NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Holzheimer RG, Mannick JA, editors. Surgical Treatment: Evidence-Based and Problem-Oriented. Munich: Zuckschwerdt; 2001.

Cover of Surgical Treatment

Surgical Treatment: Evidence-Based and Problem-Oriented.

Show details

Neuroendocrine tumors metastatic to the liver

, M.D., , M.D., and , M.D.

Author Information

Neuroendocrine tumors are rare neoplasms presenting complex challenges to diagnosis and treatment. Even in case of metastatic spread to the liver there are some important differences in the nature of these tumors compared to gastrointestinal and pancreatic adenocarcinomas. They typically show 1) specific clinical symptoms and signs and a protracted course, 2) production of hormones and or amines, which are responsible for the specific clinical symptoms and signs and act as tumor markers, and 3) tumor growth inhibition and symptom relief may be achieved by specific blocking agents. About one-half of all neuroendocrine tumors are described as non-functioning meaning that the patients do not have hormone-related symptoms.

Primary tumors and frequency of neuroendocrine liver metastases

Spread to the liver may occur from neuroendocrine tumors of the foregut, midgut as well as hindgut (table I).

Table I. Classification of neuroendocrine tumors including carcinoids.

Table I

Classification of neuroendocrine tumors including carcinoids.

Neuroendocrine liver metastases are usually multiple and of varying size. In most cases both liver lobes are affected, but miliary seeding throughout the liver is seen only occasionally.

The carcinoid is the most common neuroendocrine tumor causing liver metastases, especially when of midgut origin. However, although appendix is the most common location for midgut carcinoids it most unusually gives rise to hepatic spread. Similarly, hindgut carcinoids, usually of rectal location, if less than 2 cm rarely metastasize to the liver.

Neuroendocrine pancreatic tumors vary in malignant potential (table II) and frequency of liver secondaries (table III).

Table II. Estimated risk of malignant transformation, characteristic neuropeptide and clinical signs and symptoms in neuroendocrine tumors.

Table II

Estimated risk of malignant transformation, characteristic neuropeptide and clinical signs and symptoms in neuroendocrine tumors.

Table III. Frequency of liver metastases from neuroendocrine tumors. The Lund series.

Table III

Frequency of liver metastases from neuroendocrine tumors. The Lund series.

Insulinomas seldomly cause liver metastases whereas such lesions are seen in more than 50% of the patients with other types of neuroendocrine pancreatic tumors by the time of diagnosis (table III).


Functioning neuroendocrine tumors are characterized by inappropriate release of biologically active polypeptides or amines causing the typical clinical presentation (table I).

Therefore, determination of the “hypersecreted” hormone or amine is mandatory.

Four non-invasive imaging studies are currently in use to detect the presence and to describe the extent of neuroendocrine liver metastases.

Ultrasonography harbors a low mean sensitivity in detecting such liver deposits which may be either hypoechoic or hyperechoic.

As neuroendocrine metastases are hypervascular they may be isodense with the normal liver parenchyma if contrast infusion is used during computerized tomography (CT). However, bolus intravenous infusion of contrast material and dynamic scanning through the liver during both arterial and portal phases of hepatic enhancement make it less likely that metastases will be missed.

Overall, CT has a sensitivity of 70–85%.

When appropriate pulse sequences are used, magnetic resonance imaging (MRI) is at least as good as CT for the detection of liver metastases in patients with neuroendocrine tumors.

Somatostatin receptor scintigraphy (SRS) is helpful in localizing primary neuroendocrine tumors especially those arising in the foregut. It is also valuable for detection of metastatic disease particularly, however, if present outside the liver. Positive experience with the above techniques varies between centers, so no general recommendations can be done regarding their use.

Selective hepatic arteriography demonstrates liver metastases with a sensitivity of 65% and is capable of visualizing lesions smaller than 5 mm in diameter. However, being an invasive investigation it has been largely supplanted by the newer above-mentioned less invasive modalities.

