Presence and extent of liver metastases (LM) are major prognostic

Transcription

1 ORIGINAL ARTICLE Liver Transplantation for Neuroendocrine Tumors in Europe Results and Trends in Patient Selection A 213-Case European Liver Transplant Registry Study Y.P. Le Treut, MD, E. Grégoire, MD, J. Klempnauer, MD, J. Belghiti, MD, E. Jouve, MSc, J.Lerut,MD,PhD,FACS, D. Castaing, MD, O. Soubrane, MD, O. Boillot, MD, PhD, G. Mantion, MD, PhD, K. Homayounfar, MD, M. Bustamante, MD, D. Azoulay, MD, PhD, Ph. Wolf, MD, PhD, M. Krawczyk, MD, A. Pascher, MD, PhD, MBA, FEBS, B. Suc, MD, L. Chiche, MD, PhD, J. Ortiz de Urbina, MD, FACS, V. Mejzlik, MD, M. Pascual, MD, JPA Lodge, MD, FRCS, S. Gruttadauria, MD, PhD, FACS, F. Paye, MD, PhD, F.R. Pruvot, MD, PhD, S. Thorban, MD, FACS, A. Foss, MD, PhD, and R. Adam, MD, PhD for ELITA Objective: The purpose of this study was to assess outcomes and indications in a large cohort of patients who underwent liver transplantation (LT) for liver metastases (LM) from neuroendocrine tumors (NET) over a 27-year period. Background: LT for NET remains controversial due to the absence of clear selection criteria and the scarcity and heterogeneity of reported cases. Methods: This retrospective multicentric study included 213 patients who underwent LT for NET performed in 35 centers in 11 European countries between 1982 and One hundred seven patients underwent transplantation before 2000 and 106 after Mean age at the time of LT was 46 years. Half of the patients presented hormone secretion and 55% had hepatomegaly. Before LT, 83% of patients had undergone surgical treatment of the primary tumor and/or LM and 76% had received chemotherapy. The median interval between diagnosis of LM and LT was 25 months (range, months). In addition to LT, 24 patients underwent major resection procedures and 30 patients underwent minor resection procedures. Results: Three-month postoperative mortality was 10%. At 5 years after LT, overall survival (OS) was 52% and disease-free survival was 30%. At 5 years from diagnosis of LM, OS was 73%. Multivariate analysis identified 3 pre- From the Hôpital La Conception, Marseille, France; Medizinische Hochschule, Hannover, Germany; Hôpital Beaujon, Clichy, France; Cliniques Universitaires St Luc, Brussels, Belgium; Hôpital Paul Brousse, Villejuif, France; Hôpital Cochin, Paris, France; Hôpital Edouard Herriot, Lyon, France; Hôpital Jean Minjoz, Besançon, France; Georg August Universität, Göttingen, Germany; Complexo Hospitalario Universitario, Santiago de Compostela, Spain; Hôpital Henri Mondor, Créteil, France; Hôpital de Hautepierre, Strasbourg, France; Medical University, Warsaw, Poland; Charité Campus Virchow-Klinikum, Berlin, Germany; Hôpital Rangueil, Toulouse, France; Hôpital Côte de Nacre, Caen, France; Hospital de Cruces, Bilbao, Spain; Center of transplantation, Brno, Czech Republik; CH Universitaire Vaudois, Lausanne, Switzerland; St James University Hospital, Leeds, United Kingdom; ISMETT, Palermo, Italy; Hôpital St Antoine, Paris, France; Hôpital Claude Huriez, Lille, France; Technische Universität, München, Germany; and Rikshospitalet, Oslo, Norway. Part of this work was presented at the International Liver Congress, Barcelona, April 19 22, 2012, and at the 10th World Congress of the International Hepato- Pancreato-Biliary Association, Paris, July 1 5, Disclosure: The authors declare no conflicts of interest. Reprints: Pr. YP Le Treut, MD, Aix-Marseille Université, Service de chirurgie générale et transplantation hépatique, Hôpital de La Conception, 147 Bd Baille, Marseille, France. yves-patrice.letreut@ap-hm.fr. Copyright C 2013 by Lippincott Williams & Wilkins ISSN: /13/ DOI: /SLA.0b013e31828ee17c dictors of poor outcome, that is, major resection in addition to LT, poor tumor differentiation, and hepatomegaly. Since 2000, 5-year OS has increased to 59% in relation with fewer patients presenting poor prognostic factors. Multivariate analysis of the 106 cases treated since 2000 identified the following predictors of poor outcome: hepatomegaly, age more than 45 years, and any amount of resection concurrent with LT. Conclusions: LT is an effective treatment of unresectable LM from NET. Patient selection based on the aforementioned predictors can achieve a 5-year OS between 60% and 80%. However, use of overly restrictive criteria may deny LT to some patients who could benefit. Optimal timing for LT in patients with stable versus progressive disease remains unclear. Keywords: liver transplantation, multivariate analysis, neuroendocrine tumors, prognostic study (Ann Surg 2013;00: 1 9) Presence and extent of liver metastases (LM) are major prognostic factors in patients with neuroendocrine tumors (NET). 1 3 Because of slow growth and relatively low-grade malignancy, unresectable metastatic NET is considered as the only acceptable indication for liver transplantation (LT) in patients with metastatic cancer. However, due to scarcity and heterogeneity of reported cases of LT for NET, controversy continues surrounding patient selection and timing for LT. Further complicating the debate is the fact that a wide variety of treatment alternatives have been proposed including surgical, medical, radiological, and nuclear medical modalities. The true benefit of LT remains unknown, and comparison of outcomes of LT with those of other treatments is needed, especially for metastatic welland moderately differentiated NET, that is, NET G1/G2 of the new WHO classification. 4 Various studies allowing analysis of selection criteria for LT in patients with NET have been published. The one and only prospective study reported excellent outcomes but at the price of highly restrictive selection criteria. 5,6 Many single-center retrospective studies are available but their small sample size- rarely more than 10 cases 7 14 precludes meaningful conclusions. A few retrospective multicentric studies have been published but selection criteria have differed greatly between centers. Registry-based analyses lack the necessary specificity for analysis of problems related to NET. 18,19 The purpose of this report was to describe a large retrospective cohort of LT for NET. It was specifically designed to analyze immediate and long-term outcome, to determine the main prognostic factors, and to study the evolution of results over time. Annals of Surgery Volume 00, Number 00,

