E-Mail karger@karger.com
ENETS Consensus Guidelines
Neuroendocrinology 2016;103:186–194
DOI: 10.1159/000443172
ENETS Consensus Guidelines for High-Grade
Gastroenteropancreatic Neuroendocrine Tumors
and Neuroendocrine Carcinomas
R. Garcia-Carbonero
a
H. Sorbye
b
E. Baudin
c
E. Raymond
d
B. Wiedenmann
e
B. Niederle
f
E. Sedlackova
g
C. Toumpanakis
h
M. Anlauf
i
J.B. Cwikla
j
M. Caplin
k
D. O’Toole
l
A. Perren
m
all other Vienna Consensus Conference participants
a
Medical Oncology Department, Hospital Universitario Doce de Octubre, Madrid , Spain;
b
Department of Oncology,
Haukeland University Hospital, Bergen , Norway;
c
Institut Gustave Roussy, Villejuif , and
d
Oncologie Médicale,
Hôpitaux Universitaires Paris Nord Val de Seine, Paris , France;
e
Department of Hepatology and Gastroenterology,
Campus Virchow Klinikum, Charité Universitätsmedizin Berlin, Berlin , Germany;
f
Department of Surgery, Medical
University of Vienna, Vienna , Austria;
g
Department of Oncology, First Faculty of Medicine and General Teaching
Hospital, Prague , Czech Republic;
h
Neuroendocrine Tumour Unit, Royal Free Hospital, London , UK;
i
Institut für
Pathologie und Zytologie, St. Vincenz Krankenhaus, Limburg , Germany;
j
Department of Radiology, Faculty of
Medical Sciences, University of Warmia and Mazury, Olsztyn , Poland;
k
Neuroendocrine Tumour Unit, Royal Free
Hospital, London , UK;
l
NET Centre, St. Vincent’s University and Department of Clinical Medicine, St. James Hospital
and Trinity College, Dublin , Ireland;
m
Institute of Pathology, University of Bern, Bern , Switzerland
are not a homogenous entity and can be further subclas-
sified into biologically relevant subgroups. A separation
based on the proliferative index (Ki-67 >55%) was shown
to have clinical implications regarding response to che-
motherapy and prognosis: NEC with Ki-67 >55% re-
sponded better to platinum-based chemotherapy and,
nevertheless, had a 4 months’ shorter median survival
than G3 NEN in the lower proliferative range (20–55%)
[1] . More recent publications show that morphological
differentiation and Ki-67 are able to separate prognostic
groups among G3 cases, and therefore a separation of
well-differentiated G3 NET from poorly differentiated
G3 NEC is emerging
[2–4] . The exact criteria need to be
defined both on the morphological and on the molecular
level. The spectrum of mutations of well-differentiated
Introduction
Neuroendocrine carcinomas (NEC) are rare in the
gastrointestinal (GI) tract, whereas they are frequent in
the form of small cell carcinoma (SCLC) in the lung.
Therefore, most of the suggested guidelines arise from
analogy to SCLC. As for other extrapulmonary primary
tumor locations, published data on NEC of the GI tract
are scarce. This guideline encompasses all WHO grade 3
(G3) gastroenteropancreatic (GEP) neoplasms; however,
in the future, G3 neuroendocrine neoplasms (NEN) will
probably be separated according to differentiation, as ex-
plained below, and potentially according to organ of ori-
gin, such as for well-differentiated neuroendocrine tu-
mors (NET) G1/G2.
According to the WHO classification of 2010, NEC are
defined as poorly differentiated NEN with Ki-67 >20%
and hence G3. Increasing evidence suggests that G3 NEN
Published online: January 5, 2016
Dr. Rocio Garcia-Carbonero
Medical Oncology Department
Hospital Universitario Doce de Octubre
Avenida de Cordoba km 5.4, ES–28041 Madrid (Spain)
E-Mail rgcarbonero
@ gmail.com
© 2016 S. Karger AG, Basel
0028–3835/16/1032–0185$39.50/0
www.karger.com/nen
For an alphabetical list of all other Vienna Consensus Conference par-
ticipants, see Appendix.