Preoperative histological or cytological diagnosis can be obtained by percutaneous puncture guided by ultrasonography or CT. The sensitivity of such procedures for the diagnosis of “endocrine tumor” is over 70%. A definitive drawback with the method is the occurrence of false negative results, which may be as common as one in four. Diagnostic laparoscopy is useful not only for staging liver involvement and peritoneal deposits but also for direct biopsy or targeting the biopsy needle.

Our present knowledge of the accuracy of available diagnostic tools rests on weak grounds as the restricted number of patients with neuroendocrine liver metastases preclude appropriate scientific quality control of the tests.

What is true, however, is that the ideal technique - non-invasive, sensitive, specific, safe, and inexpensive - does not exist. As no tests are accepted to the extent that their use is generally recommended policies vary among centers depending on the skill and interest of the individual examiners.

Surgical management

Resection of liver metastases from neuroendocrine tumors is attractive because the natural history of these tumors is protracted (table IV), clinical severity of the specific clinical signs (table I) correlates with tumor volume, and local and intrahepatic growth characteristics may allow complete resection.

Table IV. Survival in patients with neuroendocrine tumors and liver metastases. The Lund series.

Table IV

Survival in patients with neuroendocrine tumors and liver metastases. The Lund series.

Liver resection should, however, if done for cure be considered only after the primary tumor and its local extension have been controlled. In patients whose symptomatology cannot be managed by medical treatment liver surgery may be helpful even if not all hepatic tumor tissue can be excised. For palliative reasons cytoreductive hepatic surgery is recommended if careful preoperative imaging suggests that the primary and regional diseases are controlled and 90% or more of all hepatic metastases can probably be removed.

Whereas the reported symptomatic response to resection of neuroendocrine metastases uniformly has been prompt and gratifying, the survival benefit has not been clearly demonstrated.

However, the low incidence of these tumors and the protracted natural course represent real obstacles to a trial directly contrasting complete resection with no treatment in rigorously comparable patients. Recently, a significantly longer actuarial survival was found after curative resection (75% 5-year survival) as compared to unresected (40% 5-year survival) patients in a non-randomized study comprising carcinoids and islet cell tumors and where the two groups did not differ in age, pathology, primary tumor site, serum alkaline phosphatase, or percentage of the liver involved (grade C). In another study 5-year survival after potentially curative liver resection was 60% whereas 5-year survival following palliative resection was calculated to be almost 40% (Grade C). However, this kind of resective surgery should be undertaken only in experienced units where an operative mortality of less than 5% can be assured. Thus, there are studies suggesting not only symptom relief but also prolonged survival following potentially curative liver resections. Although none of the studies is conclusive we recommend resection, also for prolonged survival, in selected patients with neuroendocrine liver metastases.

If metastatic liver invasion is extensive and precludes adequate cytoreductive resection cryosurgery may be used either alone or in combination with resection. Experience reported today from non-randomized trials tell about a dramatic symptom relief and a more than 80% reduction in circulating tumor markers. More data is needed for evaluation of the future potential of this treatment.

Reports of liver transplantation for metastatic neuroendocrine tumors have been confined to small numbers of patients and short follow-up. It seems, though, as if the results are better in carcinoid patients than in those with pancreatic neuroendocrine tumors. As for liver transplantation done for other malignant diseases the recurrence rate is high also if the indication is neuroendocrine tumors.

Most authors agree that liver transplantation should be applied with caution in patients with metastatic liver disease, even when from a neuroendocrine primary.

This opinion is based not only on the reported results but also on the cost and on the lack of donor livers.

Liver transplantation should, however, be considered in young patients with low operative risk but only after failure of all available treatments and no evidence of other distant metastases.

