Key words
MEN1 - mortality - pancreas - neuroendocrine tumors
Introduction
Multiple endocrine neoplasia type 1 (MEN1) is an autosomal dominant disorder caused
by pathogenic germline variants in the MEN1 gene (11q13) encoding the menin
protein [1]. The prevalence of MEN1 ranges between
1 and 10 in 100 000 people in the general population [2]. The syndrome is characterized by the occurrence of primary
hyperparathyroidism (PHPT) and pancreatic islet- and anterior pituitary tumors [3]. The most common manifestation of MEN1,
occurring in almost all patients, is PHPT and is usually the earliest manifestation
in patients, presenting before the 6th decade of life [4].
The reported prevalence for entero-pancreatic neuroendocrine tumors (NET) in MEN1
ranges between 30–80% [5], while
the rate of MEN1 among patients with pancreatic NET (PNET) varies depending on the
functional status of the tumor, ranging between 3–15% in
nonfunctioning PNET, to 20–25% in gastrin secreting PNETs [2]. The majority of MEN1-related PNET are
gastrinomas, insulinomas, and nonfunctioning tumors [5]. Other less frequently detected PNET types in MEN1 include
glucagonomas, vasoactive intestinal peptide secreting tumors, and somatostatinomas
[5]. The reported prevalence of pituitary
tumors among patients with MEN1 varies between 10 and 60%, and it may be the
first clinical manifestation in approximately 25% of de-novo cases [6]. Microscopically, MEN1-related PNETs occur in
80–100% of patients with MEN1. Hence they are often multiple,
located throughout the pancreas but become symptomatic only in a minority of
patients (0–13%)[7]. Patients with
MEN1 are prone to develop other neoplasms, including adrenal tumors, thyroid
adenomas, thymic, gastric, and/or bronchial NET, brain meningiomas and
ependymomas, skin lesions and smooth-muscle tumors, and also breast cancer [8]
[9].
Diagnosis with MEN1 may be defined based on one of the following criteria: Clinical
diagnosis, in the presence of two out of the three main MEN1-related manifestations
(primary hyperparathyroidism, anterior pituitary tumor, and/or
entero-pancreatic NET); Familial diagnosis, in the presence of one MEN1-related
manifestation when at least one first-degree relative is also diagnosed with MEN1,
or Genetic diagnosis, when a germline pathogenic variant is detected in the
MEN1 gene [5].
Patients with MEN1 have shorter longevity compared with the general population [10]. In the past, Zollinger–Ellison
syndrome (ZES) was the major cause of death in MEN1, but proton-pump inhibitors
practically eliminated the catastrophic manifestations of ZES, such as perforated
ulcers and massive upper gastrointestinal bleeding. This led to MEN1-related PNET
becoming the main cause of death in this syndrome [11]
[12]. Yet, whether MEN1-related
PNETs confer a higher mortality risk than their sporadic counterparts is still to be
determined.
In the current work, we compared the mortality risk between sporadic and MEN1-related
PNET and aimed to identify risk factors for mortality among patients with
MEN1-related PNET to allow clinical identification of patients that may benefit from
a more aggressive and early intervention.
Materials and Methods
This was a retrospective, population-based, case-control study based on the
Surveillance, Epidemiology and End Results (SEER) database. The SEER database,
maintained by the National Cancer Institute, provides cancer incidence and survival
data from population-based registries, includes more than 3 million cases, and
accounting for approximately 30% of the United States population. In the
current study, we analyzed the November 2018 update providing information for cases
diagnosed between 2000 and 2016. Data retrieved for the current analysis included
patient demographics (age at diagnosis, gender, and ethnicity), tumor
characteristics (stage, grade, and surgical intervention), and outcomes (survival
since diagnosis and cause of death). The SEER database is publicly available, open
for analysis by the scientific community, and patients’ data are anonymized.
Hence, institutional ethical approval was not required.
Cohort construction
The current analysis included patients documented in the SEER database, diagnosed
with either pituitary adenomas or PNET of any type, either functional or
non-functional. The diagnoses were sorted and identified based on the ICD-O-3
[13] histologic codes (Table 1S) and based
on the NET anatomic site. Since diagnosis with pituitary adenoma and PNET
defines MEN1, we identified patients with both diagnoses as those harboring
MEN1-related PNET – the control group comprised of patients with PNET
alone ('Sporadic PNET').