Downloaded by:
Universitätsbibliothek Bern
130.92.9.56 - 12/21/2016 4:11:17 PM
High-Grade GEP Neuroendocrine
Tumors and NEC
Neuroendocrinology 2016;103:186–194
DOI: 10.1159/000443172
187
pancreatic NET is different from that of pancreatic NEC
[5] , suggesting different ways of tumorigenesis. However,
to date there are no solid data that adequately address the
implications of these observations in terms of treatment
effect of the different available regimens.
Epidemiology
The GEP tract is the most common site of extrapulmo-
nary NEC, accounting for 35–55% of all NEC originating
from the lung. Only about 5% of all GI NEN have Ki-67
>20%
[6, 7] . This frequency might differ by organ, with
about 7% in the pancreas
[7] and up to 40% in the colon
[8] . GEP NEC are therefore very rare neoplasms repre-
senting <1% of all GI malignancies. Up to 85% have me-
tastases at the time of diagnosis (65% distant)
[1, 6] . Me-
tastases are most frequently found in the liver (70%) fol-
lowed by the lung (15%), bone (15%) and brain (4%)
[1] .
No gender difference has been identified. The mean age
at diagnosis is 60 years
[1] .
Clinicopathological Features
As the great majority of these tumors are not associ-
ated with a hormonal syndrome (<5%), and more than
two thirds of all patients present with advanced disease,
clinical presentation is dominated by tumor-derived site-
specific symptoms and the constitutional syndrome char-
acteristics of advanced cancer (anorexia, weight loss and
fatigue). Depending upon tumor location, a wide variety
of symptoms may occur. The neuroendocrine nature of
these tumors is generally not suspected from the clinical
presentation, although as in SCLC paraneoplastic syn-
dromes may occur in some patients (i.e., Cushing or In-
appropriate ADH Secretion Syndromes). A detailed an-
amnesis and physical examination are fundamental to ap-
propriately guide diagnostic procedures.
According to the WHO 2010 classification, NEC are
poorly differentiated highly aggressive neoplasms, some-
times with organoid features, marked nuclear atypia and
multifocal necrosis
[9] . A diffuse expression of neuroen-
docrine markers (diffuse for synaptophysin, focal for
chromogranin A, and the latter may be absent) separates
the entity pathologically from poorly differentiated carci-
noma.
The grading introduced by ENETS in 2006
[10] of
NEC is by definition G3, either based on a proliferation
index >20% or >20 mitoses in 10/HPF. This proposition
has been adopted by the WHO classification and was
shown repeatedly to be clinically applicable in predicting
a very aggressive subset of NEN
[7] .
Prognosis and Survival
Survival is poor in NEC, ranging from 38 months for
patients with localized disease to 5 months in the meta-
static setting according to the SEER population registry
data, which involved 2,546 patients diagnosed with GI
NEC from 1973 to 2012 in the USA
[11] . Median surviv-
al in the metastatic setting may be as short as 1 month for
patients receiving only best supportive care, up to 12–19
months for those treated with best available therapy
[1,
12]
. Only 5% of all patients are long-time survivors [7] .
Progression-free survival after cisplatin-based chemo-
therapy and overall survival differs according to the loca-
tion of the primary tumor, with poorer reported out-
comes in esophageal, colonic and rectal NEC compared
to gastric and pancreatic ones in some large European
series
[1] . In contrast, survival of pancreatic NEC was
poorer in Japanese patients
[13] . A poor performance sta-
tus, high proliferation rate, elevated baseline lactate dehy-
drogenase (LDH) and thrombocytosis are other factors
that have also been associated with a worse prognosis.
Diagnostic Procedures
Biochemical Tests
Plasma chromogranin A may be elevated in up to two
thirds of patients with advanced NEC
[1] , although the
levels are generally lower than those observed in well-dif-
ferentiated tumors
[14, 15] . In contrast, the levels of oth-
er tumor markers such as neuron-specific enolase (NSE)
are higher in poorly differentiated tumors than in NET,
and they are significantly associated with survival. How-
ever, the role of circulating tumor markers to predict and
monitor outcome has not been properly assessed in ex-
trapulmonary NEC. Screening for other hormonal mark-
ers is not justified unless clinically indicated.