2 Le Treut et al Annals of Surgery Volume 00, Number 00, 2013 METHODS Patient Identification Patients were identified from the European Liver Transplant Registry (ELTR). All patients who had a diagnosis of Secondary Carcinoma classified E9 (carcinoid) and E10 (other neuroendocrine) and underwent LT between November 1, 1982 and December 31, 2009 were included. This study was approved by and performed under the auspices of the Board of European Liver and Intestine Transplant Association, the governing society of ELTR. Data Collection After patient identification, the reporting transplant centers were contacted to obtain detailed case information. After verification of diagnosis to rule out possible errors (eg, sarcoma, colorectal metastases), a case record form containing 55 items was used to record clinical variables, including patient demographics, pre-lt diagnosis/management of primary tumor (PT) and LM, indication for LT, details about the LT procedure, pathology data, postoperative course, recurrence (date, site, and treatment), and survival. Copies of operative and pathologic reports were obtained for each patient. Tumor were classified according to the new WHO classification for endocrine tumors. 4 Case record forms were processed at the coordinating center, Hôpital de La Conception, Aix Marseille Université, Marseille. Processing included reviewing to ensure internal consistency and entry into a database (Fig. 1). The indication for LT was decided at each center. For study purposes, indications were reclassified into 3 main groups, that is, hormonal syndrome including patients presenting life-threatening or debilitating hormone-related symptoms, tumor bulk including patients presenting pain or debility associated with enlargement of the liver, and oncological including patients with low-grade symptoms. Extent of liver involvement was evaluated at each center on the basis of computed tomographic scan slices obtained before LT and expressed as approximate percentage of liver replacement. Hepatomegaly was retrospectively defined as enlargement of the explanted liver by 25% or more in relation to the patient s predicted liver volume determined using the formula proposed by Vauthey et al 20 : standard liver volume = body surface area (m 2 ) In 8 cases, calculation was impossible because the required data were missing. However, 5 of these 8 patients had clinical evidence of hepatomegaly. Reporting of markers such as Ki-67 status or chromogranin A level was too rare to allow analysis. Also due to insufficient reporting, postoperative events with the exception of early retransplantation were not taken into account. Statistical Analysis Categorical variables expressed as numbers and percentages were compared using the χ 2 test or Fischer exact test. Continuous variables expressed as mean ± SD or median (range) were compared using the Student t test or Mann-Whitney U test as appropriate. Percentages were rounded to whole numbers. The last follow-up examination was performed on December 31, Survival rates, that is, overall survival (OS) and disease-free survival (DFS) were determined using the Kaplan-Meier method, and groups were compared using the log-rank test. Prognostic analysis was based solely on pre-lt variables because they are the only ones useful for patient selection. Variables with a P value of less than 0.10 in the univariate analysis were included in multivariate Cox regression analysis performed with backward elimination. Hazard ratios (HRs) and 95% confidence intervals (CI) were calculated. P values of less than 0.05 were considered significant in all analyses. Statistical analysis was carried out using SAS version 9.2 (SAS Institute Inc, Cary, NC). FIGURE 1. Selection of patients from ELTR. PATIENTS Population and Disease Between November 1982 and December 2009, a total of 213 patients underwent of LT for NET in 35 centers in 11 European countries. Six centers, that is, Hannover, Clichy, Marseille, Brussels, Villejuif, and Paris Cochin reported 10 patients or more (maximum, 29). Ten centers reported 1 or 2 patients each. These 213 cases accounted for almost half of the cases of LT for NET performed in Europe over the study period (Fig. 1). One hundred seven patients underwent transplantation before 2000 and 106 after There were 99 females and 114 males. Mean age at the time of LT was 46 ± 11 years [median: 48 years (16 71)]. PT site locations are given in Table 1. The PT was detected at the time of or after LT in 12 cases and remained unknown despite intensive pre-lt explorations in 17 cases. Four patients did not undergo further investigation before LT because of presumption of hepatocellular carcinoma. The PT site detected before LT was in duodenum/ pancreas in 93 cases and bronchial/digestive tract in 91 cases. The PT 2 C 2013 Lippincott Williams & Wilkins

