatment of symptomatic giant hemangioma [ 10 – 12 ]. In the absence of signifi cant symptoms, however, the rationale for surgical resection is less clear. The natural progression of hemangiomas and the risk of lifethreatening complications, namely rupture and bleeding, is not well established. Several authors have advocated for preventive surgical resection or enucleation based on tumor size, rate of growth and perceived risk of rupture despite a lack of supporting evidence. Also, progress in surgical technique, anesthesia and perioperative care has reduced the risk of death and major complications following liver resection , further lowering the risks associated with surgical intervention. An evidence- based approach to surgical management of asymptomatic giant hemangioma s must balance an estimate of the risk of expectant management in the absence of well-defi ned natural history of the disease with peri-operative morbidity and risks of surgical intervention. Search Strategy A literature search of English language publications from 1990 to 2014 was used to identity published data on management of liver hemangiomas using the PICO outline (Table 2.1 ). PubMed database was searched using the following terms [number of results]: “liver, hemangioma, surgery , observation ” [ 36 ], “liver, giant, hemangioma, surgery, observation” [ 12 ], “surgical, approach, liver, hemangioima” , “giant, liver, hemangioma, surgery, technique” [ 17 ]. Articles that did not present new data with respect to observation versus surgical management were excluded. Retrospective studies of benign liver lesions were included if a separate analysis of hemangiomas was provided. In an effort to be more comprehensive in the analysis, studies that included smaller hemangiomas (<4 cm) were considered in the review. Table 2.1 PICO table for management of giant liver hemangioma P (Patients) I (Intervention) C (Comparator group) O (Outcomes measured) Patients with giant liver hemangioma (>4 cm) without symptoms Surgical resection (anatomic resection, enucleation, liver transplanta tion ) Observation Post operative complications, persistent or recurrent symptoms, life threatening adverse events J. Seal 17 No prospective randomized studies comparing surgical resection and observation of asymptomatic giant hemangioma s were identifi ed. The PubMed search yielded six single institution and one multicenter retrospective study comparing surgical resection and observation of hemangiomas and four single center retrospective reviews of benign liver lesions that included analysis of hemangiomas. An additional two retrospective single center reviews of resection for hemangioma were included to better assess the rate of complications following resection. The data were classifi ed using the GRADE system. Results Resection Versus Observation for Giant Hemangiomas Evidence comparing resection versus observation of asymptomatic giant hemangioma s is limited to retrospective single institution studies (Table 2.2 ) [ 13 – 20 ] and sub-analyses within single center retrospective studies of benign liver tumor s (Table 2.3 ) [ 21 – 24 ]. The primary outcomes in most of these studies included perioperative morbidity and mortality , resolution of symptoms in the treatment group or development of new symptoms and adverse events in the observation cohorts. None of the studies describe deliberate changes in treatment algorithms during the study period and no explicit comparison of practices in different eras. Thus, the interpretation of the data is limited to center specifi c and often ill-defi ned practices in the management of giant hemangiomas. Nevertheless, the cumulative trends in outcomes offers insight into the risks and morbidity of resection compared with observation and the natural history of giant hemangiomas. Table 2.2 Resection versus observation for liver hemangioma Author (year) N (total) Resection Observation N % Complication N Finding Yedibela (2013) 246 103 17 % 143 56 % new symptoms, 2 deaths from rupture Giuliante (2010) 74 40 10 % 34 No signifi cant increase in size of asymptomatic tumors during observation Schnelldorfer (2010) 289 56 14 % 233 Size not associated with adverse events Yoon (2003) 115 52 25 % 63 36 % resected patients to exclude presence of malignancy Terkivatan (2002) 49 11 27.3 % 38 Symptoms resolved in 12 patients Pietrabissa (1996) 78 16 NA 62 No rupture in observation group Yamagata (1991) 33 13 23 % 20 Observation group tumors <5 cm diameter 2 Is Surgery Indicated for Asymptomatic Giant Hepatic Hemangioma? 