Review Article
Therapeutic Strategies for Neuroendocrine Liver Metastases Andrea Frilling, MD, PhD1; and Ashley K. Clift, BA2
Patients who have neuroendocrine tumors frequently present with liver metastases. A wide panel of treatment options exists for these patients. Liver resection with curative intent achieves the best long-term results. Highly selected patients may be considered for liver transplantation. Substantial recurrence rates reported after surgical approaches call for neoadjuvant and adjuvant concepts. Liver-directed, locally ablative procedures are recommended for patients with limited, nonresectable tumor burden. Angiographic liver-directed techniques, such as transarterial embolization, transarterial chemoembolization, and selective internal radiotherapy, offer excellent palliation for patients with liver-predominant disease. Peptide receptor radionuclide therapy is a promising palliative procedure for patients with hepatic and/or extrahepatic metastases. The efficacy of these treatment options needs to be evaluated in randomized trials. Somatostatin analogues have demonstrated effectiveness not only for symptomatic relief in patients with secreting tumors but also for the control of proliferation in small intestinal neuroendocrine tumors and most recently also in those originating from the pancreas. Chemotherapy is an option mainly for those with pancreatic neuroendocrine tumors and high-grade tumors irrespective of the origin. Novel drugs targeting specific pathways within the tumor cell have produced improved progression-free survival compared with placebo in patients with pancreatic neuroendocrine tumors. Despite such a diverse armamentarium, there is uncertainty with regard to the optimal treatment regimens. Newly introduced molecular-based markers, along with the conduction of clinical trials comparing the efficacy of treatment modalities, offer a chance to move the treatment of neuroendocrine tumor disease toward personalized patient care. In this report, the authors review the approaches for treatment of neuroendocrine liver metastases, identify shortcomings, and anticipate future perspectives. Furthermore, clinical practice recommendations are provided for currently available treatment options. Although multiple modalities are available for the treatment of neuroendocrine liver metastases, optimal C 2014 management is unclear. The current knowledge pertaining to these treatment options is analyzed. Cancer 2014;000:000-000. V American Cancer Society. KEYWORDS: neuroendocrine tumors, liver, metastases, treatment.
INTRODUCTION Neuroendocrine (NE) tumors (NETs) of the gastroenteropancreatic system comprise a heterogeneous group of neoplasms that commonly present with metastases at initial diagnosis. Although the precise quantification of the incidence and prevalence of NETs is problematic because of discordance between cancer registries, data collected since 2000 suggest an incidence from 1.3 to 5.7 per 100,000 population per year,1,2 with a prevalence estimate by the third National Cancer Survey and the US Surveillance, Epidemiology, and End Results (SEER) Program of 35 per 100,000.3 The prevalence of NE liver metastases (LMs) remains vague, because most registries do not explicitly refer to LMs and frequently classify them under the umbrella term distant metastases. According to the SEER database, the prevalence of NE LMs is 27%3 compared with 40% to 95% reported by specialized centers.4 The predilection for hepatic metastasis varies between primary tumor sites. Although from 28.3% to 77% of patients with pancreatic NETs and from 67% to 91% of patients with small bowel NETs are diagnosed with LMs,4,5 liver involvement is rarely observed in patients with gastric, appendiceal, or rectal NETs. Alongside tumor grade, LMs represent 1 of the most powerful predictors of survival.5-7 For patients who had localized colorectal NETs, the reported 5-year survival rate ranged from 75% to 88% compared with 30% for those who had LMs.8-10 The reported long-term survival rate for patients with gastrinoma was 95% at 20 years in the absence of LMs and only 15% at 10 years if LMs occurred.11 Surgery epitomizes the management of NE LMs, offering immediate and effective control of symptoms with a prospect for long-term survival. Over the last 2 decades, a panel of novel, promising, nonsurgical therapies has been incorporated into the armamentarium of NE LMs, as critically analyzed in 4 recent, excellent, systematic reviews.12-15 There is a striking lack of quality evidence from large-scale, randomized clinical trials comparing different treatment approaches to
Corresponding author: Andrea Frilling, MD, PhD, Department of Surgery and Cancer, Imperial College London, Hammersmith Campus, Du Cane Road, London W12 0HS, UK; Fax: (011) 00442033133963; [email protected] 1 Department of Surgery and Cancer, Imperial College London, London, United Kingdom; 2School of Medicine, Imperial College London, London, United Kingdom.
DOI: 10.1002/cncr.28760, Received: January 20, 2014; Revised: April 9, 2014; Accepted: April 10, 2014, Published online Month 00, 2014 in Wiley Online Library (wileyonlinelibrary.com)
Cancer
Month 00, 2014
1
Review Article
Figure 1. A management algorithm for neuroendocrine liver metastases. CgA indicates chromogranin A; MRI, magnetic resonance imaging; 68Ga DOTA, 68Gallium dodecanetetra-acetic acid; 18FDG, 18F-fluorodeoxyglucose; PET, positron emission tomography, CT, computed tomography; FNAB, fine-needle aspiration biopsy; NEN, neuroendocrine neoplasm; G1/G2, grades 1 and 2; G3, grade 3; 1/-, with or without; TAE/TACE, transarterial embolization/transarterial chemoembolization; SIRT, selective internal radiotherapy; PRRT, peptide receptor radionuclide therapy; CRR, cytoreductive resection; LT, liver transplantation; SSA, somatostatin analogues; P, use in pancreatic neuroendocrine tumors; CLD & SS, combination of liver-directed and systemic strategies.
