Telephone: 46 31 342 1778; FAX: 46 31 822104
Orthotopic liver or multivisceral transplantation as treatment of metastatic neuroendocrine tumors
Article first published online: 22 FEB 2007
Copyright © 2007 American Association for the Study of Liver Diseases
Volume 13, Issue 3, pages 327–333, March 2007
How to Cite
Olausson, M., Friman, S., Herlenius, G., Cahlin, C., Nilsson, O., Jansson, S., Wängberg, B. and Ahlman, H. (2007), Orthotopic liver or multivisceral transplantation as treatment of metastatic neuroendocrine tumors. Liver Transpl, 13: 327–333. doi: 10.1002/lt.21056
- Issue published online: 22 FEB 2007
- Article first published online: 22 FEB 2007
- Manuscript Accepted: 30 OCT 2006
- Manuscript Received: 24 JUN 2006
- Swedish Research Council
- Swedish Cancer Society
Liver transplantation can be a therapeutic option for individual patients with neuroendocrine tumors metastatic only to the liver. In this consecutive series of 15 patients (5 multivisceral and 10 orthotopic liver transplantations) with well-differentiated carcinoids, or endocrine pancreatic tumors, we allowed higher proliferation rate (Ki67 <10%), large tumor burden, and higher age than previous studies. Liver transplantation offered good relief of symptoms, long disease-free intervals, and potential cure in individual patients. The survival of grafts and patients compared well with transplantation for benign disease. The overall 5-year survival was 90%. The recurrence-free survival of both multivisceral and liver transplantation related to the time after transplantation (about 20% at 5 years) despite inclusion of patients with higher risk. In conclusion, the critical prognosticators for long-term outcome still remain to be defined. The experience with multivisceral transplantation for patients with endocrine tumors of the pancreatic head is still limited. Liver Transpl 13:327–333, 2007. © 2007 AASLD.
Metastases of neuroendocrine tumors (NET), such as carcinoids and endocrine pancreatic tumors (EPT), can be limited to the liver for long periods. Thus a therapeutic window for liver transplantation in selected patients with tumors inaccessible for curative/cytoreductive surgery or not responding to medical therapy or interventions (local ablation/hepatic arterial embolization) is opened. Current data suggest that the primary tumor and regional metastases should be removed by a first operation, which allows assessment of tumor grade and stage. Patients with favorable biological features (well-differentiated endocrine carcinoma [WDEC], with low proliferation and stable disease) and no detectable extrahepatic metastases may even be cured by transplantation. However, critical analysis of hitherto published single-center series included 108 patients with only 5 actual 5-year disease-free survivors.1 The long-term prognosis is difficult to predict in patients not responding to conventional therapy prior to transplantation, as well as in patients with production of life-threatening hormones, such as insulin, catecholamines, or histamine, who may have been transplanted for palliative reasons.2 An upper age limit of 55 years has been suggested, and some centers will not accept a tumor mass exceeding 50% of liver volume,3 but no strict studies on tumor volume and outcome of orthotopic liver transplantation (OLT) are available. Poorly differentiated endocrine carcinomas (PDEC) seldom present at resectable stage, and for this type of highly proliferative tumor, transplantation is contraindicated.4
In 1994 Bechstein and Neuhaus5 reviewed the world literature of 30 patients with NET treated with OLT (15 carcinoids and 15 EPT). The 1-year survival was 52%, with transplantation-associated deaths in half of the patients within 8 months. Four years later, much more favorable survival figures were reported. The actual 5-year survival in a German series from Hannover was 80%; only 1 out of 12 patients with OLT had died of recurrent tumor.6 Similar figures were reported for patients with carcinoids in contrast to the poor survival of patients with EPT in a French multicenter study of 31 patients.7 On the other hand, in an English single-center series, carcinoid patients had high recurrence rates and shorter survival probably due to relatively long interval between diagnosis and OLT.8 In 1998, Lehnert9 reviewed the first 103 reported European cases with OLT as treatment for NET with estimated 5-year overall and disease-free survival of 47% and 24%, respectively. Survival was favorably influenced by young age, absence of hilar lymph node metastases, and less extensive upper-abdominal surgery. Certain tumor sites (lung and small bowel) and treatment with somatostatin analogues prior to surgery may be other positive prognostic factors. Recently a clinical management protocol was suggested.10 The selection criteria included complete resection of the primary tumor prior to OLT and tumors with low proliferation, but with lymph node metastases resectable at OLT, were accepted. Rectal carcinoid as a primary tumor or PDEC tumors, were among the exclusion criteria. With regard to multivisceral transplantation (MVTx), the prognosis for endocrine tumors widely surpassed that for other tumor types, such as sarcoma, hepatocellular cancer, or cholangiocarcinoma.11 MVTx as treatment for NET has been performed only a few times worldwide.