Endoscopic and Imaging Procedures
Endoscopic examination of the primary tumor site is
recommended, which is also useful to obtain a biopsy for
histological diagnosis. If this is not feasible, endoscopic
ultrasound-guided or percutaneous procedures can be
useful. Once the histological diagnosis of a G3 NEC has
been confirmed, complete staging using whole-body CT
Downloaded by:
Universitätsbibliothek Bern
130.92.9.56 - 12/21/2016 4:11:17 PM
Garcia-Carbonero etal.
Neuroendocrinology 2016;103:186–194
DOI: 10.1159/000443172
188
scan or MRI should be performed to assess the extent of
disease and to design the most appropriate therapeutic
strategy. A lung primary should be reasonably excluded
(negative imaging studies of the lung). FDG-PET may be
useful if radical surgery is being pursued or if clarification
of equivocal findings on conventional imaging may
change the therapeutic approach. Radiolabeled soma-
tostatin analogue scans are not routinely recommended,
as poorly differentiated tumors generally do not express
somatostatin receptors. However, data from large series
indicate positive SRI findings in a substantial proportion
of patients with certain primary tumors (up to 45% of
pancreatic NEC), particularly those with proliferative in-
dexes in the low range of G3, and may differ by histolog-
ical subtype (45% of small cell vs. 32% of large cell NEC)
[1] . In the absence of neurological symptoms, brain CT
or MRI are not recommended, as the incidence of brain
metastasis in extrapulmonary NEC is rather low (<5%)
[1] . Bone scans are neither indicated if there is no clinical
or biochemical suspicion of bone metastasis. In the pres-
ence of elevated LDH, peripheral blood leukoerythro-
blastosis or thrombocytopenia, a bone marrow biopsy
may be considered.
Minimal Consensus Statement
Clinical signs and symptoms should guide the appropriate
diagnostic procedures (summarized in fig. 1). Chromogranin A
and NSE testing is not mandatory, although they may be useful
if elevated at diagnosis. A proper assessment of their utility in
extrapulmonary NEC is, however, pending. Other hormone tests
are not routinely recommended.
A minimal diagnostic workup should include site-specific
endoscopic assessment with tumor biopsy, and whole-body CT
scan (and/or MRI) for tumor staging. In patients with metastatic
disease, an ultrasound-guided percutaneous biopsy may be per-
formed if feasible. Somatostatin receptor scintigraphy is not rou-
tinely indicated but may be considered in tumors with prolifera-
tive indexes in the low range of G3 (Ki-67 <55%). Bone scans or
brain imaging (CT or MRI) should not be performed in the ab-
sence of site-specific symptoms. FDG-PET may be considered in
patients in whom radical surgery is being pursued or if clarifica-
tion of equivocal findings on conventional imaging may change
the therapeutic approach. FDG-PET may be useful in resectable
cases for whole body assessment.
Histopathology and Genetics of Poorly
Differentiated NEC
Histopathologically, NEC show a neuroendocrine
phenotype by immunohistochemistry, in large cell NEC
a positivity for synaptophysin is mandatory, chromo-
granin A staining is variable and may be weak or absent.
Rarely, both markers may be negative in small cell NEC
(<5%
[1] ). Other neuroendocrine markers such as NSE or
CD56 are less specific and must be used with caution. Ki-
67 is by definition >20%
[10] and in half of the cases, it is
>55%
[1] . Punctate or geographic necrosis is frequent.
Reporting of the immunohistochemical results above as
well as the proliferative index by mitosis is essential. So-
matostatin receptor 2A (Sstr 2a) immunohistochemistry
is optional
[16, 17] . Over 90% of G3 NEC do not produce
hormones
[17] .
In the setting of a carcinoma of unknown primary, the
expression of transcription factors such as Ttf1, Cdx-1 or
Isl1 cannot be used to help localize the site of the primary
tumor
[18] .
Care must be taken to differentiate NEC from poorly
differentiated adenocarcinoma, especially in certain or-
gans such as the pancreas, where a differential diagnosis
with acinic cell carcinoma may be particularly challeng-
ing
[19] . NEC are separated into large cell and small cell
types; however, no clear clinicopathological differences
between the two types have been shown for the pancreas
[19] .