18 The most recent report comparing resection and observation by Yedibela et al. [ 14 ] reviewed 307 hemangiomas referred for evaluation. Elective surgery was performed in 103 patients, 62 of which were symptomatic and 41 asymptomatic. Of the asymptomatic patients undergoing resection, 11 were for diagnostic uncertainty, 9 for tumor enlargement and 21 for patient fear of future complications. There was no mortality in the post-operative period and symptom resolution was achieved in 88 % of symptomatic patients. In contrast to other studies, they report a 9 % rate of major complications in the observation group with two fatalities from traumatic hemangioma rupture. In this study, the overall trend was toward lower rates of complications in the surgical treatment group. Giuliante et al. [ 13 ] reviewed 74 consecutive referrals for hemangioma evaluation, with 34 undergoing observation and 40 treated surgically. As a regional hepatobiliary referral center, the authors discuss the bias that most giant hemangioma s were referred for symptoms, large size or diagnostic uncertainty and that the overall operative rate of 54 % is likely infl ated with underrepresentation of asymptomatic lesions. The primary indications for resection were abdominal symptoms, diagnostic uncertainty and tumor growth. Tumor growth has been used by some centers as an indication for resection, but the trajectory of growth over long term follow-up and its implication for risk of tumor rupture is uncertain. In this study, only 7 of 14 tumors demonstrating growth underwent resection, 5 of which had additional indications for resection (abdominal pain and Kasabach-Merritt syndrome). Of the seven tumors demonstrating growth in the observation group, the tendency to enlarge decreased over time and among all tumors in the observation group, there was no signifi cant increase in mean tumor size or development of new symptoms during the follow-up period. Another single institution review by Terkivatan et al. [ 19 ] supported the safety of expectant management of asymptomatic giant hemangioma s. During the study period, 11 patients were treated with resection , all of whom had abdominal symptoms attributed to the hemangioma except for one asymptomatic lesion that exhibited 5 cm of growth over 36 months follow-up. Thirty-eight patients were managed with observation . Importantly, there was no signifi cant difference in the mean diameter of tumors in the resection group compared with the observation group. Twelve of the observation patients had abdominal pain on initial assessment that was not Table 2.3 Studies of benign liver lesions: analysis of hemangiomas Author (year) N (total) Resection Observation N % Complication N Finding Mezhir (2013) 151 60 NA 91 4.4 % of observed tumors were ultimately resected Terkivatan (2001) 103 25 24 % 78 25 cm tumor followed without need for resection Charny (2001) 97 39 21 % 58 63.3 % of pre-operative biopsies were indeterminate or incorrect Weimann (1997) 238 103 18.8 % 135 No episodes of bleeding in observation group J. Seal 19 attributed to the hemangioma and, in each case, the pain resolved during follow-up. None of the asymptomatic patients developed new symptoms during a mean follow- up period of 59 months. Schnelldorfer et al. [ 16 ] report the only contemporary series to assess the impact of hemangioma management on quality of life . A quality of life survey was administered with responses from 289 of 492 patients treated for hemangioma at their center, including patients treated with surgical resection or observation . In the observation group, 20 % of patients developed new-onset symptoms with 2 % being life-threatening. Post-operative complications occurred in 14 % of resected patients with 7 % being life threatening. Interestingly, the size of the hemangioma was not associated with adverse events in either group, lending credence to the notion that risk of rupture is not necessarily related to tumor size or rate of growth. The subjective quality of life survey was similar for resected and observed patients leading the authors to conclude that observation is preferred in most patient and surgical resection should be reserved for patient with symptoms or hemangioma-related complications. The consideration of tumor size and rate of growth have both been proposed as risk factors for tumor rupture and, thus, indications for resection in asymptomatic patients. Iwatsuki et al. [ 25 ] suggested benign hepatic lesions exceeding 10 cm should be resected based increased risk for internal bleeding, further growth or rupture, although this was not based on conclusive prospective observational data. In the more recent literature, an absolute size threshold as an indication for resection is not favored. The indication for resection of giant hemangioma was most often the development of symptoms. The rate of growth of the tumor remains a controversial indication for resection, although such recommendations are made without a clearly defi ned relationship between rate of growth and risk of rupture. Pietrabissa et al. [ 17 ] proposed resection for “rapidly growing” asymptomatic hemangiomas defi ned as a minimum 25 % increase in largest diameter over a period of 6 