|
Discussion |
Departments of Clinical and Surgical Sciences (Surgery), Pathology1 and Radiology2, Royal Infirmary, Lauriston Place, Edinburgh EH3 9YW, Scotland
J.R.Coll.Surg.Edinb., 47 April 2002, 495-499
Background: Liver metastases from neuroendocrine tumours may give rise to symptoms due to hormone production or mass effect. Accepted management options include administration of somatostatin-analogues, selective chemoembolisation or hepatic resection. The aim of this study was to review the management of hepatic neuroendocrine metastases in our unit. Methods: Patients with neuroendocrine tumours presenting between 1989 and 1999 were identified from pathology, radiology and surgical databases. Case notes were retrospectively reviewed for demographic data, treatment modality and outcome. Response to treatment was based on biochemistry, radiology or symptoms, and response rates were defined accordingly. Results: Thirty patients with a mean age of 55 years presented with, or later developed liver metastases. The most frequent presenting symptoms were abdominal pain (63%), diarrhoea (40%), weight loss (33%) and flushing (13%). Five patients underwent liver resection with complete symptomatic response, nine underwent chemoembolisation with a 75% response rate (either biochemically, radiologically or symptomatic) and fifteen were treated with a somatostatin-analogue, with a response rate of 86%. Median survival from detection of metastases was 45 months. Conclusions: Liver resection provides good symptomatic relief, but it is only indicated in a small proportion of patients with metastatic neuroendocrine tumours. Both chemoembolisation and somatostatin-analogues offer useful symptomatic control for these patients with good survival prospects.
This paper has been published in abstract form in British Journal of Surgery 2000; 87, S1: 55.
Keywords: carcinoid, neuroendocrine tumours, liver, surgery, hepatectomy, chemo-embolisation, octreotide
Neuroendocrine malignancies of the gastrointestinal tract are relatively rare.1 Typically, they are classified according to their site of origin as either foregut (including lung), mid-gut or hindgut tumours.2,3 This heterogeneous group of tumours can be further divided into those with endocrine activity and a group that is hormonally inactive.4 Hormone production can give rise to local or incapacitating systemic endocrinopathy and is dependent on tumour volume.5 These tumours are usually slow growing lesions with a low tendency to metastasise.
Liver metastases from neuroendocrine tumours are very uncommon, but may give rise to significant symptoms due to excess hormone production or mass effect.2 Therefore, treatment should aim to control tumour growth and excess hormone production.4 It is possible to provide symptom relief and possibly tumour growth inhibition by administration of specific blocking agents.2 Current accepted management options include administration of somatostatin analogues, chemotherapy, or selective chemoembolisation of liver lesions. In recent years, the concept of hepatic resection with either curative or palliative intent has attracted considerable attention.5
There is ongoing debate as to the optimal therapeutic strategy for hepatic neuroendocrine metastases and there is a paucity of data comparing various treatment modalities. In the present retrospective study, we have reviewed our experience of the management of hepatic metastases of neuroendocrine tumours in a specialist centre during the past 10 years.
The medical records of all patients with biopsy-proven hepatic metastases of neuroendocrine tumours, treated between January 1989 and December 1999, were reviewed. Patients were selected from computerised pathology (1990-1999) and surgical (1989-1999) databases and verified from files kept by one of the radiologists (DNR). Case notes were reviewed for demographic data, clinical presentation, treatment modality, laboratory data and outcome. Since our department serves as a tertiary referral centre for hepatobiliary malignancies, data on the source of referral were obtained. In addition, the pathology database was screened to obtain data on the overall incidence of gastrointestinal neuroendocrine tumours (i.e. with or without liver metastases) in the Edinburgh area.
Tumours were classified according to the site of the primary tumour, presence or absence and type of hormone production, as well as specific staining techniques on resected specimens. Argyrophilic (Grimelius) and argentaffin (diazo) staining reactions were applied routinely to all specimens. Further classification was achieved if necessary by immunohistochemistry including chromogranin A, neuron specific enolase and cytokeratin.6 Patients were considered to have a hormonally active tumour if clinical signs and symptoms were consistent with known endocrinopathy or if there was biochemical evidence of neuroendocrine tumour activity.
Figure 1: Survival following detection of hepatic metastases from gastrointestinal neuroendocrine tumours (Kaplan-Meier analysis, SPSS9.0) The continuous line shows data for neuroendocrine tumours with carcinoid differentiation (group 1, N=21), the dotted line shows data for the undifferentiated neuroendocrine tumours (group 2, N=8) (p=0.09, group 1 versus 2). Asterisks represent those patients that are still alive.