We report our experience of OLT or MVTx as treatment for NET in a consecutive series of 15 patients from 1997 to 2005. All procedures were performed with curative intent in patients with WDEC tumors with a Ki67 index <10%.
PATIENTS AND METHODS
Between January 1997 and September 2005, we performed OLT (n = 10; median age, 51.5 years [range, 39-64 years]) or MVTx (n = 5; median age, 43 years [range 38-57 years]) with curative intent in 15 patients with metastatic spread of NET to the liver (Table 1). All patients were investigated preoperatively by octreotide scintigraphy (including single photon emission computed tomography) and spiral computed tomography/magnetic resonance tomography (MRT) to validate that metastatic disease was confined to the liver. Core needle biopsies from liver tumors were examined by a single pathologist (O.N.) and verified as WDEC according to World Health Organization 2000 criteria. The average proliferation index (indicated by Ki67 immunostaining using the MIB 1 antibody) did not exceed 10% in any tumor specimen (Table 1); in 2 patients (nos. 1 and 6), “hot spots” of up to 15% MIB 1-positive tumor cells were observed.
|Type of Tx||Patient||Age at Tx (y)||Gender||Tumor Type||Ki 67 (%)||Tumor Volume* (>50%)||Main Hormone||CgA||Time After Tx||Tumor Status|
|Before Tx||After Tx||First Rec. (months)||Total Follow-up (months)|
|5||43||M||EPT||10 9 ± 1%||+||Gastrin||110||36||48||66 22.4 ± 11.8||AWD|
|15||52||M||EPT||10 3.7 ± 1.2||+||PP||50||31||88 53.8 ± 9.5||AWD|
Nine patients (4 EPT with body-tail location, 2 ileal [midgut] carcinoids, 1 bronchial [foregut] carcinoid, 1 rectal [hindgut] carcinoid, and 1 NET with unknown primary tumor [most likely a carcinoid]) underwent removal/exploration for the primary tumor and excision of regional lymph-node metastases 3 months to 10 years prior to OLT. One patient (no. 15) underwent OLT on suspicion of hepatocellular carcinoma, later proven to be an EPT. He underwent tail resection shortly after OLT. Four EPT patients had liver metastases at initial presentation, while 1 patient (no. 12) developed metastases 5 years after intentionally curative surgery. Two of the carcinoid patients (nos. 10 and 11) did not present with liver metastases but developed them 7-8 years after primary surgery (Fig. 1).
Three of the EPT patients had severe (insulinoma in patient 6) or moderate hormonal symptoms (glucagonoma in patient 7 and gastrinoma in patient 13) prior to transplantation; the glucagonoma and gastrinoma patients were treated with octreotide to reduce hormonal levels. Three carcinoid patients (patients 9, 11 and 14) had hormonal symptoms, but the patient with rectal tumor (no. 10) and the patient with unknown primary (no. 8) were symptom free. The symptomatic carcinoid patients were treated with octreotide, and 2 had additional treatment with interferon-alpha. Patient 11 had a severe foregut carcinoid syndrome with histamine hypersecretion and was treated with octreotide, prednisone and histamine 1 & 2-receptor blockade. Patient 14, who had a midgut carcinoid syndrome, had been treated over a 10-year period with repeat hepatic arterial embolizations, octreotide and interferon-alpha due to tumor progression limited to the liver (Fig. 1). Patient 8 (with unknown primary tumor) had been treated with debulking liver surgery and chemotherapy prior to transplantation, and patient 13 (with gastrinoma) had undergone left hemihepatectomy for tumor reduction. Out of 10 OLT patients, 7 had tumor burden exceeding 50% of the liver volume (Table 1 and Fig. 1).
In 5 patients (those with EPT with head location), the primary tumor, lymph nodes and liver metastases were removed in the same setting as MVTx. All patients were without hormonal symptoms but had large tumor burden (>50% of liver volume) at presentation. Two patients (nos. 1 and 5) were treated with a somatostatin analogue (octreotide) and 2 (nos. 2 and 5) were treated with interferon-alpha prior to transplantation. Multivisceral transplantation included removal of the stomach, liver, transplants of duodenum, and pancreas. In the first 2 patients, we saved the native small bowel. They received the liver, stomach, pancreas, duodenum, and upper third of the small bowel. Anastomoses were performed to the superior mesenteric artery and vein distally to the pancreas of the recipient. Outflow was secured with standard end-to-end cava-to-cava anastomosis. In the remaining 3 patients, we also removed the native small bowel and transplanted back the entire donor small bowel en bloc with the stomach, liver, pancreas, and duodenum using an aortic conduit anastomosed to the infrarenal aorta and a side-to-side cava anastomosis (piggyback).