3 Annals of Surgery Volume 00, Number 00, 2013 Liver Transplantation for Neuroendocrine Tumors TABLE 1. Distribution of Cases According to Primary Tumor Location and Secretion Status Primary Tumor Site Functional Tumors Nonfunctional Tumors Total Bronchial tree Stomach Duodenum Pancreas Jejunum Ileum Right colon Sigmoid colon Rectum Common bile duct Undetected Total Ten carcinoids. One carcinoid. Two gastrinomas. Eighteen gastrinomas, 6 insulinomas, 6 VIPoma, 4 polymorphs, 3 glucagonomas, and 3 carcinoids. Two carcinoids, 1 GFRHoma, 1 VIPoma, 1 gastrinoma. Thirty-eight carcinoids One carcinoid. Three carcinoids, 1 gastrinoma. Including 4 cases detected after LT. Including 5 cases detected during (2) or after (3) LT. Including 2 cases detected during (1) or after (1) LT. Including 1 case detected after LT. site was undetected before LT in the remaining 29 cases ( ). Almost 50% of patients presented hormonal secretion.. LM was classified as synchronous in 119 cases and metachronous in 49 cases [median time interval, 40 months (3 211)]. In the last 44 cases, LM was diagnosed before detection of the PT. The median time interval between diagnosis of LM and LT was 25 months (1 149 months). The extent of liver involvement based on data available in 200 cases was less than 20% in 31 cases, 20% to 40% in 41, 40% to 60% in 41, 60% to 80% in 54, and more than 80% in 31 (median, 50%). Pretransplant History Before LT, a total of 178 patients (83%) underwent surgery for treatment of either the PT (n = 158) or LM (n = 58) (Table 2). The most frequent procedures were small bowel resection in 53 cases, distal pancreatectomy with or without splenectomy in 47, and Whipple procedure in 21. Major hepatic resection was performed in 23 cases. In addition, 2 patients required tricuspid valve replacement for carcinoid heart disease. A total of 161 patients (76%) received chemotherapy before LT, including somatostatin analogs in 63 patients and transarterial chemoembolization in 76 cases (35%, 1 14 procedures). Overall, 12 patients (6%) underwent no treatment before LT. Based on information available in 192 patients, somatostatin receptor scintigraphy (octreotide scanning) was performed in 131 patients. Indication for LT The main indications for LT and the time intervals between diagnosis of LM and LT are listed in Table 3. The indication was oncological treatment in 54%, tumor bulk in 24%, and hormonal syndrome in 17%. The LT procedure was performed for presumptive hepatocellular carcinoma in 4 cases and for severe complications after transarterial chemoembolization in 9 cases. TABLE 2. Surgical Procedures Before Transplantation Surgical Procedures for the Primary Tumor 158 Patients Pulmonary lobectomy 16 Partial or total gastrectomy 8 Whipple procedure 21 Total pancreatectomy 3 Distal pancreatectomy (± splenectomy) 47 Other pancreatectomy 3 Transduodenal tumorectomy 1 Bile duct excision 1 Small bowel resection (± right colectomy) 53 Colectomy 6 Proctectomy 3 Transanal tumorectomy 1 Surgical Procedures for the Metastases 56 Patients Major hepatic resection 23 Minor hepatic resection 35 Hepatic resection of unknown extent 5 Hepatic artery ligation 3 Sixteen staging laparotomies excluded. Including 17 repeat hepatic resections. Including 1 case of total hepatectomy secondary to hepatic artery and portal vein injury during Whipple procedure, followed by emergency rescue LT. TABLE 3. Distribution of Cases According to Indication for Transplantation Main Reason for LT N Time Interval to LT Median (Range) Oncological mo (1 149) Tumor bulk mo (1 122) Hormonal syndrome mo (4 110) Presumption of hepatocellular carcinoma 4 1, 5, 7, and 10 mo Hepatic failure after TACE 4 24, 39, 53, and 54 mo Ischemic cholangitis after TACE 5 12, 49, 60, 62, and 74 mo Other 1 3mo Total mo (1 149) Time interval from diagnosis of LM to LT. Rescue LT after resection of PT using Whipple s procedure complicated by vascular injuries. TACE indicates transarterial chemoembolization. Transplantation Procedures In 193 cases (90%), LT was performed using a deceased donor graft including right split liver graft in 6 cases and multivisceral graft in 6 cases. A living donor graft was used in 13 cases including right liver in 11 cases and left liver in 2 cases. A domino procedure was performed in 7 cases. Mean operative time was 7.8 ± 3.1 hours. Mean cold ischemic time was 8.7 ± 3.6 hours. In 84 patients, LT was associated with major or minor extrahepatic resection (Table 4). Upper abdominal exenteration consisted of en bloc resection of the liver, stomach, pancreas, spleen, and right and transverse colon in all cases. 21 Abdominal exenteration was associated with total small bowel resection in 2 cases. A total of 61 patients underwent splenectomy before or at the time of LT. Of the 196 procedures performed for treatment of the PT, 158 took place before LT, 29 at the time of LT, and 9 after LT. In the remaining 17 cases, the PT was undetected. C 2013 Lippincott Williams & Wilkins 3

4 Le Treut et al Annals of Surgery Volume 00, Number 00, 2013 TABLE 4. Surgical Procedures Performed in Addition to LT, According to the Period of the Study Surgical Procedure (n = 107) (n = 106) Total (n = 213) Major procedures patients Upper abdominal exenteration Whipple procedure Atypical gastropancreatic resection 2 2 Complex inferior vena cava reconstruction 2 2 Total gastrectomy 1 1 Minor procedures patients Extended lymph node dissection Diaphragm resection Peritonectomy 4 4 Distal pancreatectomy ± splenectomy Small bowel resection Adrenalectomy Splenectomy 1 1 Partial gastrectomy 1 1 Section of arcuate ligament 1 1 In addition, 2 cases underwent resection at postoperative day 1 and 1 case at postoperative day 14 after LT, 2 in the first period and 1 in the second period. Pathological Findings Based on data available in 206 cases, the mean weight of the explanted liver was 2905 ± 2412 g (extremes, ,400). It was significantly higher in patients whose indication for LT was tumor bulk (5619 ± 3306 g) than in patients whose indication was hormonal syndrome (2456 ± 1122 g) or oncological treatment (1806 ± 630 g). According to the arbitrary