Statistical analysis
The analysis was performed on R Studio version 1.2.5001. Tumor grades were
classified according to the SEER database definition, as the current definitions
by the WHO for pancreatic neuroendocrine tumors grading have not been used
during most of the study period and are not documented. Staging is reported
based on the American Joint Cancer Classification (AJCC) 6th Edition [14] and according to the SEER staging
definitions (localized, regional metastases, and distant metastases). Survival
analysis was based on the SEER data and on the time since diagnosis until death
or last follow-up. Continuous variables were compared using the
Student’s t-test, and categorical parameters were compared using
the chi-square test. Continuous variables are presented as mean±standard
deviation (SD) unless stated otherwise, and categorical parameters are presented
as n (%). Change in management efficacy over the study period was
adjusted by defining the earliest year of diagnosis per group, and the latter
year was set as a threshold for inclusion for both groups. Survival analysis was
performed using the Kaplan–Meier method and compared by the log-rank
test. Multivariable analysis was done using the proportional Cox regression
analysis, controlling for covariates. A two-tailed p-value
< 0.05 was considered statistically significant.
Results
The study cohort comprised of 569 PNETs cases (56.6% male), 27 of them
(4.7%) have MEN1 based on clinical criteria. The mean age at diagnosis was
61.0±13.2 years with a median of 62.0 [range 18.0, 89.0]. PNET were located
at the head (27.2%), body (13.5%), or tail of the pancreas
(31.5%). Pancreatic-head NET were characterized by a higher grade
(33.5% well-differentiated in the pancreatic head vs. 49.2% in
body/tail PNET, p < 0.001) and stage (33.5%
localized in pancreatic head NET vs. 50.8% in body/tail PNET,
p=0.001), and surgical intervention were performed less often for pancreatic
head versus body/tail PNET (45.8 vs. 67.6%, p
< 0.001, respectively). Considering the higher rate of PNET
multiplicity on MEN1, we compared patients with a single record of PNET to those
with multiple PNET records. When comparing multiple vs. single events, neither age
at diagnosis, gender, stage, grade or surgical treatment were significantly
different. However, a higher fraction of patients with multiple PNETs were of
hispanic ancestry (28.6%) compared to their representation among those with
a single PNET (8.9%, p=0.009). The cohort characteristics and
comparison between patients with sporadic vs. MEN1-related PNET are detailed in
[Table 1].
Table 1 Demographic and tumor characteristics of entire
cohort.
|
Overall
|
Sporadic
|
MEN1
|
p
|
|
n
|
569
|
542
|
27
|
|
|
MEN1 n (%)
|
27 (4.7)
|
NA
|
NA
|
|
|
Age (years)
|
|
|
|
|
|
Mean (SD)
|
61.00 (13.24)
|
61.59 (12.78)
|
49.22 (16.72)
|
<0.001
|
|
Median (range)
|
62.00 (18.0, 89.0)
|
62.50 (22.0, 89.0)
|
54.0 (18.0, 72.0)
|
<0.001
|
|
Male sex n (%)
|
322 (56.6)
|
305 (56.3)
|
17 (63.0)
|
|
|
Ethnicity n (%)
|
|
|
|
|
|
Hispanic
|
55 (9.7)
|
51 (9.4)
|
4 (14.8)
|
0.627
|
|
Caucasian
|
416 (73.1)
|
400 (73.8)
|
16 (59.3)
|
0.028
|
|
African American
|
59 (10.4)
|
53 (9.8)
|
6 (22.2)
|
0.553
|
|
Asian or Pacific Islander
|
35 (6.2)
|
35 (6.5)
|
0 (0.0)
|
0.15
|
|
Other
|
4 (0.7)
|
3 (0.6)
|
1 (3.7)
|
0.081
|
|
Location in pancreas n (%)
|
|
|
|
0.812
|
|
Head
|