Pancreatic NEC show a genetic profile different from
NET with frequent mutations in p53 and RB
[5] and a
much higher mutation rate (in review), similar to pulmo-
nary small cell carcinoma. Furthermore, up to 40% of all
NEC present a minor component of adenocarcinoma
(colon
[20] , stomach [21] ) or squamous cell carcinoma
(esophagus, anus). If the non-endocrine component ex-
ceeds 30%, the neoplasm is classified as mixed adenoneu-
roendocrine carcinoma. Differentiation together with
proliferation and mutation spectrum will be important in
discriminating G3 NET from G3 NEC in the future
[2–4] .
Minimal Consensus Statement
A routine pathological report should include morphology
(large cell vs. small cell and differentiation), staining for chromo-
granin A and synaptophysin and Ki-67 estimate or/and mitotic
count.
Treatment
Evidence to support treatment recommendations for
GEP G3 NEC is scarce and derives from limited retro-
spective series and very few small non-controlled clini-
cal trials. Most investigators, therefore, treat this entity
in analogy to the much more common SCLC due to
Downloaded by:
Universitätsbibliothek Bern
130.92.9.56 - 12/21/2016 4:11:17 PM
High-Grade GEP Neuroendocrine
Tumors and NEC
Neuroendocrinology 2016;103:186–194
DOI: 10.1159/000443172
189
their histological and clinical resemblance. Bearing
these caveats in mind, guidance is hereby provided
(fig.2). Nevertheless, generating prospective and pref-
erably controlled data is greatly needed and encouraged
in this setting.
Surgery
Curative surgery is usually attempted in localized dis-
ease, although retrospective series indicate that it is rare-
ly curative as a sole therapeutic modality
[22] . Given the
high relapse rate observed following radical surgery, most
clinicians would advocate platinum-based adjuvant ther-
apy in this setting. Data reported by Casas et al. of a large
series of esophageal small cell carcinomas support this
approach
[23] . In this study, survival was 20 months for
patients who received systemic chemotherapy in addition
to local treatment versus only 5 months for those who
were treated with local therapy only, and the type of treat-
ment was found to be an independent prognostic factor
in multivariate analysis. Some authors propose neoadju-
vant chemotherapy followed by definitive surgery, al-
though data to support this approach are scarce
[24] . In
patients with important comorbidities or where the tu-
mor’s anatomical site makes surgical resection not advis-
able due to high morbidity (i.e., esophagus), a definitive
course of radiotherapy and chemotherapy is a reasonable
treatment strategy.
In the context of advanced metastatic disease, debulk-
ing or cytoreductive surgery and surgical resection of me-
tastasis are not recommended. Other ablative strategies of
liver metastasis (i.e., radiofrequency ablation, TACE) are
also discouraged.
Medical Therapy
Chemotherapy is an essential part of the multimodal-
ity approach for localized NEC and the mainstay of care
in advanced disease. Survival of patients with metastatic
NEC treated with chemotherapy varies widely (from 7 to
19 months) but shows a substantial improvement over
that reported for patients that receive only best support-
ive care (1 month). No randomized studies, however,
have properly addressed the magnitude of this effect so
far. Swift referral for consideration of palliative chemo-
therapy is recommended as performance status deterio-
ration may occur rapidly and preclude further therapy.
Based on their established role in metastatic SCLC, cis-
platin and etoposide (EP) have been one of the most
widely used regimens in GEP NEC ( table1 )
[1, 13, 25–
29]
, with response rates in the largest most recent series
of ∼ 30% and median survival of around one year. In one
of the largest series published to date, Sørbye et al.
[1] ob-
served that Ki-67 was significantly associated with re-
sponse to chemotherapy. Indeed, patients with Ki-67 val-
ues >55% had a greater response rate (42 vs. 15%) al-
Recommended
- CT or MRI
(chest/abdomen/pelvis)
Only if clinically indicated
- FDG-PET – if surgery is
being considered
- Additional according to
symptoms
Initial diagnosis
NEC and G3 NET
Liver and kidney function
Other biochemical
markers if clinically
indicated
Synaptophysin,
chromogranin A
Ki-67
SSTR 2a (optional)
Resected disease
- Contrast CT scan every
3–6 months for 2–3 years,
and then every 6–12
months
Advanced disease
- Contrast CT scan every
2–3 months if on therapy
Follow-up
Biomarkers if elevated at
diagnosis
Other tests only if
clinically indicated
Imaging
Laboratory
Pathology
Fig. 1. Diagnostic algorithm for NEC and
G3 NET.