definition used in this study, 117 patients (55%) presented hepatomegaly. The number of LM was less than 10 in 31 cases (15%) including 18 cases of recurrent LM after hepatic resection. The number of LM was 10 or more in 182 cases (85%) including 103 cases characterized as multiple and 18 cases described as countless. Maximum tumor size was 23 cm. Surgical margins were negative (R0) in 193 cases (90%), microscopically invaded (R1) in 13, and macroscopically invaded (R2) in 7. For calculation of DFS, patients with undetected PT were included with R0 cases rather than R2 cases. Histologically, tumors were classified as well differentiated in 168 cases, moderately differentiated in 25, and poorly differentiated in 22. Based on data available in 157 cases, lymph node involvement was present in 66 cases and absent in 91. RESULTS Postoperative Outcome Three-month postoperative mortality was 10% (21/213), mainly due to surgical-related complications, including intraabdominal hemorrhage in 7 cases, hepatic artery thrombosis in 3, primary nonfunction in 3, portal vein thrombosis in 2, peritonitis in 1, and acute pancreatitis in 1. Risk factors for 3-month mortality are listed in Table 5. Patient age, time of LT before or after 2000, indication for LT, and extent of tumor invasion were not significantly related to 3-month mortality. The mean length of postoperative hospitalization was 33 ± 25 days [median, 25 (0 180)], and mean duration of intensive care unit stay was 10 ± 14 days [median, 6 (1 104)]. Retransplantation was performed in 24 cases (11%) including 16 during the first 3 postoperative months. The indications for retransplantation were hepatic artery thrombosis in 10 cases, primary nonfunction in 6, portal or hepatic vein thrombosis in 2, biliary complications in 1, recurrent hepatitis C virus in 1, chronic rejection in 1, and other causes in 3. TABLE 5. Risk Factors of 3-Month Mortality Factor No. Deaths % P RR Early retransplantation 6/16 38% Other cases 15/197 8% Upper abdominal exenteration 4/10 40% Other cases 17/203 8% Splenectomy (at any time) 13/61 21% No splenectomy 8/152 5% Operative duration >10 hr 8/38 21% Operative duration <10 hr 9/127 7% Resection margins R1 R2 5/20 25% Resection margins R0 16/193 8% Hepatomegaly 16/117 14% No hepatomegaly 5/96 5% Surgery in addition to LT 13/84 15% No surgery in addition to LT 8/129 6% Long-term Outcome No patient was lost from follow-up. The mean duration of follow-up was 56 ± 49 months (extremes, months). In 172 cases (81%), the last follow-up examination was carried out at least 5 years after LT. A total of 103 patients died more than 3 months after LT. In 86 of these cases, the cause of death was recurrent disease at a median delay of 52 months after LT (4 165 months). The remaining 17 patients died without recurrence because of late postoperative complications in 6 cases, complications of late retransplantation in 4, infectious complications in 3, and other causes in 4. Overall, 37 patients (17%) died from early or late complications of LT without evidence of recurrent disease. The median duration of survival for patients who died without recurrent disease was 8 months (4 165 months). Eighty-nine patients were alive at the end of the follow-up, without recurrence in 63 cases and with recurrence in 26 cases. The median duration of OS after LT was 67 months. The OS rates at 1, 2, 3, 4, and 5 years were 81%, 73%, 65%, 55%, and 52%, respectively. The median duration of DFS was 24 months. The DFS rates at 1, 2, 3, 4, and 5 years were 65%, 49%, 40%, 33%, and 30%, respectively. The OS rate calculated from the time of diagnosis of LM was 73%. 4 C 2013 Lippincott Williams & Wilkins

5 Annals of Surgery Volume 00, Number 00, 2013 Liver Transplantation for Neuroendocrine Tumors For the 29 patients in whom PT was undetected at the time of LT, the OS rate was 59%. In the 12 patients in whom PT was discovered and removed during (n = 3) or after (n = 9) LT, the 5-year OS rate was 65%. Intervals for discovery of the PT after LT ranged from 6 to 46 months. The PT site was in the digestive tract in 8 cases and in the pancreas in 4 cases (Table 1). For the 17 patients in whom the PT was never identified, the 5-year OS rate was 54% (difference between groups, Non significant). Prognostic Factors and Trends in Patient Selection The results of univariate analysis of prognostic factors are presented in Table 6. Neither age regardless of cutoff (45, 50, or 55 years) nor time interval between diagnosis and LT was correlated to long-term survival. Estimated extent of liver involvement was less predictive than hepatomegaly. Only factors available before LT were used in multivariate analysis. Three were found to be independent predictors of poor prognosis, i.e., major resection in addition to LT (HR: 3.1, 95% CI: , p < ), poor tumor differentiation (HR: 2.7, 95% CI: , p < ), and hepatomegaly (HR: 2.3, 95% CI: , p < ). Long-term results improved significantly after 2000 (59% versus 46% OS), with no significant difference in 3-month postoperative mortality (8/106 vs 13/107). The second period was also characterized by a doubling of the time interval between diagnosis and LT and a significant increase in the rate of 5-year survival from diagnosis of LM (62% to 84%). Comparison of the patient population treated before and after 2000 revealed a number of differences (Table 7), with significantly fewer patients presenting poor prognosis factors in particular, the 3 independent predictors. There were also fewer patients with hepatomegaly, and only 1 patient underwent major resection in addition to LT. The number of poorly differentiated tumors decreased but not significantly. A separate multivariate analysis was performed on the 106 patients treated after Results showed the following