Patient 13 was initially treated with cyclosporine A, steroids, and azathioprine. The initial immunosuppression for the other OLT patients was with tacrolimus and steroids with the addition of antilymphocyte globulin as induction (patients 8, 14, and 15). One patient (no. 1) with MVTx was treated with tacrolimus, steroids, and mycophenolate mofetil, and 1 patient (no. 3) was treated with a steroid-free protocol with tacrolimus and induction with high doses of antilymphocyte globulin. The other MVTx patients were initially treated with tacrolimus, steroids, and daclizumab induction (nos. 1 and 4), and rapamycin (no. 2), or tacrolimus, steroids, mycophenolate mofetil, and antilymphocyte globulin (no. 5). Episodes of acute rejection were treated with steroid bolus, and mycophenolate mofetil was added to the maintenance immunosuppression in most cases. Immunosuppression was not routinely modified to rapamycin after recurrence.
Prior to transplantation, all EPT patients were screened for peptide hormones in plasma by radioimmunoassay (gastrin, insulin, proinsulin, C-peptide, glucagon, vasoactive intestinal peptide, pancreatic polypeptide, human chorionic gonadotrophin subunits, calcitonin, and substance P), and carcinoid patients were screened for the main histamine and serotonin metabolites in urine (methylimidazole acetic acid and 5-hydroxyindole acetic acid, respectively) to find tumor-specific markers. All patients were screened for the general neuroendocrine tumor marker chromogranin A in plasma. Since many immunosuppressed patients have reduced kidney function and decreased elimination of low-molecular-weight proteins, peptide hormones in plasma may be elevated in the absence of recurrent tumor. Therefore, since 1999 we have added determination of chromogranin B in plasma to the screening, as this marker is regarded to be less sensitive to reduced glomerular filtration.12 Tumor-specific markers were followed every 3 months after transplantation.
At 6-month intervals, the patients underwent octreotide scintigraphy and spiral computed tomography. If there were discrepancies between these imaging procedures, MRT was performed. These examinations could be performed at shorter intervals if biochemical signaling of tumor recurrence was recorded.
Treatment at Tumor Recurrence
The purpose of the careful monitoring was to facilitate for surgical treatment, or other interventions, by early detection of tumor recurrence. If such therapy was not possible to perform, the patients were treated with somatostatin analogues/chemotherapy according to the Mayo program13 adjusted to hepatorenal function. Two patients (nos. 10 and 14), 1 with recurrent lymph node metastases and intact liver graft and the other with both lymph-node and liver metastases, received somatostatin-receptor-targeted radiotherapy with 177Lu-DOTA-Tyr3-octreotate (Table 2 and Fig. 2).
|October 2000 (A)*||December 2001 (B)†||November 2003||July 2005 (C)‡|
|CgA/plasma (<45 units/L)||230||73||36||170|
|CgB/plasma (<2 units/L)||2.6||1.1||1.1||n.d.|
|5-HIAA/urine (<5 mmol/mol creatinine)||13.1||6.7||6.2||21.6|
The probability of escaping tumor recurrence after transplantation (MVTx or OLT) was studied as a function of age at transplantation, gender, Ki 67 index, time from diagnosis to transplantation, tumor mass exceeding 50% of liver volume, or not, and time after transplantation, as well as Kaplan-Meier estimates of survival (overall and recurrence free).
Outcome in Patients Treated With OLT
The mean observation time for these patients was 53.8 ± 9.5 months (SEM). Of the 10 OLT patients, 8 are alive and active; the insulinoma patient (no. 6) died of recurrent tumor after 7 months, and the gastrinoma patient (no. 13) died of other causes 74 months after OLT. The latter patient had 5 operations after OLT: revision of hepatic duct stricture, left-sided adrenalectomy due to recurrent tumor, excision of lymph node metastases in the left renal hilum, and neck exploration on 2 occasions. One of the cervical explorations was negative for tumor with discrete lymphoproliferative disease in excised nodes despite uptake at octreotide scintigraphy. The tumor markers were negative on all occasions and preoperative diagnosis was based on scintigraphy and radiological observations. The patient died unexpectedly from pneumonia, and at autopsy no residual tumors were detected. Patient 15 with nonfunctional EPT had normal tumor markers 3 months after OLT but showed a slow increase of markers possibly related to reduced kidney function. After 31 months, he had increased levels of the tumor-specific markers pancreatic polypeptide and proinsulin and scintigraphically/radiologically-proven recurrent disease in the liver and retroperitoneal lymph nodes, which led to hepatic resection and metastasectomy. Five years after OLT he developed bilobar liver metastases, which have been treated for palliation by repeat embolization and radiofrequency ablation; he has been observed for 88 months. Patient 12 with nonfunctional EPT developed a small scintigraphic uptake close to the pancreatic head after 13 months; this lymph-node metastasis was removed, and the patient is tumor-free after a total observation of 70 months. Only 1 patient with EPT (no. 7, with glucagonoma) treated with OLT is without signs of tumor recurrence; present observation time is 12 months. The patient with unknown primary tumor (no. 8) treated with chemotherapy prior to transplantation developed recurrence in the graft after 9 months, which was recently resected; she has no evidence of disease after a total observation time of 17 months (Table 1 and Fig. 1).