Patients fit for major surgery were considered for curative hepatic resection of their metastatic neuroendocrine tumours in the absence of extrahepatic dis-ease. Similarly, palliative liver resection was considered in patients with adequate functional reserve if non-surgical treat-ment failed to control symptoms adequately. Octreotide was preferred as primary treatment but chemoembolisation was employed in patients with multiple bilobar metastases not amenable to surgical treatment and with poor control of their symptoms by octreotide. When appropriate, patients were assessed for carcinoid-induced right-sided heart disease prior to hepatic resection, and valvular disease was then managed by non-surgical means. In patients undergoing resection, octreotide was routinely given peri-operatively to prevent carcinoid crisis.
Selective hepatic arterial chemoembolisation was performed to control symptoms if patients were considered to be unsuitable for surgery or if surgical or other medical treatment had failed. 7 The procedure involved preliminary angiography to identify the hepatic arterial anatomy and assess the patency of the portal vein. Following selective catheterisation a combination of doxorubicin (40 mg) and iodised oil (10-20 ml) (Lipiodol; Laboratoire Guerbet, France) was delivered into the tumour circulation.
In large hypervascular lesions a particulate agent, Contour (Merck Pharmaceuticals, UK) or Spongostan fragments (Johnson & Johnson, UK) were used to occlude the flow to the lesion. Chemoembolisation was repeated at 3-monthly intervals for a maximum of three sessions, depending on the response, as judged by clinical or biochemical improvement and radiologically by measurement on computerised tomography (CT) scans.8
Response to treatment was defined primarily on control of symptoms, where these were present pre-operatively, and based on biochemistry results (mainly decreased urinary 5-hydroxyindole acetic acid (HIAA) levels) or on radiology (decreased tumour size). Follow-up data on survival were collected from patient records and from information provided by their general practitioners.
Statistics
Statistical analysis was performed using the Statistical Package for the Social Sciences, version 9.0 for Windows 98 (SPSS, Chicago, Ill., USA). Kaplan-Meier analysis was applied to calculate actuarial survival curves and the log-rank test was used to test for differences in survival. Data a represented as means (±SD) unless otherwise indicated. P<0.05 was taken to indicate statistical significance.
Patients
Between January 1989 and December 1999, thirty patients underwent treatment in our unit for hepatic neuroendocrine metastases. Of these, 9 patients presented with liver secondaries from neuroendocrine tumours within the population of Edinburgh and 21 patients (70%) were referred from elsewhere. In the study period, a total of 112 cases of primary gastrointestinal neuroendocrine tumours were identified (including the 30 with hepatic metastases).
The mean age of the 30 patients with liver metastases at initial presentation was 55 (±14; range 23-81) years. There were 17 males and 13 females. The tumours were classified as carcinoid (n=21), undifferentiated neuroendocrine carcinoma (n=8) and gastrinoma (n=1). 3 Seventeen patients (57%) had liver metastases at presentation, whilst the remaining 13 patients (43%) developed metachronous lesions. The mean interval between detection of the primary tumour and the hepatic deposits was 31 months. Liver metastases were multiple in the majority of patients (93%). The site of the primary tumour was ileocaecal in 33%, pancreas in 27%, stomach in 10%, colon in 10% and unknown in 20% of cases. The most frequent presenting symptoms were abdominal pain (63%), diarrhoea (40%) and weight loss (33%). Flushing was a presenting symptom in 13% of patients and an additional 10% of patients later developed flushing.
There were no patients with multiple endocrine neoplasias in this series. Due to the retrospective nature of the study, quantitative or qualitative results on serum hormones or urinary excretion of hormone metabolites following referral were available in only a limited number of patients. For example, urinary 5-HIAA results were available in 60% of patients with carcinoid tumours and levels were elevated in 72% of these. Histological staining was positive for argyrophilia in 8 of 11 cases (73%), argentaffin staining was positive in 5 of 8 cases (63%) and chromogranin A staining was positive in 11 of 14 cases (79%). Neuron specific enolase staining was positive in both cases where it was tested and cytokeratin staining was positive in 4 of 6 cases (67%). There was no difference in staining between hormonally active and inactive tumours.
Figure 2a Figure 2bFigure 2a and b: CT-scan showing multiple hypervascular metastases in both lobes of the liver in a patient who had previously undergone resection of an ileal carcinoid tumour
Management
Five patients (17%) underwent major liver resection for symptomatic carcinoid tumours. Three patients underwent extended right hepatectomy (segments IV-VIII), one patient underwent central liver resection (segments IV, V and VIII) and one patient underwent segmentectomy IV and multiple local resections. Four hepatectomies were performed with palliative and one with curative intent. There was no intraor postoperative mortality and no episode of carcinoid crisis, although intra-operative blood loss was significant in this patient group, in which some degree of right-sided heart disease was usually present. One patient underwent a relaparotomy for post-operative bleeding from the right hepatic artery stump and bile leakage following an extended right hepatectomy. The patient subsequently made an uneventful postoperative recovery and survived for a further four years. All patients had a complete response in terms of symptom control and only one patient required later treatment with octreotide for symptomatic recurrence.