independent predictors of poor prognosis: hepatomegaly (HR: 2.6, 95% CI: , p = 0.006), age over 45 years (HR: 2.0, 95% CI: , p = 0.073), and resection in addition to LT (HR: 1.9, 95% CI: , p = 0.058). Based on cumulation of these factors in the same patient, a 4-point prognostic scale was devised (Fig. 2). Five-year OS and DFS rates were 79% and 57%, respectively, for the 58 patients presenting 0 or 1 factor versus 38% and 19% for the 48 patients presenting 2 or 3 factors. Differences in both OS and DFS between groups were significant. DISCUSSION This ELTR study presents by far the largest cohort of LT for NET in the literature. Its greatest advantage is to describe a wide range of clinical, pathological, and therapeutic factors potentially affecting long-term outcome. Despite several weaknesses, that is, retrospective design, 3-decade-long study period, and failure to collect cases from some expert centers, the results of this study are meaningful. The most striking finding is that OS rate in the whole series was more than 50% at 5 years, thus validating the use of LT for patients with cancer. 22 Despite this validation, the actual benefit of LT needs to be proven. Early studies reported 5-year survival rates ranging from 20% to 30% in patients with untreated metastatic NET. 23,24 Currently, no patient with NET goes untreated. Analysis of more than 35,000 patients with NET within the Surveillance, Epidemiology and End Results database showed that 5-year OS rate was 35% for patients with metastatic well-differentiated G1/G2 NET. Median survival had dramatically increased to 39 months for patients diagnosed during the period, possibly due to the use of somatostatin analogue since In specialized centers, 5-year OS rate from the diagnosis of LM in patients presenting metastatic well-differentiated NET after nontransplant treatment now exceeds 50%. 3,26 However, it should be recalled that LT is usually performed after all other treatments have been exhausted. 16,27 In this ELTR study, the median value of diagnosis-to-lt interval has risen 34 months for patients undergoing LT since The 5-year OS rate from diagnosis of LM was 73% for the whole series and 84% for patients undergoing LT since These rates are far higher than the OS rates of nontransplant treatment, suggesting a benefit of transplantation in patients with NET. The 3 poor prognostic factors identified in this Europe-wide study are consistent with those described in 2 earlier French multicentric studies. Association of LT with major resection, especially upper abdominal exenteration was the most unfavorable factor in both previous studies. 15,16 Poor survival in these cases is partly related to the increased postoperative mortality rate, but it also reflects the extensive tumor load necessitating such procedures. 17 Similarly, it is likely that the impact of minor procedures on prognosis observed in our last 106 cases is related to the extrahepatic spread. Hepatomegaly, arbitrarily defined as enlargement of the explanted liver by 25% or more above the patient s predicted liver volume, somehow a surrogate of liver involvement, 5 was already identified as a predictor of poor prognosis in the second French series. 16 The third unfavorable predictor, that is, poor tumor differentiation appeared in the form of noncarcinoid tumors in the first French multicentric series. 15 Since then, carcinoid tumors have been classified as low-grade NET in the WHO classification. 5 Because of small sample size, poor differentiation no longer appeared as an independent factor in multivariate analysis of the 106 patients operated after Our study also shows that 5-year OS rate has improved over time up to 59% for patients operated on during the last 10 years versus 46% for patients operated previously. This improvement was associated with changes in both patient selection and surgical techniques used by European teams (Table 7). Regarding patient selection, recent experience was characterized by fewer patients with poorly differentiated NET, hepatomegaly, or for tumor bulk indications. It is also noteworthy that the number of patients undergoing medical or surgical treatment before LT has increased as witnessed by the longer interval between diagnosis of LM and LT (median, 34 vs 17 months). Regarding surgical technique, a major change since 2000 is avoidance of upper abdominal exenteration and Whipple procedure (1 case). Analysis using our prognostic scale indicates that OS and DFS rates could be further improved to 80% and 60%, respectively, by excluding patients with 2 or 3 poor prognostic factors, because a single factor had no impact on survival (Fig. 2). Comparison of this study with other multicentric studies (Table 8) showed similarities regarding survival. The 5-year OS and DFS rates described herein were consistent with those reported in 3 American studies, 18,19,28 that is, 52% versus 49% and 30% versus 32%, respectively. Comparison also demonstrated similarity regarding prognostic factors. As in this report and our previous studies, 15,16 upper abdominal exenteration, or Whipple procedure, and/or multivisceral graft were adverse prognostic factors in the Lehnert s compiled series 17 and in 1 US multicentric series. 28 As in our last 106 cases, age more than 50 years was identified as a poor prognostic factor, with an HR comparable to that described by Lehnert. 17 In 1 United Network of Organ Sharing (UNOS) database study, multivariate analysis identified only nonspecific prognostic factors such as albumin or bilirubin level because of a lack of pertinent clinical and pathological data. 19 In the other UNOS database study, univariate analysis showed higher 5-year OS in patients who waited more than 2 months before undergoing LT: 63% versus 36%. This suggests that it may be appropriate to wait for disease to stabilize before considering patients for LT. 18 The only prospective series on the topic was published by the Milan group that has for 15 years been performing LT with curative intent using strict selection criteria, that is, low-grade NET, PT C 2013 Lippincott Williams & Wilkins 5