guide treatment decisions, and most available experience is based primarily on retrospective analyses of case series without controls. Indeed, 2 Cochrane reviews of treatments for NE LMs reported that a lack of robust evidence precluded their analyses.16,17 In the current review, we collate the currently available evidence on treatment of NE LMs and suggest avenues for future progress. Diagnosis
Contrast-enhanced ultrasonography (CEUS), multiphase helical computed tomography (CT) with a multirow detector scanner, and diffusion-weighted (DW) magnetic resonance imaging (MRI) are currently considered the morphologic imaging modalities of choice for the detection of NE LMs. Compared with CEUS, CT, T2-weighted MRI, and dynamic gadolinium-enhanced MRI, DW MRI detects smaller and more liver deposits.18 Meticulous pathologic examination of resected liver specimens revealed that preoperative imaging resulted in the understaging of NE LMs in >50% of patients.19 Ultrasound-guided or 2
CT-guided biopsy is integral for diagnosis confirmation and tumor grading (low grade [grade 1], antigen identified by a monoclonal antibody Ki-67 [Ki-67] index 2%; intermediate grade [grade 2], Ki-67 index 3%-20%; high grade [grade 3], Ki-67 index >20%).20 Tumor grade provides valuable prognostic information and influences treatment trajectories (Fig. 1). NE LMs may be classified morphologically as type I (single metastasis), type II (isolated metastatic bulk accompanied by smaller deposits), or type III (disseminated metastatic spread).21 Somatostatin receptor-based imaging exploits the observation that the majority of NETs express somatostatin receptors, enabling the visualization of primary and metastatic lesions while concurrently probing for potential amenability to receptor-targeted treatment. Other functional imaging techniques target tumor metabolism. An array of tracer molecules may be used, including 111In-octreotide; the 68 Ga-conjugated radiopharmaceuticals, including 0 1 3 dodecanetetra-acetic acid (DOTA) -Phe -Tyr octreotide (DOTATOC), DOTA0-Tyr3 octreotate (DOTATATE), Cancer
Month 00, 2014
Treatment of Neuroendocrine Liver Metastases/Frilling and Clift
DOTA021-Nal3 octreotide (DOTANOC), and 64CuDOTA; 18F-fluorinated levodopa (18F-DOPA); and 11C–5hydroxytryptophan (11C-5-HTP). The more frequently used 68 Ga-DOTATOC positron emission tomography (PET)/ CT imaging identifies LMs with a sensitivity of 82% to 100% and a specificity of 67% to 100% and identifies extrahepatic disease in grade 1 NETs with a sensitivity of 85% to 96% and a specificity of 67% to 90%.22 It has been demonstrated that results from 68Ga-DOTATOC PET/CT imaging modify the treatment decision in approximately 33% of patients.23,24 Higher grade NETs (grade 2-3) are staged better using 18F-fluorodeoxyglucose (FDG)-PET/CT. Currently used imaging response criteria only insufficiently reflect the distinct characteristics of NETs. Although tumor size may remain stable on morphologic imaging, changes in activity demonstrated by functional imaging may indicate treatment efficacy. Although it is generally considered a biomarker for NETs, the peptide chromogranin A (CgA) is not helpful in diagnosing LMs. However, it can be useful for assessing entire tumor burden, longitudinal monitoring, and estimating prognosis. It has been demonstrated that a CgA decrease >80% after resection of NE LMs predicts both a complete symptomatic response and disease stabilization.25 More recently, a panel of novel, promising NET biomarkers, such as a multitranscript molecular signature encompassing 51 marker genes,26 paraneoplastic antigen Ma2 autoantibodies,27 circulating tumor cells,28 and metabolic spectra,29 has been introduced. The clinical utility of such biomarkers will require further validation in larger series. Treatment Liver resection
Hepatectomy represents the mainstay of treatment for patients with grade 1 or 2 NE LMs (Table 1). Because substantial numbers of patients present with type III LMs on initial diagnosis, only approximately 20% to 30% are candidates for resection. Selected patients with primarily unresectable, bilobar metastases may benefit from preoperative right portal vein embolization with consecutive left liver lobe hypertrophy41 or 2-step resections.42,43 Radiation lobectomy with 90Y-radiolabeled microspheres has the potential to attain local tumor control and hypertrophy of the anticipated future liver remnant in patients with large right liver lobe metastases and small-for-size left liver lobes.44 Difficulties in planning complex surgical procedures can be alleviated by using computer-assisted, 3-dimensional CT technology, which provides precise imaging of liver structures and an accurate volumetric calculation.45 A recent systematic review that analyzed results from 29 reports revealed median 1-year, 3-year, 5-year, and Cancer
Month 00, 2014
