See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/282770072

The value of a rapid contrast-enhanced angio-MRI protocol in the detection of

head and neck paragangliomas in SDHx mutations carriers: a retrospective
study on behalf of the PGL.EV...

Article  in  European Radiology · October 2015

DOI: 10.1007/s00330-015-4024-5

CITATIONS READS

12 77

17 authors, including:

Francoise Borson-Chazot Pascal Rousset

Hospices Civils de Lyon (Centre Hospitalier Universitaire de Lyon) HCL - Centre Hospitalier Lyon Sud - Université Claude Bernard Lyon 1
385 PUBLICATIONS   5,711 CITATIONS    111 PUBLICATIONS   601 CITATIONS   

SEE PROFILE SEE PROFILE

Aurore Caumont-Prim Anne Hernigou

Hôpital Européen Georges-Pompidou (Hôpitaux Universitaires Paris-Ouest) Hôpital Européen Georges-Pompidou (Hôpitaux Universitaires Paris-Ouest)
14 PUBLICATIONS   260 CITATIONS    159 PUBLICATIONS   1,296 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Pleuro-pulmonary diseases in women View project

Thyroid cancer in France View project

All content following this page was uploaded by Pascal Rousset on 14 February 2016.

The user has requested enhancement of the downloaded file.

Eur Radiol
DOI 10.1007/s00330-015-4024-5

HEAD AND NECK

The value of a rapid contrast-enhanced angio-MRI

protocol in the detection of head and neck paragangliomas
in SDHx mutations carriers: a retrospective study on behalf
of the PGL.EVA investigators*
Guillaume Gravel 1 & Patricia Niccoli 2 & Vincent Rohmer 3,4 & Guy Moulin 5 &
Françoise Borson-Chazot 6,7,8 & Pascal Rousset 9 & Anne Pasco-Papon 10 &
Claude Marcus 11 & Frédérique Dubrulle 12 & Hervé Gouya 13 & François Bidault 14 &
Benoit Dupas 15 & Jean Gabrillargues 16 & Aurore Caumont-Prim 20,21 & Anne Hernigou 1 &
Anne-Paule Gimenez-Roqueplo 17,18,19 & Philippe Halimi 1,19

Received: 16 April 2015 / Revised: 21 July 2015 / Accepted: 9 September 2015

# European Society of Radiology 2015

Abstract Methods This retrospective sub-study is based on the multi-

Objectives To assess the performance of a simplified MRI
protocol consisting of a contrast-enhanced three-dimensional
MR angiography (CE-MRA) in association with a post-
contrast T1-weighted sequence (T1WIV) for the detection of
HNPGLs in SDHx mutation carriers.
center PGL.EVA cohort, which prospectively enrolled SDHx
mutation carriers from 2005 to 2009; 157 index cases or rel-
atives were included. CE-MRA and the T1WIV images were
read solely with knowledge of the clinical data but blind to the
diagnosis. Sensitivity, specificity and likelihood ratios for the

Anne-Paule Gimenez-Roqueplo and Philippe Halimi jointly directed this

work.
Electronic supplementary material The online version of this article
(doi:10.1007/s00330-015-4024-5) contains supplementary material,
which is available to authorized users.

* Anne-Paule Gimenez-Roqueplo 8
INSERM UMR1052, UMR CNRS 5286, Cancer Research Center of
anne-paule.gimenez-roqueplo@egp.aphp.fr Lyon, F-69008 Lyon, France
9
Groupement Hospitalier Est, Service de Radiologie, Hospices civils
1 de Lyon, F-69003 Lyon, France
Assistance Publique-Hôpitaux de Paris, Service de Radiologie,
10
Hôpital Européen Georges Pompidou, F-75015 Paris, France Service de Radiologie, Centre Hospitalier Universitaire d’Angers,
2 F-49933 Angers, France
Assistance Publique-Hôpitaux de Marseille, Service
11
d’Endocrinologie, Diabète et Maladies Métaboliques, Centre Service de Radiologie, Centre Hospitalo-Universitaire de Reims,
Hospitalier Universitaire la Timone, F-13000 Marseille, France F-51000 Reims, France
3 12
Service d’Endocrinologie, Diabétologie, Nutrition, Centre Service de Radiologie, Centre Hospitalo-Universitaire de Lille,
Hospitalier Universitaire d’Angers, F-49933 Angers, France F-59000 Lille, France
4 13
LUNAM Université, INSERM, U1063, F-49933 Angers, France Assistance Publique-Hôpitaux de Paris, Service de Radiologie,
5 Hôpital Cochin, F-75006 Paris, France
Assistance Publique-Hôpitaux de Marseille, Service de Radiologie,
14
Centre Hospitalier Universitaire la Timone, Service de Radiologie, Institut Gustave Roussy,
F-13000 Marseille, France F-94800 Villejuif, France
6 15
Hospices civils de Lyon, Groupement Hospitalier Est, Fédération Service de Radiologie, Centre Hospitalo-Universitaire de Nantes,
d’Endocrinologie, F-69003 Lyon, France F-44000 Nantes, France
7 16
Faculté de Médecine Lyon-Est, Université de Lyon, Service de Neuroradiologie, Centre Hospitalo-Universitaire de
F-69372 Lyon, France Clermont-Ferrand, F-63000 Clermont Ferrand, France