months. The rationale is based on the speculation that the risk for tumor rupture or diagnostic error is higher in patients with rapid growth. Although well reasoned, no clinical support is provided to substantial the claim of increased risk of rupture or diagnostic error with rapid growth, leaving the natural disease course of giant hemangiomas unresolved. Diagnostic uncertainty has been used as an indication for resection of vascular lesions of the liver when a defi nitive diagnosis cannot be made from radiologic studies, particularly in the setting of known extra-hepatic malignancies or risk factors for the development of liver malignancy ( cirrhosis , hepatitis, steatosis). With advances in imaging modalities and experience at high-volume centers, diagnostic uncertainty is much less common. In a subgroup analysis of a retrospective review of benign liver lesions at Memorial Sloan-Kettering Cancer Center, Mezhir et al. [ 22 ] report diagnostic uncertainty as the indication for surgical resection in 53.3 % of asymptomatic hemangiomas from 1992 to 2009. Importantly, though, only 12.5 % of patients resected for diagnostic uncertainty were seen in the last 10 years of the study, suggesting a signifi cant impact of advances in medical imaging. 2 Is Surgery Indicated for Asymptomatic Giant Hepatic Hemangioma? 20 Treatment of Giant Hemangiomas: Operative Approaches and Non-surgical Therapies The morbidity and mortality of hemangioma resection is an important consideration when determining appropriate treatment for giant hemangioma s, particularly asymptomatic lesions. Complications following surgical resection range from 10 % to 27.3 % in recent studies comparing resection and observation . In a recent multi- institutional review of surgical management of hepatic hemangiomas [ 26 ], the overall rate of Clavien grade 3 complication or higher was 5.7 % and included bile leak s and bleeding. The 30-day post-operative mortality was 0.8 % (N = 2). Although hepatic resection s for hemangioma can be performed safely, post-operative complications can be signifi cant with a nearly 1 % risk of death. The documented risks of surgery must be carefully weighted against the severity of symptoms and the very low risk of complications associated with observation. Evolution in operative techniques and application of non-surgical therapies may lead to a less invasive and safer approach to management of hemangiomas in the future. Though technical aspects of resection may vary greatly between centers, a single center study by Lerner et al. [ 27 ] documented an intra-institutional evolution toward enucleation over resection for large hemangiomas over time with more liberal use of infl ow occlusion, less intra-operative blood loss and less complications. Liver transplantation [ 10 – 12 , 28 ] for life threatening symptoms attributed to giant hemangioma s has been reported, further expanding therapeutic options to include lesions anatomically unresectable by conventional approaches. Laparoscopic [ 29 – 32 ] and robotic [ 33 ] approaches have been reported by several centers, although the impact of a minimal access approach in reducing operative risk or complications is not defi ned. The use of several non-surgical therapies for symptomatic giant hemangioma s has been reported. Transarterial embolization is widely used both in the setting of acute management of ruptured hemangiomas and as a pre-operative treatment to reduce vascular infl ow and decrease size [ 34 – 38 ]. Radiofrequency ablation has also been reported for symptomatic control of giant hemangiomas with promising results. Gao et al. [ 39 ] report a single institution initial experience with RFA treatment in giant hemangiomas. In this series, use of RFA in lesions >10 cm diameter had a 100 % complication rate including life-threatening complications of lower esophageal fi stula and acute respiratory distress syndrome. In smaller lesions (5–10 cm diameter), RFA was successful at controlling symptoms with only minor complications. Medical therapies such as the anti-angiogenic agent bevicizumab and the tyrosine kinase inhibitor sorafenib have also been reported to decrease the size of large hemangiomas, though published data is limited to case reports. Overall, even less invasive modes of treating giant hepatic hemangiomas will have some risk of complication that must be carefully considered in asymptomatic patients. J. Seal 21 Recommendations Surgical resection is a well established and accepted treatment of symptomatic giant hemangioma s. The relative risk of surgical intervention is balanced by a benefi t to the patient with expected improvement or resolution of symptoms. Management of asymptomatic giant hemangiomas remains controversial as the natural progression of the disease in not well described. The hypothesis that tumor size contributes to the risk of rupture is not supported in