10-year survival rates of 94% (range, 79%-100%), 83% (range, 63%-100%), 70.5% (range, 31%-100%), and 42% (range, 0%-100%), respectively.12 Resection with macroscopically clear margins (R0 and R1) was feasible in a median of 63% of patients (range, 38%-100%) and 21% of patients (range, 0%-37%), respectively. Complete and partial symptomatic responses were observed in a median of 73% (range, 23%-92%) and 14% (range, 0%74%) of patients, respectively. The perioperative morbidity and mortality rates ranged from 3% to 45% (median, 23%) and from 0% to 9% (median, 0%), respectively. Contrasting these rather encouraging results were dismal 1-year, 3-year, 5-year, and 10-year recurrence-free survival rates of 63% (range, 50%-80%), 32% (range, 24%69%), 29% (range, 6%-66%), and 1% (0%-11%), respectively. These results were in concordance with a multicentric analysis of 339 patients who underwent resection of NE LMs and had overall 5-year and 10-year survival rates of 75% and 51%, respectively, and recurrence-free survival rates at 1 year, 3 years, and 5 years of 56.9%, 24.2%, and 5.9%, respectively.46 Overall, outcomes of hepatic resection for NE LMs require cautious consideration. In some series, patients with extrahepatic metastases are included; and, in others, results from resection as the only technique are pooled with those from ablation alone and incomplete resection in combination with ablative techniques. Inclusion criteria for liver resection are not well reported, and there are no uniform follow-up protocols. Neoadjuvant and adjuvant treatment concepts for down-staging and reduction of disease recurrence have been introduced in single case reports and smaller series; however, there is an urgent need for further expansion and adoption in the standard treatment pathways. Valid assessment of the efficacy of liver resection for NE LMs remains debatable, because no prospective randomized trials to date have compared outcomes of surgery with outcomes of nonsurgical, liver-directed techniques. Mayo et al47 compared patients who underwent liver resection with those who underwent transarterial embolization (TAE). Surgically treated patients had superior median and 5-year survival compared with the TAE cohort (123 months and 74% vs 34 months and 30%, respectively). Although symptomatic patients who had >25% liver involvement benefited from surgery compared with TAE, there was no difference in long-term survival in the asymptomatic groups. The authors concluded that asymptomatic patients with large hepatic burden would be better treated using TAE, because they would not benefit from an invasive surgical approach. Improved 3
Review Article TABLE 1. Outcomes From Liver Resection for Neuroendocrine Liver Metastases—Selected Single-Center Studies Published Since 2000
Reference
No. of Patients Undergoing Resection
No. of R0 Resections
5-Year OS, %
5-Year DFS, %
74 41 27 18 47 170 16 31 20 16 34 26
28 26 23 15 25 75 16 10 20 16 15 13
NR 88 100 86 74
NR 31 96 90 66 76
Saxena 201130 Scigliano 200931 Frilling 200921 Gomez 200732 Elias 200333 Sarmiento 200334 Norton 200335 Nave 200136 Coppa 200137 Yao 200138 Chamberlain 200039 Pascher 200040
82 86 67 70 85 NR
Other Survival Data After R0 Resection Median OS, 73 mo
29
NR
Median survival, 70 mo
30-Day Morbidity and Mortality Mortality, 1.3% (4% at 60 d) Morbidity, 14%; mortality, 0% Morbidity, 7.4%; mortality, 0% Morbidity, 22%; mortality, 5.6% Morbidity, 45%; mortality, 5% Morbidity, 14%; mortality, 1.2% Morbidity, 19%; mortality. 0% Morbidity, 13% NR Morbidity, 12%; mortality, 0% Mortality, 6% Mortality, 0%
Abbreviations: DFS, disease-free survival; NR, not reported; OS, overall survival; R0, resection with negative margins.
survival outcomes of patients who underwent liver resection, compared retrospectively with those who underwent TAE (76% vs 50% at 5 years) and those who underwent medical treatment (83% vs 39% at 3 years), also were reported by Chamberlain et al.39 Liver resection should be considered as the treatment of first choice for patients who are fit to undergo hepatic resection for grade 1 or 2 LMs and have no extrahepatic disease. Those who have extrahepatic lesions also may be considered if their tumor burden is entirely resectable. A complete tumor resection should be anticipated. Debulking liver surgery may be indicated in individual patients with hormonally active tumor burden who are refractory to nonsurgical treatment options. Liver transplantation
Patients with unresectable NE LMs may be considered as potential candidates for liver transplantation. Higher tumor grade, nonportal tumor drainage, extrahepatic metastases (with the exception of resectable perihilar lymph node metastases), and advanced carcinoid heart disease are generally accepted contraindications. Other criteria, including age 55 years and a hepatic tumor load 50%, have been suggested by some centers48; however, their validity was contradicted by others.49 In singlecenter series published since 2000, the 3-year and 5-year overall survival rates ranged from 57% to 100% and from 33% to 90%, respectively, and disease-free survival rates from 11% to 77% at 5 years have been reported50-53 (Table 2). Although short-term results have improved because of better selection of transplantation candidates, refinement of surgical techniques, and the introduction of novel immunosuppressive regimens, the high recur4