Nine patients (30%) were treated by chemoembolisation with a 75% response rate (mean of 1.7 embolisations per patient). Eight of these patients were classified as having symptomatic carcinoid tumours and one as having an undifferentiated, hormonally inactive tumour. Of the patients that underwent chemoembolisation, 6 were treated with octreotide either before or after chemoembolisation. Overall, fifteen patients (50%) were treated with the somatostatin-analogue octreotide at some stage of their disease, with a response rate of 86%.
Outcome
Eight patients are still alive and well; 95% of deaths were tumour-related. For the whole group of thirty patients, the median survival after initial presentation was 60 ±11 months (95% confidence interval (CI): 39-81 months). There was no statistically significant difference between the subgroup with carcinoid tumours (n=21) and the subgroup with undifferentiated neuroendocrine tumours (n=8) (Fig. 1). Median survival after detection of liver metastases was 45 months (95% CI: 27-63 months). For this analysis, ±9 the single patient with metastatic gastrinoma was excluded. Median survival following detection of liver metastases in the carcinoid group was 52 ±10 months (95% CI: 33-71 months), whereas in the undifferentiated group it was 20 ±5 months (11-29 months), (p=0.09). There was no statistically significant difference in median survival between patients treated surgically (59±10 months), by embolisation (24±13 months; p=0.11 versus surgery) or by octreotide (37±16 months; p=0.33 versus surgery), or between octreotide treatment and embolisation.
In this study of 30 patients with hepatic metastases of neuroendocrine gastrointestinal tumours, the median survival following medical, radiological or surgical treatment is similar to a large American series5. Median survival for non-surgical treatment of hepatic metastases from colorectal cancers has been reported to be approximately 10 months, and the prognosis of liver secondaries of non-colorectal, non-neuroendocrine origin is even worse. 9-11 The median survival of 45 months in the present series is in keeping with the belief that the prognosis of liver metastases of gastrointestinal neuroendocrine origin, even without surgical treatment, is generally favourable.
Although this is a small series of patients and there may be considerable selection bias, it should be noted that 9 patients presented with neuroendocrine liver metastases from the local referral area in the study period. During these 10 years, a total of 91 primary or secondary neuroendocrine gastrointestinal malignancies were registered in the pathology database of our institution, suggesting that at least ten percent of patients with such primary malignancies will develop hepatic metastases over time. This is in sharp contrast with the metastatic potential of other gastrointestinal tumours such as colorectal carcinoma.
Modern percutaneous or intra-operative ultrasonography and contrast-enhanced spiral CT-scanning (figure 2A&B) are usually sufficiently sensitive and specific to confirm the diagnosis of hepatic metastases of neuroendocrine primaries. Mesenteric angiography (figure 3) will generally only be required as a preliminary to embolisation. Magnetic resonance imaging is a promising diagnostic modality in the workup of neuroendocrine hepatic metastases but its exact role, compared with CT-scanning has yet to be defined. Recently, indium-labelled octreotide, meta-iodobenzylguanidine (MIBG) scanning and positron emission tomography scanning have all been suggested to be very promising in detecting primary tumours and the presence of extrahepatic disease.3, 4, 12, 13
Figure 3: Mesenteric angiogram with selective imaging of the hepatic arterial blood supply. In the right lobe of the liver multiple hypervascular areas compatible with metastases of a carcinoid tumour can be seen
Somatostatin receptor scintigraphic scanning (octreoscan) can be used to direct therapy of primary and metastatic neuroendocrine tumours, because 80% of neuroendocrine tumours have somatostatin receptors.4, 14-16 The presence of a receptor for somatostatin may provide an indication for the response to octreotide treat-ment.14 Plasma or urine levels of hormones or their metabolites (e.g. 5HIAA) and plasma chromogranin A levels may be used to confirm the diagnosis and monitor the effect of treatment.3, 4
Currently, most patients with metastatic neuroendocrine tumours are considered for treatment with the somatostatin analogue, octreotide. Octreotide is effective in reducing symptoms and may arrest tumour growth in some patients.4 It has been suggested that high dose somatostatin analogue treatment may be effective in symptom control in patients failing to respond to standard treatment, possibly by inducing apoptosis.4 Recently, the long-acting somatostatin analogue, octreotide LAR (Lanreotide, Sandostatin LAR), a micro-encapsulated long-acting formulation suitable for once-monthly intramuscular injection, has been shown to be as effective as octreotide in the control of symptoms in patients with malignant carcinoid syndrome.17
The potential role of radiolabeled MIBG and octreotide in the treat-ment of primary or metastatic neuroendocrine tumours has yet to be confirmed.3, 18, 19 Good symptom control has been achieved by some workers with repeated administration of interferon, but this treatment has more side effects than octreotide and is not yet considered routine treatment.4, 12
Hepatic artery (chemo-) embolisation may be useful to provide symptom relief 1, 7 as these tumours are hypervascular and almost exclusively supplied by the arterial system. Our experience confirms that it can be performed with minimal morbidity and no mortality.7,20, 21 Systemic chemotherapy for neuroendocrine metastases has considerable toxic side effects and its effectiveness is still controversial.1, 3-5
In the present study, patients that were not amenable to surgical treatment were either treated by chemoembolisation or octreotide. During the early study period, more patients were treated initially by chemo-embolisation given that this treatment had been shown to be more effective than other non-surgical therapies.20 As octreotide gained wider acceptance as an effective alternative,22 patients who were not candidates for surgical treatment were treated more often by either subcutaneous or intramuscular octreotide.