the literature and should not be used alone as an indication for surgical resection . Retrospective studies from many centers demonstrate that expectant management of asymptomatic giant hemangiomas is safe. Surgical resection or an alternative treatment modality should be considered if symptoms develop. Though several centers consider rapid growth of tumors during observation periods as an indication for resection, there is no evidence to suggest that growth alone presents additional risk to the patient if a defi nitive radiographic diagnosis has been made. A Personal View of the Data Surgical resection of giant symptomatic hemangiomas is widely accepted as the risks of surgery are balanced by a direct benefi t of symptom relief. Though resolution of symptoms after resection is not universal and varies in published reports, it is reasonable to expect a high rate of symptom resolution if the initial symptoms were appropriately attributed to the hemangioma. To that end, symptomatic patients with giant hemangioma s should be carefully assessed to ensure the symptoms, in particular non-specifi c abdominal pain , can be attributed to effects of the mass. The retrospective series highlighted in this chapter support the safety and prudence of deferring surgical resection in asymptomatic patients with giant hemangiomas. Yedibela et al. reported a higher rate of adverse events in patients under observation including two deaths from rupture, compared with other contemporary series that report no adverse events. Overall, the evidence indicates observation of even very large hemangiomas is safe. There is no evidence to support size alone as an indication for surgical resection in asymptomatic lesions. Impending complications such as vascular compression or gastrointestinal obstruction may be appropriate. Patient anxiety has been reported as an indication for surgical resection of asymptomatic lesions in some series. While anxiety can certainly be a signifi cant symptom, it is the responsibility of the surgeon to reassure patients regarding the very low risk of adverse events associated with observation and to manage patient anxiety non-operatively. Finally, the risk of rupture related to tumor growth is unclear, but to date no series has demonstrated that rapid growth of hemangiomas presents an increased risk if a defi nitive radiologic diagnosis has been made. Rapid growth may prompt further investigation to confi rm the diagnosis and exclude concern for malignancy, but there is insuffi cient evidence to support surgical resection on the basis of growth alone. 2 Is Surgery Indicated for Asymptomatic Giant Hepatic Hemangioma? 22 Recommendations • Symptomatic liver hemangiomas should be surgically resected if symptoms are attributed to the hemangioma itself (evidence quality high, strong recommendation). • Asymptomatic giant hemangioma s can be safely observed regardless of tumor size (evidence quality moderate, strong recommendation). • The size of the hemangioma (greatest diameter) alone should not be used as an indication for resection . (evidence quality moderate, strong recommendation) • Rate of growth alone should not be used as an indication for surgical resection (evidence quality low, weak recommendation). References 1. Hoekstra LT, Bieze M, Erdogan D, Roelofs JJTH, Beuers UHW, van Gulik TM. Management of giant liver hemangiomas: an update. Expert Rev Gastroenterol Hepatol. 2013;7:263. 2. Belghiti J, Cauchy F, Paradis V, Vilgrain V. Diagnosis and management of solid benign liver lesions. Nat Rev Gastroenterol Hepatol. 2014;11:737–49. 3. Toro A, Mahfouz A-E, Ardiri A, et al. What is changing in indications and treatment of hepatic hemangiomas. A review. Ann Hepatol. 2014;13:327. 4. Losanoff JE, Millis JM. Liver hemangioma complicated by obstructive jaundice. Am J Surg. 2008;196:e3. 5. Tang L, Zhou W-P. Education and imaging. Hepatobiliary and pancreatic: large cavernous hemangioma with obstructive jaundice. J Gastroenterol Hepatol. 2009;24:930. 6. Kim DY, Pantelic MV, Yoshida A, Jerius J, Abouljoud MS. Cavernous hemangioma presenting as Budd-Chiari syndrome. J Am Coll Surg. 2005;200:470. 7. Akbulut S, Yilmaz M, Kahraman A, Yilmaz S. Bilateral lower limb edema caused by compression of the retrohepatic inferior vena cava by a giant hepatic hemangioma. Int Surg. 2013;98:229. 8. Aydin C, Akbulut S, Kutluturk K, Kahraman A, Kayaalp C, Yilmaz S. Giant hepatic hemangioma presenting as gastric outlet obstruction. Int Surg. 2013;98:19. 9. Donati M, Stavrou GA, Donati A, Oldhafer KJ. The risk of spontaneous rupture of liver hemangiomas: a critical review of the literature. J Hepatobiliary Pancreat Sci. 2011;18:797–805. 10. Meguro M, Soejima Y, Taketomi A, et al. Living donor liver transplantation in a patient with giant hepatic hemangioma complicated by Kasabach-Merritt syndrome: report of a case. Surg Today. 2008;38:463. 