rence rate remains a matter of concern. Patient-specific biomarkers are needed for the identification of those who gain a long-term benefit from the procedure. Like patients who are followed after transplantation for hepatocellular carcinoma, down-staging while on the waiting list should become part of the transplantation protocol for NE LMs. Liver transplantation should be offered not as an ultima ratio solution after failure of all other treatment attempts but, rather, as a well timed procedure in the still controllable stage of the disease. A recent retrospective analysis of 213 patients with NET undergoing liver transplantation in European centers between 1982 and 2009 revealed 1-year, 2-year, 3-year, and 5-year overall survival rates of 81%, 73%, 65%, and 52%, respectively, and a disease-free survival rate of 30% at 5 years.57 Hepatomegaly, age >45 years, and concomitant transplantation and resection were identified as unfavorable prognostic factors. It remains a matter of debate whether surgical morbidity rather than tumor biology contributed to the poor outcome in patients who underwent concomitant procedures. The 3-month postoperative mortality rate was 10% and was mainly attributed to surgical complications, including intra-abdominal hemorrhage, hepatic artery thrombosis, portal vein thrombosis, peritonitis, pancreatitis, and primary nonfunction. Among the risk factors for 3-month mortality were early retransplantation, upper abdominal exenteration, splenectomy, operative duration >10 hours, and hepatomegaly. Seventeen patients died while recurrence free because of late morbidity related to transplantation. In total, early or late complications of transplantation accounted for death in 37 recurrence-free patients (17%). The median survival in that cohort was 8 months (range, Cancer
Month 00, 2014
Treatment of Neuroendocrine Liver Metastases/Frilling and Clift
TABLE 2. Liver Transplantation for Neuroendocrine Metastases—Selection of Single-Center Series Published Since 2000 OS, % No. of Patients
Reference Bonaccorsi-Riani 201050 Olausson 200749 Marin 200754 Mazzaferro 200748 van Vilsteren 200655 Frilling 200651 Florman 200452 Cahlin 200356 Rosenau 200253 Coppa 200137 Pascher 200040
9 15a 10 24 19 15 11 7 19 9 4
1-Year
2-Year
3-Year
88
77
86
57
DFS, % 5-Year
10-Year
33 90
1-Year 67 70
33
5-Year
10-Year
11 20
38 90
88 78.3 73
3-Year
77 80 69.4
67.2 36
48.3
80 89 100 100
100 75
100 75
80 70
50
56
21 53
21
Abbreviations: DFS, disease-free survival; OS, overall survival. a These included 5 patients who underwent visceral transplantations.
4-165 months). In US transplantation centers, comparable outcomes for patients who underwent transplantation for hepatocellular carcinoma and those undergoing liver transplantation for NE metastases were reported.58,59 A review of the United Network for Organ Sharing database59 comprised 150 patients who underwent liver transplantation either in isolation or as a component of multivisceral transplantation. Survival rates at 1 year, 3 years, and 5 years for patients who received only a liver graft were comparable to those obtained in patients who underwent multivisceral transplantation (81% vs 80%, 65% vs 65%, and 49% vs 48%, respectively). The disease-free survival rates at 1 year, 3 years, and 5 years were 77%, 50%, and 32%, respectively, in 83 evaluable patients. In the group of patients who underwent transplantation for carcinoids, the patient and graft survival rates were comparable at 1 year, 3 years, and 5 years (76%, 55%, and 47%, respectively, and 71%, 51%, and 44%, respectively). Overall, results from multivisceral transplantations for NE LMs appear to be contradictory. Although promising results were reported for highly selected patients by some groups,49,59,60 others have discouraged this approach.61 Liver transplantation should be offered to strictly selected patients with nonresectable LMs. Grade 3 tumors, nonportal systemic tumor drainage, extrahepatic metastases with the exception of perihilar lymph node metastases, and carcinoid heart disease in general are considered exclusion criteria. Because of the imminent shortage of cadaveric organ donors, consideration of living-donor liver transplantation should be implemented in the evaluation process. Cancer
Month 00, 2014
Locally ablative techniques
In patients with hepatic metastases who are ineligible for complete resection, ablative modalities can be used either as stand-alone therapy (through percutaneous, laparoscopic,62 or open approaches) or as surgical adjuncts.63 Various methods are in clinical use, including radiofrequency ablation,64 microwave ablation,65 laser ablation,66 and cryoablation.67 Although radiofrequency ablation has been used most extensively, microwave ablation provides comparable efficacy but is less time consuming and, with laser application, allows for more accurate, MRI-guided monitoring of treatment. Overall 5-year survival rates range from 37% to 57%, and the most favorable results are reported in patients who have limited hepatic tumor burden, dominant liver lesions measuring 1 cm.65,67,68 In the largest reported series on 89 patients with NE LMs who underwent 119 laparoscopic radiofrequency ablation treatment sessions, patients achieved a median overall survival of 6 years after the first treatment session and a median disease-free survival of 1.3 years. Local recurrence within the liver occurred in 23% of patients, 63% developed new lesions, and extrahepatic disease became evident in 53% at a median follow-up of 3063 months. Amelioration of symptoms was achieved in 97% of patients. The postinterventional morbidity rate in that study was 6%, and the 30-day mortality rate was 1%.68 Locally ablative techniques represent a safe and repeatable palliative treatment for patients who have a limited number of unresectable LMs that preferably measure
Therapeutic Strategies for Neuroendocrine Liver Metastases Andrea Frilling, MD, PhD1; and Ashley K. Clift, BA2