Two of the patients with carcinoid tumors (nos. 9 and 11) have remained tumor free during an observation period of 57 and 70 months, respectively. Two carcinoid patients with long intervals between primary surgery and OLT developed recurrent disease; 1 (no. 14, with midgut carcinoid) had multiple lymph-node metastases but normal liver graft 10 months after OLT, and the other (no. 10, with hindgut carcinoid) developed liver metastases and a pelvic recurrence 43 months after OLT. These patients received 3-4 intravenous courses of 177Lu-DOTA-Tyr3-octreotate (7.0-7.5 GBq each) by Professor Erich Krenning and Dr. Dik Kwekkeboom in Rotterdam, Holland. The pelvic tumor was resected prior to radiotherapy to avoid a “sink effect” and facilitate for optimal delivery of radionuclide to the residual liver tumors. Both patients gradually normalized their tumor markers and showed marked objective tumor responses (Table 2 and Fig. 2). The first patient developed liver graft metastases more than 4 years later, and the second patient is in remission almost 2 years after targeted radiotherapy.
Outcome in Patients Treated With MVTx
The mean observation time for the MVTx patients was 22.4 ± 11.8 (SEM). Of the 5 patients with EPT, 2 (nos. 1 and 2) treated with MVTx died of transplantation-associated causes (lymphoproliferative disease and bleeding due to arteritis, respectively) within 4 months after surgery. Patient 4 died of recurrent tumor 27 months after MVTx, and patient 3 is tumor free after 12 months. Patient 5 underwent exploration of the mesenteric root 3 years after MVTx due to positive scintigraphic findings (normal tumor markers) but with negative tumor findings at surgery. The positive scintigraphy was possibly due to an infected aneurysm, which was resected. Almost 4 years after MVTx, this patient had tumor recurrence and is alive with poor renal function and thus cannot be treated with chemotherapy or somatostatin-receptor-mediated radiotherapy.
Detection of Tumor Recurrence
Tumor recurrence was first detected by octreotide scintigraphy, later verified by radiological investigations, in the vast majority of patients. In 2 patients (nos. 5 and 13), the surgical exploration was solely based on suspicion of tumor uptakes at scintigraphy with negative tumor findings (infected aneurysm and lymphoproliferative disease, respectively). In 1 patient (no. 15), marked biochemical signalling paralleled the appearance of scintigraphic uptakes.
Recurrence Rate and Survival
No significant influence on the risk of recurrence (MVTx or OLT) was found for age at transplantation, gender, Ki 67 index, time from diagnosis to transplantation, or tumor mass exceeding 50% of liver volume, or not. A hazard function for recurrence was estimated to exp (−0.9708 to 0.0375 × t), where t is years after transplantation (Fig. 3). The estimated median time to recurrence was 1.9 years. None of the 4 recurrence-free patients have died, while 5 out of 11 patients with recurrence died during the follow-up period.
The probability of escaping tumor-related death after OLT was 0.9 after 6 years: 1.0 for carcinoid tumors and 0.8 for EPT (Fig. 4). Of the 5 MVTx-treated patients, 2 died of transplantation-associated causes within the first 4 months, and 1 died of recurrent tumor after 27 months. Two patients are alive; 1 with recurrent tumor after 4 years has now been observed for 66 months, and the other is tumor free 1 year after MVTx.