11. Ferraz AAB, Sette MJA, Maia M, et al. Liver transplant for the treatment of giant hepatic hemangioma. Liver Trans. 2004;10:1436. 12. Vagefi PA, Klein I, Gelb B, et al. Emergent orthotopic liver transplantation for hemorrhage from a giant cavernous hepatic hemangioma: case report and review. J Gastrointest Surg. 2011;15:209. 13. Giuliante F, Ardito F, Vellone M, et al. Reappraisal of surgical indications and approach for liver hemangioma: single-center experience on 74 patients. Am J Surg. 2011;201:741. 14. Yedibela S, Alibek S, Müller V, et al. Management of hemangioma of the liver: surgical therapy or observation? World J Surg. 2013;37:1303. J. Seal 23 15. Yamagata M, Kanematsu T, Matsumata T, Utsunomiya T, Ikeda Y, Sugimachi K. Management of haemangioma of the liver: comparison of results between surgery and observation. Br J Surg. 1991;78:1223. 16. Schnelldorfer T, Ware AL, Smoot R, Schleck CD, Harmsen WS, Nagorney DM. Management of giant hemangioma of the liver: resection versus observation. J Am Coll Surg. 2010;211:724. 17. Pietrabissa A, Giulianotti P, Campatelli A, et al. Management and follow-up of 78 giant haemangiomas of the liver. Br J Surg. 1996;83:915. 18. Herman P, Costa MLV, Machado MAC, et al. Management of hepatic hemangiomas: a 14-year experience. J Gastrointest Surg. 2005;9:853. 19. Terkivatan T, Vrijland WW, Den Hoed PT, et al. Size of lesion is not a criterion for resection during management of giant liver haemangioma. Br J Surg. 2002;89:1240. 20. Yoon SS, Charny CK, Fong Y, et al. Diagnosis, management, and outcomes of 115 patients with hepatic hemangioma. J Am Coll Surg. 2003;197:392. 21. Charny CK, Jarnagin WR, Schwartz LH, et al. Management of 155 patients with benign liver tumours. Br J Surg. 2001;88:808. 22. Mezhir JJ, Fourman LT, Do RK, et al. Changes in the management of benign liver tumours: an analysis of 285 patients. HPB. 2013;15:156. 23. Weimann A, Ringe B, Klempnauer J, et al. Benign liver tumors: differential diagnosis and indications for surgery. World J Surg. 1997;21:983. 24. Terkivatan T, de Wilt JW, de Man RA, et al. Indications and long-term outcome of treatment for benign hepatic tumors: a critical appraisal. Arch Surg. 2001;136:1033. 25. Iwatsuki S, Todo S, Starzl TE. Excisional therapy for benign hepatic lesions. Surg Gynecol Obstet. 1990;171:240. 26. Miura JT, Amini A, Schmocker R, et al. Surgical management of hepatic hemangiomas: a multi-institutional experience. HPB. 2014;16:924. 27. Lerner SM, Hiatt JR, Salamandra J, et al. Giant cavernous liver hemangiomas: effect of operative approach on outcome. Arch Surg. 2004;139:818. 28. Longeville JH, de la Hall P, Dolan P, et al. Treatment of a giant haemangioma of the liver with Kasabach-Merritt syndrome by orthotopic liver transplant a case report. HPB Surg. 1997;10:159. 29. Acharya M, Panagiotopoulos N, Bhaskaran P, Kyriakides C, Pai M, Habib N. Laparoscopic resection of a giant exophytic liver haemangioma with the laparoscopic Habib 4× radiofrequency device. World J Gastrointest Surg. 2012;4:199. 30. Gadiyaram S, Shetty N. Laparoscopic resection of giant liver hemangioma using laparoscopic Habib probe for parenchymal transection. J Minim Access Surg. 2012;8:59. 31. Lanthaler M, Freund M, Nehoda H. Laparoscopic resection of a giant liver hemangioma. J Laparoendosc Adv Surg Tech A. 2005;15:624. 32. Patriti A, Graziosi L, Sanna A, Gullà N, Donini A. Laparoscopic treatment of liver hemangioma. Surg Laparosc Endosc Percutaneous Tech. 2005;15:359. 33. Giulianotti PC, Addeo P, Bianco FM. Robotic right hepatectomy for giant hemangioma in a Jehovah’s witness. J Hepatobiliary Pancreat Sci. 2011;18:112. 34. Lupinacci RM, Szejnfeld D, Farah JFM. Spontaneous rupture of a giant hepatic hemangioma. Sequential treatment with preoperative transcatheter arterial embolization and conservative hepatectomy. G Chir. 2011;32:469. 35. Panis Y, Fagniez PL, Cherqui D, Roche A, Schaal JC, Jaeck D. Successful arterial embolisation of giant liver haemangioma. Report of a case with fi ve-year computed tomography follow- up. HPB Surg. 1993;7:141. 36. Suzuki H, Nimura Y, Kamiya J, et al. Preoperative transcatheter arterial embolization for giant cavernous hemangioma of the liver with consumption coagulopathy. Am J Gastroenterol. 1997;92:688. 2 Is Surgery Indicated for Asymptomatic Giant Hepatic Hemangioma? 24 37. Vassiou K, Rountas H, Liakou P, Arvanitis D, Fezoulidis I, Tepetes K. Embolization of a giant hepatic hemangioma prior to urgent liver resection. Case report and review of the literature. Cardiovasc Interv Radiol. 2007;30:800. 38. Zhou J-X, Huang J-W, Wu H, Zeng Y. Successful liver resection in a giant hemangioma with intestinal obstruction after embolization. World J Gastroenterol. 2013;19:2974. 39. Gao J, Ke S, Ding X, Zhou Y, Qian X, Sun W. Radiofrequency ablation for large hepatic hemangiomas: initial experience and lessons. Surgery. 2013;153:78.