Patients who have neuroendocrine tumors frequently present with liver metastases. A wide panel of treatment options exists for these patients. Liver resection with curative intent achieves the best long-term results. Highly selected patients may be considered for liver transplantation. Substantial recurrence rates reported after surgical approaches call for neoadjuvant and adjuvant concepts. Liver-directed, locally ablative procedures are recommended for patients with limited, nonresectable tumor burden. Angiographic liver-directed techniques, such as transarterial embolization, transarterial chemoembolization, and selective internal radiotherapy, offer excellent palliation for patients with liver-predominant disease. Peptide receptor radionuclide therapy is a promising palliative procedure for patients with hepatic and/or extrahepatic metastases. The efficacy of these treatment options needs to be evaluated in randomized trials. Somatostatin analogues have demonstrated effectiveness not only for symptomatic relief in patients with secreting tumors but also for the control of proliferation in small intestinal neuroendocrine tumors and most recently also in those originating from the pancreas. Chemotherapy is an option mainly for those with pancreatic neuroendocrine tumors and high-grade tumors irrespective of the origin. Novel drugs targeting specific pathways within the tumor cell have produced improved progression-free survival compared with placebo in patients with pancreatic neuroendocrine tumors. Despite such a diverse armamentarium, there is uncertainty with regard to the optimal treatment regimens. Newly introduced molecular-based markers, along with the conduction of clinical trials comparing the efficacy of treatment modalities, offer a chance to move the treatment of neuroendocrine tumor disease toward personalized patient care. In this report, the authors review the approaches for treatment of neuroendocrine liver metastases, identify shortcomings, and anticipate future perspectives. Furthermore, clinical practice recommendations are provided for currently available treatment options. Although multiple modalities are available for the treatment of neuroendocrine liver metastases, optimal C 2014 management is unclear. The current knowledge pertaining to these treatment options is analyzed. Cancer 2014;000:000-000. V American Cancer Society. KEYWORDS: neuroendocrine tumors, liver, metastases, treatment.
INTRODUCTION Neuroendocrine (NE) tumors (NETs) of the gastroenteropancreatic system comprise a heterogeneous group of neoplasms that commonly present with metastases at initial diagnosis. Although the precise quantification of the incidence and prevalence of NETs is problematic because of discordance between cancer registries, data collected since 2000 suggest an incidence from 1.3 to 5.7 per 100,000 population per year,1,2 with a prevalence estimate by the third National Cancer Survey and the US Surveillance, Epidemiology, and End Results (SEER) Program of 35 per 100,000.3 The prevalence of NE liver metastases (LMs) remains vague, because most registries do not explicitly refer to LMs and frequently classify them under the umbrella term distant metastases. According to the SEER database, the prevalence of NE LMs is 27%3 compared with 40% to 95% reported by specialized centers.4 The predilection for hepatic metastasis varies between primary tumor sites. Although from 28.3% to 77% of patients with pancreatic NETs and from 67% to 91% of patients with small bowel NETs are diagnosed with LMs,4,5 liver involvement is rarely observed in patients with gastric, appendiceal, or rectal NETs. Alongside tumor grade, LMs represent 1 of the most powerful predictors of survival.5-7 For patients who had localized colorectal NETs, the reported 5-year survival rate ranged from 75% to 88% compared with 30% for those who had LMs.8-10 The reported long-term survival rate for patients with gastrinoma was 95% at 20 years in the absence of LMs and only 15% at 10 years if LMs occurred.11 Surgery epitomizes the management of NE LMs, offering immediate and effective control of symptoms with a prospect for long-term survival. Over the last 2 decades, a panel of novel, promising, nonsurgical therapies has been incorporated into the armamentarium of NE LMs, as critically analyzed in 4 recent, excellent, systematic reviews.12-15 There is a striking lack of quality evidence from large-scale, randomized clinical trials comparing different treatment approaches to
Corresponding author: Andrea Frilling, MD, PhD, Department of Surgery and Cancer, Imperial College London, Hammersmith Campus, Du Cane Road, London W12 0HS, UK; Fax: (011) 00442033133963; [email protected] 1 Department of Surgery and Cancer, Imperial College London, London, United Kingdom; 2School of Medicine, Imperial College London, London, United Kingdom.
DOI: 10.1002/cncr.28760, Received: January 20, 2014; Revised: April 9, 2014; Accepted: April 10, 2014, Published online Month 00, 2014 in Wiley Online Library (wileyonlinelibrary.com)
Cancer
Month 00, 2014
1
Review Article
Figure 1. A management algorithm for neuroendocrine liver metastases. CgA indicates chromogranin A; MRI, magnetic resonance imaging; 68Ga DOTA, 68Gallium dodecanetetra-acetic acid; 18FDG, 18F-fluorodeoxyglucose; PET, positron emission tomography, CT, computed tomography; FNAB, fine-needle aspiration biopsy; NEN, neuroendocrine neoplasm; G1/G2, grades 1 and 2; G3, grade 3; 1/-, with or without; TAE/TACE, transarterial embolization/transarterial chemoembolization; SIRT, selective internal radiotherapy; PRRT, peptide receptor radionuclide therapy; CRR, cytoreductive resection; LT, liver transplantation; SSA, somatostatin analogues; P, use in pancreatic neuroendocrine tumors; CLD & SS, combination of liver-directed and systemic strategies.