155 (27.2)
|
147 (27.1)
|
8 (29.6)
|
|
|
Body
|
77 (13.5)
|
75 (13.8)
|
2 (7.4)
|
|
|
Tail
|
179 (31.5)
|
170 (31.4)
|
9 (33.3)
|
|
|
Other parts
|
28 (4.9)
|
28 (5.2)
|
0 (0.0)
|
|
|
Overlapping lesions
|
42 (7.4)
|
39 (7.2)
|
3 (11.1)
|
|
|
Unknown
|
88 (15.5)
|
83 (15.3)
|
5 (18.5)
|
|
|
Grade n (%)
|
|
|
|
0.605
|
|
I (Well differentiated)
|
231 (40.6)
|
217 (40.0)
|
14 (51.9)
|
|
|
II (Moderately differentiated)
|
66 (11.6)
|
62 (11.4)
|
4 (14.8)
|
|
|
III (Poorly differentiated)
|
23 (4.0)
|
22 (4.1)
|
1 (3.7)
|
|
|
IV (Undifferentiated anaplastic)
|
12 (2.1)
|
12 (2.2)
|
0 (0.0)
|
|
|
Unknown
|
237 (41.7)
|
229 (42.3)
|
8 (29.6)
|
|
|
Stage n (%)
|
|
|
|
0.257
|
|
Localized
|
235 (41.3)
|
221 (40.8)
|
14 (51.9)
|
|
|
Regional
|
121 (21.3)
|
114 (21.0)
|
7 (25.9)
|
|
|
Distant
|
184 (32.3)
|
180 (33.2)
|
4 (14.8)
|
|
|
Unknown
|
29 ( 5.1)
|
27 (5.0)
|
2 (7.4)
|
|
|
Stage AJCC n (%)
|
|
|
|
0.558
|
|
I
|
112 (19.7)
|
106 (19.6)
|
6 (22.2)
|
|
|
II
|
260 (45.7)
|
246 (45.4)
|
14 (51.9)
|
|
|
III
|
70 (12.3)
|
66 (12.2)
|
4 (14.8)
|
|
|
IV
|
127 (22.3)
|
124 (22.9)
|
3 (11.1)
|
|
|
Single event n (%)
|
548 (96.3)
|
522 (96.3)
|
26 (96.3)
|
1
|
|
Surgery performed n (%)
|
327 (57.5)
|
310 (57.2)
|
17 (63.0)
|
0.695
|
MEN1: Multiple endocrine neoplasia type 1. NA: Not applicable.
Univariate analysis – MEN1-related versus sporadic PNETs
Age at diagnosis with MEN1-related PNET was significantly younger than with
sporadic PNET (49.2±16.7 vs. 61.6±12.8 years, respectively, p
< 0.001). PNET pancreatic location, grade, and stage were
comparable between the MEN1 and sporadic groups ([Table 1]). In terms of ethnic composition, a trend toward more
patients of African American ancestry was found among patients with MEN1
(22.2%) versus sporadic PNET (9.8%, p=0.08).
Nevertheless, when comparing patients of African American ancestry to other
patients, other patient and tumor characteristics were comparable.
All-cause mortality
We did not find a significantly different risk for all-cause mortality (ACM) in
our analysis. However, a trend for lower ACM was found for patients with
MEN1-related vs. sporadic PNET (Log-rank test, p=0.09, [Fig. 1]) and especially among those with
nonfunctioning PNET (p=0.07). A similar trend was found among patients
with advanced disease, both based on the SEER staging system and based on AJCC
staging, with lower ACM risk found in patients with MEN1-related versus sporadic
PNET (p=0.08 for both comparisons, [Fig.
2]). Comparison between patients with MEN1-related and sporadic PNET,
demonstrated a trend for lower ACM in those that did not undergo surgical
intervention (p=0.08, [Fig. 3]).
Fig. 1 Kaplan–Meier analysis, comparing all-cause
mortality risk between patients with MEN1-related and sporadic
pancreatic neuroendocrine tumors. MEN1: Multiple endocrine neoplasia
type 1.
Fig. 2 Kaplan–Meier analysis, comparing all-cause
mortality risk between patients with MEN1-related and sporadic
pancreatic neuroendocrine tumors – stratified by disease stage
(localized disease – left, advanced disease – right).
MEN1: Multiple endocrine neoplasia type 1.
Fig. 3 Kaplan–Meier analysis, comparing all-cause
mortality risk between patients with MEN1-related and sporadic
pancreatic neuroendocrine tumors – stratified to patients in
which surgical intervention was performed (right) or not (left). MEN1:
Multiple endocrine neoplasia type 1.