Ischemic therapy, embolization, chemoembolization

As the liver has a dual vascular supply and liver metastases derive the majority of blood from the hepatic artery selective metastasis ischemia may be induced by occlusion of tumor vessels. With surgical hepatic artery ligation and hepatic dearterialization (which means ligation of not only the hepatic artery but of all collateral circulation to the liver and division of the hepatoduodenal ligament except the portal vein and the bile duct) significant palliation with relief of symptoms has been reported but only for a limited period. However, morbidity is high and mortality may occur. Temporary hepatic dearterialization (i.e. placement of strangulation slings around the hepatic artery, allowing temporary external occlusion of the vessel) provide the same relief of humoral symptoms and reduction of tumor growth but with significantly less morbidity. To make repeat tumor ischemia possible an implantable occluder allowing external repeated blockage of the hepatic artery flow can be put in place at a laparotomy. Ischemia is produced through the occluder for 16 hours every second month. The treatment can be continued for several years and often on an ambulatory basis. This technique caused a 50% decrease of 5-HIAA, symptom relief and reduced tumor bulk in a series of carcinoid patients. However, there is today general agreement that embolization with microspheres is preferable to ligation or occlusion of the hepatic artery. Embolization can be done selectively depending on tumor bulk and can be repeated. The response rate is 45% to 60% for tumor reduction and amelioration of clinical symptoms in parallel with reduced hormone or amine levels.

Chemoembolization implies that transcatheter arterial regional chemotherapy is immediately followed by embolization with microspheres or absorbable gelatin pledgets. Cisplatinum, mitomycin C, 5-fluorouracil, and doxorubicin have been used together with embolization in various studies. Toxicity has been considerable and long-term effects on survival have not been demonstrated probably reflecting the limited response rate of neuroendocrine tumors to conventional chemotherapy.

Thus, it remains to be shown if chemoembolization has any advantages to the effects obtained after embolization alone.

Medical therapy

Clinical trials of chemotherapy of malignant neuroendocrine tumors have been scarce because of the rarity of these tumors.

The specific drug of choice is streptozotocin developed during the 1960s for treating various non-resectable neuroendocrine pancreatic tumors, including insulinoma.

When used alone, streptozotocin generates response rates of 30% to 40%. It is more effective for islet cell tumors than for the carcinoid syndrome. When combined with 5-fluorouracil the response rate raises to 60% to 65% for a duration of about 2 years. Chemotherapy is still the first-line treatment in patients with unresectable progressive neuroendocrine pancreatic tumors. Overall, chemotherapy has turned out to be of limited value in patients with carcinoid tumors.

Somatostatin inhibits hormonal synthesis and release. It may also interfere with the peripheral effects of circulating hormones and in high-doses it has been suggested to induce apoptosis in neuroendocrine tumors.

Relief of the specific tumor symptoms occur in 35% to 80% of patients if standard doses of somatostatin are used (100 μg to 150 μg two or three times per day) and circulating hormone levels are reduced in about 60% (grade C).

Tachyphylaxis appears, however, to be a common problem as the doses used need to be increased with time. Side effects are few. Whether somatostatin has direct antitumor effects is not clearly shown. A long-acting (one month) somatostatin preparation has recently been introduced implying a more stable circulating somatostatin level and a markedly improved quality of life.

Interferon-α (INF-α) was originally introduced as a treatment of carcinoid tumors. It has, however, been reported to have beneficial effects also in patients with pancreatic neuroendocrine tumors. Reduction in circulating hormone levels are reported in 50%, in specific symptoms in 60%, and tumor reduction in 15%. Side effects are moderate.

When combined INF-α and somatostatin analogs are used improved responses are reported compared to when the drugs are used alone.

Surprisingly, only few trials on the effects of various medical treatments are prospective, randomized and these studies include a small number of patients. This is the reason why no general recommendations can be made and explains, furthermore, the diverging management policies practiced in various centers.

Management of non-functioning pancreatic endocrine tumors

Non-functioning islet cell neoplasms are defined as lesions that do not produce clinical signs and symptoms related to a regulatory peptide hormone.

However, patients with such tumors may have elevated serum concentrations of various hormones. About one-third of sporadic pancreatic endocrine tumors can be regarded as non-functioning. Typically, these tumors are recognized in patients in their 5th and 6th decades of life, although individuals with MEN-1 often develop these lesions, which are frequently multicentric, much earlier in life. More than 50% of the clinically non-functioning tumors are malignant (table III), as evidenced by local invasion of contiguous organs, lymph node metastases, or distant metastases.