6 Le Treut et al Annals of Surgery Volume 00, Number 00, 2013 TABLE 6. Prognostic Factors in Univariate Analysis Factor N Median Survival 5-Yr Survival, % P Period mo 46 P < 0.05 Period >60 mo 59 Age 50 yr 127 >60 mo 55 NS Age >50 yr mo 47 Resection of the PT before LT 158 >60 mo 56 vs others Resection of the PT during LT mo 22 P < Resection of the PT after LT 9 >60 mo 75 PT never resected 17 >60 mo 54 Onset of metastasis Prevalent mo 50 vs others NS Synchronous mo 49 Metachronous 49 >60 mo 62 Diagnosis-to-LT interval <36 mo 137 >60 mo 53 NS Diagnosis-to-LT interval 36 mo 72 >60 mo 52 Not reported 4 25 Hepatomegaly mo 39 P < Other cases 96 >60 mo 68 Indication for LT Oncological 110 >60 mo 55 vs others Tumor bulk mo 35 P < Hormonal syndrome 37 >60 mo 57 Iatrogenic complications 10 >60 mo 79 Misdiagnosis 4 >60 mo 75 Estimated tumoral invasion <20% 31 >60 mo 78 20% 40% 43 >60 mo 62 40% 60% mo 44 60% 80% mo 42 >80% mo 42 Not reported 13 >60 mo 51 50% 109 >60 mo 61 P < 0.02 >50% mo 42 Not reported 13 > 60 mo 51 PT location (definitive) Bronchial tree 16 >60 mo 53 vs P < 0.05 Stomach 8 >60 mo 75 vs others Small bowel 65 >60 mo 62 P < 0.02 Large bowel mo 40 Duodenum/pancreas mo 44 Unlocalized (never identified) 17 >60 mo 54 PT location detected before LT Bronchial/digestive tract 91 >60 mo 59 vs Duodenum/pancreas mo 43 P < 0.01 Unlocalized before LT 29 > 60 mo 59 Histology Well differentiated 191 >60 mo 55 P < 0.01 Poorly differentiated mo 27 Upper abdominal exenteration 10 6 mo 20 P < Other cases 203 >60 mo 53 Splenectomy before or at LT mo 37 P < No splenectomy 152 >60 mo 58 Major resection in addition to LT mo 21 P < Minor resection or no resection 189 >60 mo 56 Resection margins R0 193 >60 mo 56 P < R1/R mo 15 Lymph node status Positive mo 41 vs P < 0.01 Negative 91 >60 mo 62 Not reported mo 46 NS indicates non significant. 6 C 2013 Lippincott Williams & Wilkins

7 Annals of Surgery Volume 00, Number 00, 2013 Liver Transplantation for Neuroendocrine Tumors TABLE 7. Changes in Patient Characteristics and Results Over Time (n = 107) (n = 106) P Hepatomegaly present <0.001 LT for tumor bulk <0.01 LT for oncological treatment <0.001 PT resected before LT <0.05 PT never resected 4 13 <0.02 Chemotherapy before LT <0.05 TACE before LT <0.01 Median diagnosis-to-lt interval (mo) 17 (1 122) 34 (1 149) <0.001 Upper abdominal exenteration 10 0 <0.01 Major procedure in addition to LT 23 1 <0.001 Any procedure in addition to LT <0.05 Poor tumor differentiation 14 8 NS Overall 5-yr survival from LT 46% 59% <0.05 Disease-free 5-yr survival from LT 22% 39% <0.001 Overall 5-yr survival from diagnosis of LM 62% 84% <0.01 NS indicates... ; non significant TACE, transarterial chemoembolization. TABLE 8. Multicentric Series of Liver Transplantation for Metastatic Neuroendocrine Tumors Authors LeTreutetal 15 Lehnert 17 LeTreutetal 16 Sher 28 Gedaly et al 18 N Guyen et al 19 This study Type and Period of the Study N Overall 5-Yr Survival Disease-Free 5-Yr Survival Adverse Prognostic Factors ( Univariate or Multivariate Analysis) French multicentric study, % 17% Upper abdominal exenteration, noncarcinoid tumors Compilation of literature, % 24% Upper abdominal exenteration or Whipple s operation at LT (HR: 4.8), age > 50 yr (HR: 2.1) French multicentric study, 85 47% 20% Upper abdominal exenteration (HR: ), pancreatic primary tumor (HR: 2.9), hepatomegaly (HR: 2.8) US multicentric study, 83 49% NR Multivisceral grafting, resection of primary tumor at LT UNOS database analysis, % 32% Wait time <67 d UNOS database analysis, % NR Lower albumin level at LT Period % NR Low albumin and high bilirubin levels at LT, high donor creatinine ELTR multicentric study, % 30% Major procedure in addition to LT (HR: 3.1), poorly differentiated tumor (HR: 2.7), hepatomegaly (HR: 2.3) Period % 39% Hepatomegaly (HR: 2.6), any procedure in addition to LT (HR: 1.9), age 45 yr (HR: 2) ELTR indicates European Liver Transplant Registry; HR, hazard ratio in case of multivariate analysis; NR, not reported; UNOS, United Network of Organ Sharing. drained by the portal system and obligatorily removed before LT, liver involvement of 50% or less, age 55 years or less, and stable disease for at least 6 months before LT. 6 Only 10% of patients were symptomatic. Using these criteria, the Milan group has achieved remarkable results, with 96% OS and 80% DFS at the last follow-up in 30 patients. 5 However, such stringent criteria probably denied LT to many patients who might have benefited. Three arguments can be proposed to support the use of more liberal criteria. First, other groups have obtained comparable OS rates (70% 90%) in cohorts including 60% to 100% of symptomatic patients 8,10,11 and extensive liver involvement with median rates of 88% 8 or more than 50% in 80% of cases. 10 Second, 5-year survival rates after LT in the ELTR series were 53% for patients with NET from bronchial PT and 59% for patients in whom PT could not be detected before LT (Table 6). Third, if the Milan criteria (except for disease stability that was unavailable in our cohort) had been used in our ELTR series, only 38 of the last 106 patients would have been selected and the OS and DFS would have been 79% and 51%, respectively (data not shown). In comparison, if the more liberal criteria established herein had been used, an additional 20 patients (n = 58) would have been selected with no detriment to survival rates (Fig. 2). At the other end of the selection spectrum, these criteria are debatable because some multidisciplinary oncological committees apply a wait-andsee policy for management of asymptomatic patients presenting C 2013 Lippincott Williams & Wilkins 7