simplified MRI protocol were compared to the full MRI pro-
tocol reading results and to the gold standard status obtained
through the consensus of an expert committee.
Results The sensitivity and specificity of the readings of the
simplified MRI protocol were, respectively, 88.7 % (95 %
CI=78.1–95.3) and 93.7 % (95 % CI=86.8–97.7) versus
80.7 % (95 % CI=68.6–89.6) and 94.7 % (95 % CI=88.1–
98.3) for the readings of the full MRI protocol.
Conclusions The simplified post-contrast MRI with shorter
duration (5 to 10 minutes) showed no performance difference
compared to the lengthy standard full MRI and can be pro-
posed for the detection of head and neck paragangliomas
(HNPGLs) in SDHx mutation carriers.
Key Points
• Rapid angio-MRI protocol and the usual lengthy protocol
show equal diagnostic performance.
• The CE-MRA is the key sequence for the detection of
HNPGLs.
• The T1WIV sequence assists in localizing HNPGLs.
Keywords Paraganglioma . Head and neck . Magnetic
resonance angiography . Screening . SDH
Abbreviations
HNPGL Head and neck paraganglioma
CE-MRA Contrast-enhanced three-dimensional MR
angiography
T1WIV Post-contrast T1-weighted sequence
Introduction
Head and neck paragangliomas (HNPGLs) are rare tumours
with an incidence of one in 30,000 to one in 10,000 [1, 2].
They develop from the paraganglionic tissue within the carot-
id bifurcation, in the jugulotympanic region, along the nodose
ganglia of the vagus nerve, or rarely in other locations such as
the larynx, nasal cavity and thyroid gland [2, 3]. Carotid body
17
Assistance Publique-Hôpitaux de Paris, Service de Génétique,
Hôpital Européen Georges Pompidou, 20-40, rue Leblanc,
75015 Paris, France
18
INSERM, UMR970, Paris Cardiovascular Research Center,
F-75015 Paris, France
19
Sorbonne Paris Cité, Faculté de Médecine, Université Paris
Descartes, F-75006 Paris, France
20
Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges
Pompidou, Unité d’Épidémiologie et de Recherche Clinique,
F-75015, Paris, France
21
INSERM, Centre d’investigation Épidémiologique 4,
F-75015 Paris, France
Eur Radiol
paragangliomas tend to displace the external carotid artery
(ECA) anteromedially and the internal carotid artery (ICA)
posterolaterally, while vagal paragangliomas usually displace
the ICA anteriorly [4]. HNPGLs typically appear as
hypervascular masses with intense early arterial phase en-
hancement on contrast enhanced CT or magnetic resonance
angiography and with a Bsalt and pepper^ appearance on T1
and T2-weighted images on magnetic resonance imaging
(MRI), referring to high velocity arterial branch flow voids
in tumour vessels (pepper appearance) and slow flow or haem-
orrhage in the tumour (salt appearance) [5–7]. The PGL.EVA
study showed, from the largest published prospective series of
238 SDHx mutation carriers, that MRI was better to detect
HNPGLs than [123I] meta-iodobenzylguanidine (mIBG) and
somatostatin receptor scintigraphies (SRS) [8]. In the
PGL.EVA study, the magnetic resonance imaging (MRI) pro-
tocol included up to eight pre-contrast and post-contrast se-
quences. The duration of such an exam exceeds 40 minutes.
Approximately 35 % of paragangliomas (PGLs) are genet-
ically determined by a germline mutation in one of the 13
identified susceptibility genes [9–13]. Eighty percent of
inherited PGLs are caused by a germline mutation in VHL or
SDHx group of genes [13]. The SDHx (SDHD, SDHC, SDHB,
SDHA, SDHAF2) genes encode the succinate dehydrogenase
(SDH) a mitochondrial enzyme that converts succinate into
fumarate in the Krebs cycle. Inactivation of succinate dehy-
drogenase results in the accumulation of succinate responsible
of an abnormal activation of hypoxia-angiogenesis pathway
that explains the hypervascularization of PGL [14]. In a large
cohort of SDHx mutation carriers, the prevalence of the
HNPGL was 62 % and 21 % at the age of 40 years for
SDHD and SDHB mutation carriers, respectively [15]. In
SDHx mutation carriers, PGLs occur early in life (mean age
between 29 and 37.9 years) and are frequently multifocal [8,
16, 17]. HNPGLs can lead to metastases up to 30 years after
initial diagnosis [18]. Approximately 16 % of patients with an
apparently sporadic HNPGL carry a germline mutation [15].
Therefore, genetic testing guided by the clinical presentation
is warranted in every patient with PGL [13, 15, 19].
Surgery and radiotherapy are the two treatment options
used for HNPGLs. HNPGLs can cause symptoms by
compressing adjacent structures with growth [20], and conse-
quently, vascular and neurological post-operative complica-
tions depend on tumour size [21, 22]. Radiotherapy versus
surgery shows good control rates and reduced morbidity,
and can stabilize or reduce tumour growth, while surgical