This series differs from previous ones in several aspects. Patients up to 64 years old, with tumors of relatively high Ki67 index and with tumor burden exceeding 50% of liver volume (12 out of 15) were included; 12 had undergone primary surgery at our center (6 before OLT, 5 at MVTx, and 1 at 3 months after OLT). Patients with carcinoids, or tail-located EPT, were treated in 2 settings: histologically-proven radical removal of the primary and regional metastases followed by careful work-up to exclude extrahepatic metastases prior to OLT. Patients with head-located EPT require Whipple procedures, so we chose to treat them with MVTx. Age exceeding 50 years and extensive upper-abdominal surgery were significant risk factors in multivariate analysis of a large retrospective collective series of patients with NET treated by OLT.9 The MVTx patients had tumors with higher proliferation (Ki67 index), and all had large tumor burden of the liver and may thus represent a negative selection bias vs. the OLT patients.
Ki67 is a nuclear protein expressed in all proliferating cells, and the Ki67 index correlates with the clinical course of several tumor diseases. For EPT a cutoff value higher than 5% identified patients with shorter survival.14, 15 In a retrospective analysis of 19 OLT patients with NET and long follow-up, low proliferation rate (Ki 67 <5%) and regular expression of the adhesion molecule ecadherin (as a marker of metastatic potential) were associated with better survival. Analysis of the expression of the tumor suppressor p53 did not contribute to improved prognostic accuracy.16 Our results are comparable with the long-term survival of that series despite our inclusion of patients with Ki67 up to 10%. In limited series it is difficult to define risk factors for death, especially in case of high survival rate. Of the variables we analyzed as potential risk factors for recurrence (age, tumor burden, Ki67, and time from diagnosis), none had significant influence on the recurrence rate; the Ki67 index was closest to reaching statistical significance (P = 0.14). The hazard function for recurrence related to time from transplantation. In a recent series with OLT as treatment for gastrointestinal NET, the Ki67 index of 18 explanted livers varied between 0.06 and 2.35%. The follow-up in that series is still short, but there was no evident relation between survival/recurrence and low Ki67 index (<2%).10
The probability of escaping recurrence after MVTx or OLT was less than 0.2 at 5 years in our series. The results after OLT are promising, with an overall 90% 5-year survival, and thus confirm good symptom relief and long disease-free intervals in a majority of patients. The initial survival of grafts and patients do not differ from those obtained for benign disease.17 Each year, 25-30 patients with metastatic EPT or carcinoid are referred to our center from the western region of Sweden (1.6 million inhabitants). About 25% are subject to radical resection, including liver surgery; the others have their primary and regional metastases resected. The metastatic liver is then treated with palliative debulking or other interventions (hepatic arterial embolization or radiofrequency ablation) followed by interferon-alpha or chemotherapy upon progression.18 With this aggressive approach, we could report very high 5-year survival for patients with the midgut carcinoid syndrome early: 100% for hepatic resection and 63% for embolization.19 However, applying the same treatment principles for EPT, the 5-year survival was less than 20%, reflecting a biological difference between the 2 tumor types.20 This OLT series mainly included patients with EPT, with proliferation rates <5%. The carcinoid patients had lower proliferation rates (1-2%, with 1 exception), long duration of disease, and severe hormonal symptoms with 1 exception (patient 10). Some authors have excluded patients with rectal carcinoids,10 but in this series the patient with rectal carcinoid had an excellent clinical course; the patient was recurrence free during 43 months and is presently in good condition 65 months after OLT after receptor-targeted radiotherapy as completion treatment. Even with limited access to donor organs, we believe that transplantation can be an excellent treatment alternative for selected patients, such as those with metastatic EPT or carcinoids with severe hormonal symptoms. Our endocrine unit annually gets a referral of 25-30 patients with carcinoids/EPT and liver metastases for conventional surgical treatment. During the study period, 10 patients were offered the transplantation alternative and 9 agreed. From the previous 10-year period, 6 patients had disease limited to the liver and agreed to transplantation upon progression or nonresponsiveness to treatment.
Our estimated 1-year recurrence-free survival was almost 70% for MVTx and OLT vs. 77% for OLT reported by van Vilsteren et al.10 These figures compare well with 3 large series reporting 45-60% for OLT,7, 9, 16 but the short survival time reflects only the slow growth of these tumors. The high 5-year survival (90%) after OLT in this series is most likely due to a very strict protocol for follow-up with active treatment of recurrent tumor. That the majority of patients were operated on at our center with complete resections of their primary and regional metastases prior to or during transplantation possibly contributed to the favorable outcome. On the other hand, our recurrence-free survival after 5 years was similar to other reports. This fact can reflect that we have included several patients with potential risk factors, such as large tumor burden and higher proliferation, than previous series. Our data confirm that OLT is a valuable treatment for individual patients with unresectable liver metastases of NET. The experience with MVTx is still too limited for proper evaluation. The critical prognosticators to assess outcome of OLT for patients with NET are still to be defined.