guide treatment decisions, and most available experience is based primarily on retrospective analyses of case series without controls. Indeed, 2 Cochrane reviews of treatments for NE LMs reported that a lack of robust evidence precluded their analyses.16,17 In the current review, we collate the currently available evidence on treatment of NE LMs and suggest avenues for future progress. Diagnosis
Contrast-enhanced ultrasonography (CEUS), multiphase helical computed tomography (CT) with a multirow detector scanner, and diffusion-weighted (DW) magnetic resonance imaging (MRI) are currently considered the morphologic imaging modalities of choice for the detection of NE LMs. Compared with CEUS, CT, T2-weighted MRI, and dynamic gadolinium-enhanced MRI, DW MRI detects smaller and more liver deposits.18 Meticulous pathologic examination of resected liver specimens revealed that preoperative imaging resulted in the understaging of NE LMs in >50% of patients.19 Ultrasound-guided or 2
CT-guided biopsy is integral for diagnosis confirmation and tumor grading (low grade [grade 1], antigen identified by a monoclonal antibody Ki-67 [Ki-67] index 2%; intermediate grade [grade 2], Ki-67 index 3%-20%; high grade [grade 3], Ki-67 index >20%).20 Tumor grade provides valuable prognostic information and influences treatment trajectories (Fig. 1). NE LMs may be classified morphologically as type I (single metastasis), type II (isolated metastatic bulk accompanied by smaller deposits), or type III (disseminated metastatic spread).21 Somatostatin receptor-based imaging exploits the observation that the majority of NETs express somatostatin receptors, enabling the visualization of primary and metastatic lesions while concurrently probing for potential amenability to receptor-targeted treatment. Other functional imaging techniques target tumor metabolism. An array of tracer molecules may be used, including 111In-octreotide; the 68 Ga-conjugated radiopharmaceuticals, including 0 1 3 dodecanetetra-acetic acid (DOTA) -Phe -Tyr octreotide (DOTATOC), DOTA0-Tyr3 octreotate (DOTATATE), Cancer
Month 00, 2014
Treatment of Neuroendocrine Liver Metastases/Frilling and Clift
DOTA021-Nal3 octreotide (DOTANOC), and 64CuDOTA; 18F-fluorinated levodopa (18F-DOPA); and 11C–5hydroxytryptophan (11C-5-HTP). The more frequently used 68 Ga-DOTATOC positron emission tomography (PET)/ CT imaging identifies LMs with a sensitivity of 82% to 100% and a specificity of 67% to 100% and identifies extrahepatic disease in grade 1 NETs with a sensitivity of 85% to 96% and a specificity of 67% to 90%.22 It has been demonstrated that results from 68Ga-DOTATOC PET/CT imaging modify the treatment decision in approximately 33% of patients.23,24 Higher grade NETs (grade 2-3) are staged better using 18F-fluorodeoxyglucose (FDG)-PET/CT. Currently used imaging response criteria only insufficiently reflect the distinct characteristics of NETs. Although tumor size may remain stable on morphologic imaging, changes in activity demonstrated by functional imaging may indicate treatment efficacy. Although it is generally considered a biomarker for NETs, the peptide chromogranin A (CgA) is not helpful in diagnosing LMs. However, it can be useful for assessing entire tumor burden, longitudinal monitoring, and estimating prognosis. It has been demonstrated that a CgA decrease >80% after resection of NE LMs predicts both a complete symptomatic response and disease stabilization.25 More recently, a panel of novel, promising NET biomarkers, such as a multitranscript molecular signature encompassing 51 marker genes,26 paraneoplastic antigen Ma2 autoantibodies,27 circulating tumor cells,28 and metabolic spectra,29 has been introduced. The clinical utility of such biomarkers will require further validation in larger series. Treatment Liver resection
Hepatectomy represents the mainstay of treatment for patients with grade 1 or 2 NE LMs (Table 1). Because substantial numbers of patients present with type III LMs on initial diagnosis, only approximately 20% to 30% are candidates for resection. Selected patients with primarily unresectable, bilobar metastases may benefit from preoperative right portal vein embolization with consecutive left liver lobe hypertrophy41 or 2-step resections.42,43 Radiation lobectomy with 90Y-radiolabeled microspheres has the potential to attain local tumor control and hypertrophy of the anticipated future liver remnant in patients with large right liver lobe metastases and small-for-size left liver lobes.44 Difficulties in planning complex surgical procedures can be alleviated by using computer-assisted, 3-dimensional CT technology, which provides precise imaging of liver structures and an accurate volumetric calculation.45 A recent systematic review that analyzed results from 29 reports revealed median 1-year, 3-year, 5-year, and Cancer
Month 00, 2014