Downloaded by:
Universitätsbibliothek Bern
130.92.9.56 - 12/21/2016 4:11:17 PM
Garcia-Carbonero etal.
Neuroendocrinology 2016;103:186–194
DOI: 10.1159/000443172
190
though poorer survival (10 vs. 14 months) than patients
with Ki-67 values <55%. Other negative prognostic fac-
tors in this study were a poor performance status, prima-
ry colorectal tumors and elevated platelets or LDH levels,
which were all associated with decreased survival.
Alternative regimens substituting carboplatin for cis-
platin, or irinotecan for EP, have been validated in SCLC
and seem at least equivalent in terms of efficacy in lim-
ited series of GEP NEC ( table1 )
[13, 30–35] , with differ-
ent toxicity profiles. In the context of advanced stage
SCLC, a randomized study conducted in Japan demon-
strated that the combination of irinotecan and cisplatin
(IP) was associated with improved overall survival as
compared to the standard cisplatin and EP combination
[36] . Two subsequent randomized Western trials, how-
ever, failed to confirm this superiority. Both regimens
produced comparable efficacy, with less hematological
and greater GI toxicity with the IP combination (particu-
larly diarrhea and vomiting)
[37, 38] . Consistent with
these findings, large retrospective data of systemic che-
motherapy for advanced GEP NEC from 23 Japanese in-
stitutions documented that the IP regimen was associ-
ated with greater response rates (50 vs. 28%) and sur vival
(13 vs. 7 months) than the EP regimen, and this differ-
ence was more remarkable in hepatobiliopancreatic
NEC. Prognostic factors in this study included primary
tumor site (those with hepatobiliopancreatic primaries
having the worst prognosis) and elevated baseline LDH
levels, whereas treatment schedule was not an indepen-
dent predictive factor for survival. Three-drug regimens
such as cisplatin, EP and paclitaxel do not seem to sub-
stantially improve efficacy and are significantly more
toxic
[39] .
Evidence for salvage therapy in patients progressing
on first-line platinum-based regimens is very limited
( table1 )
[1, 13, 40–45] . Overall, response rates are lower
(18% in the NORDIC NEC study), although small series
have documented response rates of 23–40% with oxali-
platin-based (XELOX, FOLFOX) or irinotecan-based
(FOLFIRI, IP) regimens. Welin et al. reported a 33% re-
sponse rate with temozolomide, alone or in combina-
tion with capecitabine and bevacizumab, in a cohort of
25 patients with poorly differentiated NEC (17 of GEP
origin)
[44] . A Ki-67 index <60% was predictive for re-
sponse to treatment and survival. In contrast, no re-
sponses were observed in another series of 28 NEC cas-
es treated with temozolomide monotherapy
[45] . Re-
treatment with platinum/EP may also be considered in
patients that achieve good durable responses upfront
and progress after a treatment break of at least 3 months,
provided no cumulative toxicity (i.e., neurotoxicity, oto-
toxicity) precludes further treatment with platinum
Chemotherapy
Radiotherapy may be considered in
selected sites (bone, brain, ...) for
symptom control
Special considerations for well-differentiated G3
ض Platinum-based regimens seem to be less
active than in poorly differentiated tumors
ض Other regimens may be considered (i.e.
TMZ-based)
ض PRRT may be considered in SSTR-positive
tumors
Chemotherapy; radiotherapy
ض Cisplatin or carboplatin and etoposide are generally recommended in the adjuvant setting or as
first-line therapy in advanced disease. Alternative regimens substituting irinotecan for etoposide
may also be employed as first-line therapy in advanced disease.
ض Irinotecan or oxaliplatin-based regimens may be considered as second-line therapy.
Clinical judgement should be used.
Distant metastases
Radical surgery
+ adjuvant
chemotherapy
± radiotherapy
Resectable
Locoregional
radiotherapy
+ chemotherapy
Locoregional unresectable
NEC and G3 NET
Fig. 2. Therapeutic algorithm for NEC and G3 NET.
Downloaded by:
Universitätsbibliothek Bern
130.92.9.56 - 12/21/2016 4:11:17 PM