Whilst hepatic resections may have been associated with high morbidity and mortality rates in the past, recent reports have provided evidence that it is possible to perform resectional liver surgery safely in the United Kingdom.23, 24 Although the present study contains only a small group of selected patients, the results would seem to confirm that, in specialised units, liver surgery for neuroendocrine tumours can be performed safely with either curative or palliative intent.5 Careful consideration should be given to the patient with carcinoid disease affecting the heart since the raised central venous pressure can cause difficulties with blood loss during liver dissection. A total vascular exclusion technique is poorly tolerated by such patients.25 The current retrospective review demonstrates that hepatic resection may provide good relief of symptoms and our limited experience would suggest that it may result in improved long time survival.
The role of liver transplantation in the treatment of metastatic neuroendocrine tumours is not clear.5 A recent review suggests that results are discouraging in elderly patients, particularly when extended operations are required to deal with extrahepatic disease. On the other hand, there does seem to be a place for liver replacement in young patients with intractable pain or hormone-related symptoms when disease is confined to the liver.26
In recent years, promising data have been published on the role of open, laparoscopic and percutaneous cryotherapy and in situ radiofrequency ablation in the early treatment of hepatic neuroendocrine metastases.27, 28 It does seem more appropriate to consider hepatic resection when feasible and it is yet to be determined whether long-term control of symptoms is achievable in patients with multiple lesions when cyrotherapy is employed.29 It is possible that improved survival and symptom control will be achieved in the future by adopting a multimodal approach, applying different treatment modalities synchronously or metachronously.4, 22, 29, 30
Liver metastases from neuroendocrine tumours are rare, but may give rise to significant symptoms due to hormone production or mass effect. Liver resection provides good symptomatic relief, but it is only indicated in a small proportion of patients with metastatic neuroendocrine tumours. Both chemoembolisation and somatostatin-analogues offer useful symptomatic control for these patients with good survival prospects.
C.H.C. Dejong was a recipient of a grant from the Niels Stensen Stichting, Amsterdam, the Netherlands.
1. Moertel CG, Johnson CM, McKusick MA, et al. The
management of patients with advanced carcinoid tumors
and islet cell carcinomas. Ann Intern Med 1994; 120(4):
302-9
2. Ihse I, Persson B, Tibblin S. Neuroendocrine metastases of
the liver. World J Surg 1995; 19:76-82
3. Kulke MH, Mayer RJ. Carcinoid tumors [see comments]. N Engl J Med 1999; 340(11): 858-68
4. Eriksson B, Oberg K. Neuroendocrine tumours of the pancreas. Br J Surg 2000; 87(2): 129-31
5. Que FG, Nagorney DM, Batts KP, et al. Hepatic
resection for metastatic neuroendocrine carcinomas. Am J
Surg 1995; 169(1): 36-42; discussion 42-3
6. Schurmann G, Betzler M, Buhr HJ. Chromogranin A,
neuron-specific enolase and synaptophysin as neuroendocrine cell markers in the diagnosis of tumours of the
gastro- entero-pancreatic system. Eur J Surg Oncol 1990;
16(4): 298-303
7. Therasse E, Breittmayer F, Roche A, et al. Transcatheter
chemoembolization of progressive carcinoid liver metastasis. Radiology 1993; 189(2): 541-7
8. Wigmore SJ, Redhead DN, Yan XJ, et al. Virtual hepatic
resection using three-dimensional reconstruction of helical
computed tomography angioportograms. Ann Surg 2001;
233(2): 221-226
9. Bland KI, Karakousis CP, Copeland EM. Atlas of surgical
oncology. 1st ed. Philadelphia: W.B. Saunders Company
Ltd, 1995
10. Geoghegan JG, Scheele J. Treatment of colorectal liver
metastases. Br J Surg 1999; 86:158-169
11. Schwartz SI. Hepatic resection for noncolorectal
nonneuroendocrine metastases. World J. Surg. 1995;
19:72-75
12. Taal BG, Hoefnagel C, Boot H, et al. [Carcinoid tumour of
the intestines: developments in the Netherlands for diagnosis and palliative treatment]Ned. Tijdschr. Geneeskd.