8 Le Treut et al Annals of Surgery Volume 00, Number 00, 2013 well-differentiated NET with liver involvement of less than 50% and stable disease. 2,3,26 Current evidence 3,29 31 suggests a growing consensus around the following recommendations concerning LT in patients with NET. 1. LT is a valid option only for operable patients presenting unresectable symptomatic or asymptomatic LM of NET, including cases with recurrent LM after one or more previous liver resections (27% in this series). 2. LT is suitable only for patients in whom disease is confined to the liver. This mandatory condition underlines the need for thorough staging with frequent reevaluation using positron emission tomography/computed tomography, somatostatin receptor scintigraphy, or other scintigraphy techniques to rule out extrahepatic disease, particularly bone metastases, 3,8,31,32 and for exploratory laparotomy or laparoscopy to detect peritoneal deposits. 8,10,13 LT in patients with undetected PT (ie, 14% in this series, as compared with 13% in the general population of patients with NET 25 )remains controversial, but the 59% OS rate reported herein suggests that failure to detect the PT should not be considered as an absolute contraindication. 3. LT should only be performed in patients with well-differentiated NET (NET G1/G2). Poor tumor differentiation (neuroendocrine carcinoma G3) is unanimously considered as a contraindication for LT. 3,30 With regard to tumor proliferation, it is interesting to note that the Hannover group 11 has recommended a Ki 67 index of less than 10%. Almost all groups comply with this recommendation, but some allow higher thresholds in hotspots: 15% 10 and even up to 20%, 5 that is, the upper limit for this index for G2 NET. 4 In practice, a large proportion of patients with well-differentiated NET have a Ki 67 of more than 15%, especially when the PT is located in the pancreas LT should not be associated with major extrahepatic resection. Ideally, the PT should be removed before LT, especially if it is located in the pancreatic head. This strategy now seems to be widespread, because no upper abdominal exenteration and only 1 concurrent Whipple procedure have been reported in the last 10 years (Table 4). FIGURE 2. Overall survival of 106 patients according to the number of adverse prognostic factors: hepatomegaly, resection in addition to LT, and age more than 45 years. Top: 5- year survival rates: 77%, 79%, 39%, and 33% for 0, 1, 2, or 3 factors, respectively. Bottom: after gathering in 2 groups. 0 1 factors (n = 58) and 2 3 factors (n = 48). P < Using these recommendations, a 5-year OS rate of nearly 60% can be easily achieved as demonstrated over the last 10 years in the European experience presented. It would even seem possible to reach the thresholds of 70% OS and 50% DFS rates advocated by some for patients with cancer 5 by contraindicating patients with more than one of the unfavorable prognostic factors described in this study. From a clinical standpoint, however, it seems questionable to exclude patients older than 45 years with hepatomegaly simply to improve statistics. The most controversial point involves timing of LT. Many authors think that LT should be indicated in patients who respond to medical treatment and/or have presented stable LM for 6 to 12 months. 6,13,18,19,27,31 Others think that LT should be indicated in patients who are refractory to systemic medical treatment and present progressing hepatic tumor load. 7,8,10,17,32 This debate cannot be resolved by retrospective studies that rarely mention criteria such as LM stability and response to treatment. One US series reported a correlation between longer waiting time and better outcome, suggesting greater benefit for patients with stable disease. 18 Actually, waiting time depends mainly on organ allocation rules and does not presuppose disease stability at the time of listing. This ELTR study focused on diagnosis-to-lt interval that takes into account natural history of the disease, response to pretransplant treatment, choice of LT by the transplant team, and also waiting time. As in the other reports, 11 13,15,16 univariate analysis did not identify this interval as a prognostic factor. However, it is interesting to note that doubling of its median value in latter part of this study was accompanied by an improvement in survival not only from diagnosis of LM but also after LT. In other words, lengthening of the diagnosis-to-lt interval had no detrimental effect on the outcome of LT. This finding suggests that in asymptomatic cases, LT should be postponed until patients show signs of becoming refractory to other treatment. ACKNOWLEDGMENTS The authors thank Vincent Karam, data manager of the ELTR, for his support in identifying patients for the study, and the other contributors who assisted to data collection: C Letoublon, Hôpital Michallon, Grenoble, France; J Baulieux, Hôpital Croix Rousse, Lyon, France; J Benhamou, Hôpital Bichat, Paris, France; 8 C 2013 Lippincott Williams & Wilkins