Non-functioning tumors cause symptoms similar to those of exocrine pancreatic neoplasms, i.e. nausea, vomiting, and abdominal and/or back pain and display occasionally obstructive jaundice if located in the pancreatic head. Determination of circulating chromogranin is of value and the diagnosis may be verified by percutaneous cutting needle biopsy of the primary lesion or liver secondaries. Imaging of liver metastases follows the same guidelines as described above for functioning tumors. Liver resection is warranted only if cure is aimed at as cytoreduction is less appropriate for the truly non-functioning tumors without symptoms and signs to be palliated. Unresectable lesions may be treated with the same systemic therapy as described above for functioning tumors. An exception is though somatostatin which is mainly used for symptom relief and these patients by definition do not suffer hormone related symptoms.

Overall, non-functioning tumors seem to have poorer prognosis and lower response rate to systemic therapy than functioning tumors with comparable tumor burden.

In accordance with what is said above for the functioning tumors there are few, if any, controlled studies to support any management strategies also in patients with non-functioning neuroendocrine tumors.


Ihse I, Persson B, Tibblin S. Neuroendocrine metastases of the liver. World J Surg. (1995);19:76–82. [PubMed: 7740814]
Miller DL, Buefow PC (1998) Imaging procedures for the endocrine pancreas. In: Howard J, Idezuki Y, Ihse I, Prinz R (eds) Surgical diseases of the pancreas. Williams & Wilkins, Baltimore, pp 703–715 .
Que F G, Nagorney D M, Batts K P, Linz L J, Kvols L K. Hepatic resection for metastatic neuroendocrine carcinomas. Am J Surg. (1995);169:36–43. [PubMed: 7817996]
Chen H, Hardacre J M, Uzar A, Cameron J L, Choti M A. Isolated liver metastases from neuroendocrine tumors: does resection prolong survival? J Am Coll Surg. (1998);187:88–93. [PubMed: 9660030]
Frilling A, Rogiers X, Malago M, Liedke O M, Kaun M, Broelsch C E. Treatment of liver metastases in patients with neuroendocrine tumors. Langenbecks Arch Surg. (1998);383:62–70. [PubMed: 9627173]
Lang H, Oldhafer K J, Weimann A, Schlitt H J, Scheumann G F W, Flemming P, Ringe B, Pichlmayr R. Liver transplantation for metastatic neuroendocrine tumors. Ann Surg. (1997);225:347–354. [PMC free article: PMC1190741] [PubMed: 9114792]
Le Treut Y P, Delpero J R, Dousset B, Cherqui D, Segol P h, Mantion G, Hannoun L, Benhamou G, Launois B, Boillot P, Domergue J, Bismuth H. Results of liver transplantation in the treatment of metastatic neuroendocrine tumors. Ann Surg. (1997);225:355–364. [PMC free article: PMC1190742] [PubMed: 9114793]
Wiedermann B, Jensen R T, Mignon M, Modlin C J, Skogseid B, Doherty G, Oberg K. Consensus conference. Preoperative diagnosis and surgical management of neuroendocrine gastroenteropancreatic tumors: general recommendations by a consensus workgroup. World J Surg. (1998);22:309–318. [PubMed: 9494425]
Dousset B, Saint-Marc O, Pitre J, Soubrane O, Houssin D, Chapuis Y. Metastatic endocrine tumors: medical treatment, surgical resection, or liver transplantation. World J Surg. (1996);20:908–915. [PubMed: 8678970]
Eriksson BK, Larsson EG, Skogseid BM, Löfberg AM, Lörelius LE, Öberg KE (1998) Liver embolization of patients with malignant neuroendocrine gastrointestinal tumors. Cancer: 2293–2301 . [PubMed: 9840528]
Copyright © 2001, W. Zuckschwerdt Verlag GmbH.
Bookshelf ID: NBK6993


  • PubReader
  • Print View
  • Cite this Page

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...