10-year survival rates of 94% (range, 79%-100%), 83% (range, 63%-100%), 70.5% (range, 31%-100%), and 42% (range, 0%-100%), respectively.12 Resection with macroscopically clear margins (R0 and R1) was feasible in a median of 63% of patients (range, 38%-100%) and 21% of patients (range, 0%-37%), respectively. Complete and partial symptomatic responses were observed in a median of 73% (range, 23%-92%) and 14% (range, 0%74%) of patients, respectively. The perioperative morbidity and mortality rates ranged from 3% to 45% (median, 23%) and from 0% to 9% (median, 0%), respectively. Contrasting these rather encouraging results were dismal 1-year, 3-year, 5-year, and 10-year recurrence-free survival rates of 63% (range, 50%-80%), 32% (range, 24%69%), 29% (range, 6%-66%), and 1% (0%-11%), respectively. These results were in concordance with a multicentric analysis of 339 patients who underwent resection of NE LMs and had overall 5-year and 10-year survival rates of 75% and 51%, respectively, and recurrence-free survival rates at 1 year, 3 years, and 5 years of 56.9%, 24.2%, and 5.9%, respectively.46 Overall, outcomes of hepatic resection for NE LMs require cautious consideration. In some series, patients with extrahepatic metastases are included; and, in others, results from resection as the only technique are pooled with those from ablation alone and incomplete resection in combination with ablative techniques. Inclusion criteria for liver resection are not well reported, and there are no uniform follow-up protocols. Neoadjuvant and adjuvant treatment concepts for down-staging and reduction of disease recurrence have been introduced in single case reports and smaller series; however, there is an urgent need for further expansion and adoption in the standard treatment pathways. Valid assessment of the efficacy of liver resection for NE LMs remains debatable, because no prospective randomized trials to date have compared outcomes of surgery with outcomes of nonsurgical, liver-directed techniques. Mayo et al47 compared patients who underwent liver resection with those who underwent transarterial embolization (TAE). Surgically treated patients had superior median and 5-year survival compared with the TAE cohort (123 months and 74% vs 34 months and 30%, respectively). Although symptomatic patients who had >25% liver involvement benefited from surgery compared with TAE, there was no difference in long-term survival in the asymptomatic groups. The authors concluded that asymptomatic patients with large hepatic burden would be better treated using TAE, because they would not benefit from an invasive surgical approach. Improved 3
Review Article TABLE 1. Outcomes From Liver Resection for Neuroendocrine Liver Metastases—Selected Single-Center Studies Published Since 2000
Reference
No. of Patients Undergoing Resection
No. of R0 Resections
5-Year OS, %
5-Year DFS, %
74 41 27 18 47 170 16 31 20 16 34 26
28 26 23 15 25 75 16 10 20 16 15 13
NR 88 100 86 74
NR 31 96 90 66 76
Saxena 201130 Scigliano 200931 Frilling 200921 Gomez 200732 Elias 200333 Sarmiento 200334 Norton 200335 Nave 200136 Coppa 200137 Yao 200138 Chamberlain 200039 Pascher 200040
82 86 67 70 85 NR
Other Survival Data After R0 Resection Median OS, 73 mo
29
NR
Median survival, 70 mo
30-Day Morbidity and Mortality Mortality, 1.3% (4% at 60 d) Morbidity, 14%; mortality, 0% Morbidity, 7.4%; mortality, 0% Morbidity, 22%; mortality, 5.6% Morbidity, 45%; mortality, 5% Morbidity, 14%; mortality, 1.2% Morbidity, 19%; mortality. 0% Morbidity, 13% NR Morbidity, 12%; mortality, 0% Mortality, 6% Mortality, 0%
Abbreviations: DFS, disease-free survival; NR, not reported; OS, overall survival; R0, resection with negative margins.
survival outcomes of patients who underwent liver resection, compared retrospectively with those who underwent TAE (76% vs 50% at 5 years) and those who underwent medical treatment (83% vs 39% at 3 years), also were reported by Chamberlain et al.39 Liver resection should be considered as the treatment of first choice for patients who are fit to undergo hepatic resection for grade 1 or 2 LMs and have no extrahepatic disease. Those who have extrahepatic lesions also may be considered if their tumor burden is entirely resectable. A complete tumor resection should be anticipated. Debulking liver surgery may be indicated in individual patients with hormonally active tumor burden who are refractory to nonsurgical treatment options. Liver transplantation
Patients with unresectable NE LMs may be considered as potential candidates for liver transplantation. Higher tumor grade, nonportal tumor drainage, extrahepatic metastases (with the exception of resectable perihilar lymph node metastases), and advanced carcinoid heart disease are generally accepted contraindications. Other criteria, including age 55 years and a hepatic tumor load 50%, have been suggested by some centers48; however, their validity was contradicted by others.49 In singlecenter series published since 2000, the 3-year and 5-year overall survival rates ranged from 57% to 100% and from 33% to 90%, respectively, and disease-free survival rates from 11% to 77% at 5 years have been reported50-53 (Table 2). Although short-term results have improved because of better selection of transplantation candidates, refinement of surgical techniques, and the introduction of novel immunosuppressive regimens, the high recur4