9 Annals of Surgery Volume 00, Number 00, 2013 Liver Transplantation for Neuroendocrine Tumors K Boudjema, Hôpital Pontchaillou, Rennes, France; J Saric, Hôpital Pellegrin, Bordeaux, France; D Görog, Semmelweis University, Budapest, Hungary; F Navarro, Hôpital St Eloi, Montpellier, France; A Königsrainer, Universitätklinikum, Tübingen, Germany; U Baccarini, Policlinico Universitario, Udine, Italy; and T Meyer, Universitätklinikum, Würzburg, Germany. Y.P.L.T. initiated the study, prepared the study design, and was responsible for study management, data collection and interpretation, manuscript preparation, and submission. E.G. assisted Y.P.L.T. for data collection and interpretation. E.J. performed the statistical analysis. All other coauthors were responsible for data collection and critical review of the manuscript. The order of the coauthors is based on number of patients (3 or more) included in each center. REFERENCES 1. Pape UF, Böhming M, Berndt U, et al. Survival and clinical outcome of patients with neuroendocrine tumors of the gastroenteropancreatic tract in a German referral center. Ann NY Acad Sci. 2004;1014: Hentic O, Couvelard A, Rebours V, et al. Ki-67 index, tumor differentiation, and extent of liver involvement are independent prognostic factors in patients with liver metastases of digestive endocrine carcinomas. Endocr Relat Cancer. 2010;18 : Pavel M, Baudin E, Couvelard A, et al. ENETS Consensus Guidelines for the management of patients with liver and others distant metastases from neuroendocrine neoplasms of foregut, midgut, hindgut, and unknown primary. Neuroendocrinology. 2012;95: Rindi G, Arnold R, Bosman FT. Nomenclature and classification of neuroendocrine neoplasms of the digestive system. In: Bosman FT, Carneiro F, Hruban H, Theise ND, eds. WHO Classification of Tumours of the Digestive System. Lyon: IARC Press; 2010: Mazzaferro V. Surgical approach to the treatment of liver metastases from neuroendocrine tumors: hepatic resection and liver transplantation. Tumori. 2010;96: Mazzaferro V, Pulvirenti A, Coppa J. Neuroendocrine tumors metastatic to the liver:how to select patients for liver transplantation? J Hepatol. 2007;47: Florman S, Toure B, Kim L, et al. Liver transplantation for neuroendocrine tumors. J Gastrointest Surg. 2004;8: Frilling A, Malago M, Weber F, et al. Liver transplantation for patients with metastatic endocrine tumors: single-center experience with 15 patients. Liver Transpl. 2006;12: Marin C, Robles R, Fernandez JA, et al. Role of liver transplantation in the management of unresectable neuroendocrine liver metastases. Transplant Proc. 2007;39: Olausson M, Friman S, Herlenius G, et al. Orthotopic liver or multivisceral transplantation as treatment of metastatic neuroendocrine tumors. Liver Transpl. 2007;13: Rosenau J, Bahr MJ, von Wasielewski R, et al. Ki67, E-cadherin, and p53 as prognostic indicators of long-term outcome after liver transplantation for metastatic neuroendocrine tumors. Transplantation. 2002;73: Routley D, Ramage JK, McPeake J, et al. Orthotopic liver transplantation in the treatment of metastatic neuroendocrine tumors of the liver. Liver Transpl Surg. 1995;1: van Vilsteren FG, Baskin-Bey ES, Nagorney DM, et al. Liver transplantation for gastroenteropancreatic neuroendocrine cancers: defining selection criteria to improve survival. Liver Transpl. 2006;12: Alessiani M, Tzakis A, Todo S, et al. Assessment of five-year experience with abdominal organ cluster transplantation. J Am Coll Surg. 1995;180: Le Treut YP, Delpero JR, Dousset B, et al. Results of liver transplantation in the treatment of metastatic neuroendocrine tumors. A 31-case French multicentric report. Ann Surg. 1997;225: Le Treut YP, Gregoire E, Belghiti J, et al. Predictors of long-term survival after liver transplantation for metastatic endocrine tumors: an 85-case French multicentric report. Am J Transplant. 2008;8: Lehnert T. Liver transplantation for metastatic neuroendocrine carcinoma: an analysis of 103 patients. Transplantation. 1998;66: Gedaly R, Daily MF, Davenport D, et al. Liver transplantation for the treatment of liver metastases from neuroendocrine tumors: an analysis of the UNOS database. Arch Surg. 2011;146: N Guyen NT, Harring TR, Goss JA, et al. Neuroendocrine liver metastases and orthotopic liver transplantation: the US experience. Int J Hepatol : doi: /2011/ Vauthey JN, Abdalla EK, Doherty DA, et al. Body surface area and body weight predict total liver volume in western adults. Liver Transpl. 2002;8: Starzl TE, Todo S, Tzakis A. Abdominal organ cluster transplantation for the treatment of upper abdominal malignancies. Ann Surg. 1989;210: O Grady JG. Treatment options for other hepatic malignancies. Liver Transpl. 2000;6:S23 S Moertel CG. An Odyssey in the land of small tumor. J Clin Oncol. 1987;5: Soreide O, Berstad T, Bakka A, et al. Surgical treatment as a principle in patients with advanced abdominal carcinoids tumors. Surgery. 1992;111: Yao JC, Hassan M, Phan A, et al. One hundred years after carcinoid : epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in United States. J Clin Oncol. 2008;26: Durante C, Boukheris H, Dromain C, et al. Prognostic factors influencing survival from metastatic (stage IV) gastroenteropancreatic well-differentiated endocrine carcinoma. Endocr Relat Cancer. 2009;16: Sutcliffe R, Maguire D, Ramage J, et al. Management of neuroendocrine liver metastases. Am J Surg. 2004;187: Sher L. Association between timing of resection of primary tumor in patients undergoing liver transplantation for metastatic neuroendocrine tumor and survival. Am J Transplant. 2009;12: Pascher A, Klupp J, Neuhaus P. Endocrine tumours of the gastrointestinal tract. Transplantation in the management of metastatic endocrine tumours. Best Pract Res Clin Gastroenterol. 2005;19: Grégoire E, Le Treut YP. Liver transplantation for primary and secondary endocrine tumors. Transplant Int. 2010;23: Bonnacorsi-Riani E, Apestegui C, Jouret-Mourin A, et al. Liver transplantation and neuroendocrine tumors: lessons from a single centre experience and from the literature review. Transplant Int. 2010;23: Frilling A, Li J, Malamutmann E, et al. Treatment of liver metastases from neuroendocrine tumours in relation to the extent of hepatic disease. Br J Surg. 2009;96: C 2013 Lippincott Williams & Wilkins 9