1999; 143:445-451
13. van Heerden JA, Thompson GB. Chapter 48. The
surgical approach to islet cell tumors of the pancreas. In
Trede M, Carter DC, eds. Surgery of the pancreas. New
York: Churchill Livingstone, 1997. pp. 565-588
14. van Eijck CH, de Jong M, BreemanWA, et al. Somatostatin
receptor imaging and therapy of pancreatic endocrine
tumors. Ann Oncol 1999; 10(Suppl 4): 177-81
15. Slooter GD, Mearadji A, Breeman WA, et al. Somatostatin
receptor imaging, therapy and new strategies in patients with
neuroendocrine tumours. Br J Surg 2001; 88(1): 31-40
16. Murray AD, Brooks ME. Nuclear medicine in oncology.J R Coll Surg Edinb 2000; 45(2): 110-9
17. Rubin J, Ajani J, Schirmer W, et al. Octreotide acetate
long-acting formulation versus open-label subcutaneous
octreotide acetate in malignant carcinoid syndrome. J Clin
Oncol 1999; 17(2): 600-6
18. Taal BG, Hoefnagel CA,Valdes Olmos RA, et al. Palliative
effect of metaiodobenzylguanidine in metastatic carcinoid
tumors. J Clin Oncol 1996; 14(6): 1829-38
19. Kema IP, Willemse PH, De Vries EG. Carcinoid tumors [letter; comment]. N Engl J Med 1999; 341(6): 453-4; discussion 454-5
20. Perry LJ, Stuart K, Stokes KR, Clouse ME. Hepatic arterial
chemoembolizationformetastaticneuroendocrinetumors.
Surgery 1994; 116(6): 1111-6; discussion 1116-7
21. Proye C. Natural history of liver metastasis of
gastroenteropancreatic neuroendocrine tumors: place for
chemoembolization. World J Surg 2001; 25(6): 685-8
22. Oberg K. Advances in chemotherapy and biotherapy of
endocrine tumors. Curr Opin Oncol 1998; 10(1): 58-65
23. Finch MD, Crosbie JL, Currie E, Garden OJ. An 8-year
experience of hepatic resection: indications and outcome.
Br J Surg 1998; 85:315-319
24. Rees M, Plant G, Wells J, Bygrave S. One hundred and
fifty hepatic resections: evolution of technique towards
bloodless surgery. Br J Surg 1996; 83(11):1526-9
25. Bismuth H, Castaing D, Garden OJ. Major hepatic resection under total vascular exclusion. Ann Surg 1989; 210(1):
13-9
26. Lehnert T. Liver transplantation for metastatic
neuroendocrine carcinoma: an analysis of 103 patients.
Transplantation 1998; 66(10): 1307-12
27. Siperstein AE, Berber E. Cryoablation, percutaneous alcohol injection, and radio frequency ablation for treatment of
neuroendocrine liver metastases. World J Surg 2001; 25(6):
693-6
28. Seifert JK, Junginger T, Morris DL. A collective review of
the world literature on hepatic cryotherapy. J R Coll Surg
Edinb 1998; 43(3): 141-54
29. Usatoff V, Habib N. Hepatic malignancy: challenges and
opportunities for the surgeon. J R Coll Surg Edinb 2000;
45(2):99-109
30. Cha C, Lee FT, Jr., Rikkers LF, et al. Rationale for the combination of cryoablation with surgical resection of hepatic
tumours. J Gastrointest Surg 2001; 5(2): 206-13
Copyright: 6 March 2002
Correspondence: O.J. Garden, Regius Professor of Surgery, Department of Clinical and Surgical Sciences (Surgery), Royal Infirmary, Lauriston Place, Edinburgh EH3 9YW, Scotland. E-mail OJGarden@ed.ac.uk