rence rate remains a matter of concern. Patient-specific biomarkers are needed for the identification of those who gain a long-term benefit from the procedure. Like patients who are followed after transplantation for hepatocellular carcinoma, down-staging while on the waiting list should become part of the transplantation protocol for NE LMs. Liver transplantation should be offered not as an ultima ratio solution after failure of all other treatment attempts but, rather, as a well timed procedure in the still controllable stage of the disease. A recent retrospective analysis of 213 patients with NET undergoing liver transplantation in European centers between 1982 and 2009 revealed 1-year, 2-year, 3-year, and 5-year overall survival rates of 81%, 73%, 65%, and 52%, respectively, and a disease-free survival rate of 30% at 5 years.57 Hepatomegaly, age >45 years, and concomitant transplantation and resection were identified as unfavorable prognostic factors. It remains a matter of debate whether surgical morbidity rather than tumor biology contributed to the poor outcome in patients who underwent concomitant procedures. The 3-month postoperative mortality rate was 10% and was mainly attributed to surgical complications, including intra-abdominal hemorrhage, hepatic artery thrombosis, portal vein thrombosis, peritonitis, pancreatitis, and primary nonfunction. Among the risk factors for 3-month mortality were early retransplantation, upper abdominal exenteration, splenectomy, operative duration >10 hours, and hepatomegaly. Seventeen patients died while recurrence free because of late morbidity related to transplantation. In total, early or late complications of transplantation accounted for death in 37 recurrence-free patients (17%). The median survival in that cohort was 8 months (range, Cancer
Month 00, 2014
Treatment of Neuroendocrine Liver Metastases/Frilling and Clift
TABLE 2. Liver Transplantation for Neuroendocrine Metastases—Selection of Single-Center Series Published Since 2000 OS, % No. of Patients
Reference Bonaccorsi-Riani 201050 Olausson 200749 Marin 200754 Mazzaferro 200748 van Vilsteren 200655 Frilling 200651 Florman 200452 Cahlin 200356 Rosenau 200253 Coppa 200137 Pascher 200040
9 15a 10 24 19 15 11 7 19 9 4
1-Year
2-Year
3-Year
88
77
86
57
DFS, % 5-Year
10-Year
33 90
1-Year 67 70
33
5-Year
10-Year
11 20
38 90
88 78.3 73
3-Year
77 80 69.4
67.2 36
48.3
80 89 100 100
100 75
100 75
80 70
50
56
21 53
21
Abbreviations: DFS, disease-free survival; OS, overall survival. a These included 5 patients who underwent visceral transplantations.
4-165 months). In US transplantation centers, comparable outcomes for patients who underwent transplantation for hepatocellular carcinoma and those undergoing liver transplantation for NE metastases were reported.58,59 A review of the United Network for Organ Sharing database59 comprised 150 patients who underwent liver transplantation either in isolation or as a component of multivisceral transplantation. Survival rates at 1 year, 3 years, and 5 years for patients who received only a liver graft were comparable to those obtained in patients who underwent multivisceral transplantation (81% vs 80%, 65% vs 65%, and 49% vs 48%, respectively). The disease-free survival rates at 1 year, 3 years, and 5 years were 77%, 50%, and 32%, respectively, in 83 evaluable patients. In the group of patients who underwent transplantation for carcinoids, the patient and graft survival rates were comparable at 1 year, 3 years, and 5 years (76%, 55%, and 47%, respectively, and 71%, 51%, and 44%, respectively). Overall, results from multivisceral transplantations for NE LMs appear to be contradictory. Although promising results were reported for highly selected patients by some groups,49,59,60 others have discouraged this approach.61 Liver transplantation should be offered to strictly selected patients with nonresectable LMs. Grade 3 tumors, nonportal systemic tumor drainage, extrahepatic metastases with the exception of perihilar lymph node metastases, and carcinoid heart disease in general are considered exclusion criteria. Because of the imminent shortage of cadaveric organ donors, consideration of living-donor liver transplantation should be implemented in the evaluation process. Cancer
Month 00, 2014
Locally ablative techniques
In patients with hepatic metastases who are ineligible for complete resection, ablative modalities can be used either as stand-alone therapy (through percutaneous, laparoscopic,62 or open approaches) or as surgical adjuncts.63 Various methods are in clinical use, including radiofrequency ablation,64 microwave ablation,65 laser ablation,66 and cryoablation.67 Although radiofrequency ablation has been used most extensively, microwave ablation provides comparable efficacy but is less time consuming and, with laser application, allows for more accurate, MRI-guided monitoring of treatment. Overall 5-year survival rates range from 37% to 57%, and the most favorable results are reported in patients who have limited hepatic tumor burden, dominant liver lesions measuring 1 cm.65,67,68 In the largest reported series on 89 patients with NE LMs who underwent 119 laparoscopic radiofrequency ablation treatment sessions, patients achieved a median overall survival of 6 years after the first treatment session and a median disease-free survival of 1.3 years. Local recurrence within the liver occurred in 23% of patients, 63% developed new lesions, and extrahepatic disease became evident in 53% at a median follow-up of 3063 months. Amelioration of symptoms was achieved in 97% of patients. The postinterventional morbidity rate in that study was 6%, and the 30-day mortality rate was 1%.68 Locally ablative techniques represent a safe and repeatable palliative treatment for patients who have a limited number of unresectable LMs that preferably measure