Multivariable analysis
We performed a multivariable Cox-regression analysis, adjusting for patient and
tumor characteristics. In the adjusted survival analysis, age at diagnosis,
tumor grade, and stage were all independently associated with increased risk for
all-cause mortality ([Fig. 4]). Again, a
trend for lower ACM was found, even after adjusting for the covariates, in
MEN1-related versus sporadic PNET (Hazard ratio 0.351, 95% confidence
interval 0.105–1.17, p=0.07).
Fig. 4 Multivariable analysis for all-cause mortality in patients
with sporadic and MEN1-related pancreatic neuroendocrine tumors. MEN1:
Multiple endocrine neoplasia type 1.
Discussion
In this study, we aimed to assess tumor and patient-related factors associated with
increased risk for all-cause mortality in patients with sporadic and MEN1-related
PNETs based on the SEER database. Our analysis demonstrated that patients with
MEN1-related PNET are diagnoses at a younger age, as expected. The MEN1 group
included a higher percentage of patients of African American ancestry compared with
the sporadic PNET group. Although our comparative survival analysis did not reach
statistical significance, the MEN1 group had a consistent trend for lower all-cause
mortality risk across most analyses, including patients with advanced disease (AJCC
stage III/IV). Importantly, among patients that did not undergo surgical resection
of their tumor, those with MEN1-related PNET had better outcome compared to those
with sporadic PNET. The trend for lower mortality risk in MEN1-related PNET was also
found in the multivariable analysis.
Our finding that patients with MEN1-related PNET were diagnosed at a younger age
compared to those with sporadic PNET is expected as has been previously reported
[7]
[15]
[16] and provides support for our cohort validity.
Similarly, the increased risk for ACM by higher grade and stage in the multivariable
analysis supports our study results.
We found that among patients that were followed expectantly (with no surgical
intervention), those with MEN1-related PNET had lower ACM rates than sporadic PNET.
A possible explanation is the low risk of small MEN1-related PNET to advance over
time, as shown in previous studies of the Dutch and French MEN1 cohorts [17]. Our results support the conservative,
non-invasive approach in patients with MEN1-related NET in the lack of clear
benefit. Previously, the French research group of MEN1 showed that MEN1 patients
with small NF-PNETS who undergone pancreatic surgery did not differ significantly in
tumor progression compared to conservative treatment, and therefore suggested that
surgery is not beneficial in this population [18].
Our study demonstrates once more that NETs in the context of hereditary syndromes are
distinct from their sporadic counterparts. The genetic predisposition leads to
clinical differences that should be translated into syndrome-specific clinical
management. Specifically, we demonstrated that in patients with MEN1-related PNET,
the risk for all-cause mortality may be lower compared to patients with sporadic
PNET. Our analyses did not reach statistical significance, explained by the small
sample size, but were consistent over several subgroup analyses and in the
multivariable model. Hence, we advise choosing a cautious approach when considering
surgical intervention in patients with MEN1-related PNET.
This study is based on one of the largest cancer databases available. The SEER
database is a population-based cohort, representing the US population in an unbiased
way. The data available enabled us to perform a robust multivariable analysis and to
define the differences in the risk for mortality between MEN1-related and sporadic
PNET. Our study had several unavoidable limitations that should be considered when
interpreting our results. First, this study was a retrospective study, with all the
inherent limitations and biases of this methodology. Second, our study may
inaccurately represent the general MEN1 population for three reasons: 1. Small
sample size, 2. Possible inclusion of MEN1 phenocopies manifesting with PA and PNET,
such as MEN4, and 3. Defining MEN1 patients based on PNET and PA co-occurrence and
not based on parathyroid adenomas or family history, which are not reported in the
SEER database. Third, since MEN1 genotype is not reported in the SEER
database, the possible association between MEN1 mutation status and life
expectancy could not be assessed [19]. Finally, we
could not adjust for other comorbidities that may affect mortality risk.
In conclusion, based on a large cancer database, and while considering the above
mentioned limitations we provide further validation to the less aggressive behavior
of MEN1-related PNET, that while did not reach statistical significance, support
previous studies, and may reflect on our clinical approach with patients suffering
from this unique and rare syndrome.