resection can succeed in eradication [23]. Some authors refer
to radiotherapy as a first line treatment of choice for HNPGLs
due to its excellent long-term local control rate [24].
The surveillance frequency of head and neck areas by MRI
in SDHx mutation carriers is yet to be established, although it
has been estimated that the tumours grow slowly with a me-
dian doubling time of 4.2 years [25]. Early imaging detection
Eur Radiol
and management could prevent neurological deficit due to an
enlarging tumour mass and minimize postoperative morbidity
[21, 22]. Genetically predisposed patients with SDHx muta-
tion require frequent screening to identify small lesions and
provide early treatment. Knowing that small HNPGLs with
diameters less than 10 mm are frequent in SDHx mutation
carriers and present as an enhancement focus without the typ-
ical Bsalt and pepper^ appearance [6], our hypothesis was that
detection of an enhancement focus alone could be sufficient
for the diagnosis of HNPGLs in SDHx mutation carriers.
Therefore, our objective was to assess the diagnostic perfor-
mance of a short post-contrast MRI protocol consisting of a
contrast-enhanced three-dimensional (3D) time-of-flight angi-
ography at arterial phase (CE-MRA) and an axial plane fast
spin-echo T1-weighted sequence with fat saturation (T1WIV)
in the detection of HNPGLs in SDHx mutation carriers.
Methods
Design
The HNMRI PGL.EVA study was designed as a retrospective
sub-study of the PGL.EVA study. The PGL.EVA study [8]
was a French prospective multicenter study (http://
clinicaltrials.gov/ct2/show/NCT00188019; registration
number NCT00188019) designed to assess the sensitivity
and specificity of the four screening methods usually
available in routine practice in 2004: head and neck
gadolinium-enhanced magnetic resonance angiography,
[123I] meta-iodobenzylguanidine scintigraphy, somatostatin
receptor scintigraphy and contrast-enhanced computed to-
mography, which was used only to detect thoracic, abdominal
and pelvic paragangliomas. The study was approved by the
appropriate ethics committee (Comité de Protection des
Personnes, CPP Ouest II, Angers, France). Written informed
consent was obtained from each patient for inclusion in the
study. In the former PGL.EVA study, results from all MRI
reports were obtained at local centres and a gold standard
status was established. The MRI protocol included transverse
and coronal plane T1-weighted spin-echo images, T2-
weighted fast spin-echo images, and T2-weighted fast spin-
echo with fat saturation images. After IV contrast injection of
gadolinium chelate (0.1 mmol/kg body weight, gadoteric acid;
Dotarem Guerbet, Aulnay-sous-Bois, France), a contrast-
enhanced three-dimensional time-of-flight angiography at ar-
terial phase and an axial plane fast spin-echo T1-weighted
sequence with fat saturation were obtained.
Sites and patients
Two categories of subjects were recruited in the PGL.EVA
study: 1) patients with a previous diagnosis of paraganglioma
(index cases) and 2) apparently asymptomatic subjects identi-
fied by familial genetic testing as being at risk (relatives).
They (men or women) were eligible if they were 6 years old
or older and had previously been informed of their positive
genetic status. Exclusion criteria were refusal or inability to
understand and sign informed consent, children under 6 years
of age, pregnant and/or lactating women, and SDHD mutation
inherited from the maternal branch. Patients with multiple
bone and/or lymph node metastases were not recruited in the
PGL.EVA study. The PGL.EVA cohort was formed by 258
subjects recruited prospectively in 23 French medical centres
from 6 June 2005 to 22 December 2009. As reported previ-
ously, 18 were not enrolled and two were excluded due to
technical problems with CD writings [8]. From the 238 sub-
jects included in the PGL.EVA study, 81 were excluded from
the sub-study due to missing CE-MRA and/or T1WIV se-
quences. Thus, 157 index cases or relatives carrying mutations
in SDHD, SDHB or SDHC genes were finally included in the
HNMRI PGL.EVA sub-study (Fig. 1).
Gold standard
As previously reported, the gold standard status for the diag-
nosis or exclusion of paragangliomas was obtained through
the consensus of an expert committee for each enrolled sub-
ject. It was based on the image readings of MRI, mIBG scan
and SRS, along with clinical, biological and genetic data [8].
Reading images
Three radiologists (two senior radiologists with more than
20 years experience and a junior radiologist with 2 years ex-
perience) retrospectively read solely the CE-MRA and the
T1WIV sequences from the head and neck MRI of all patients.
The retrospective readings were done with knowledge of the
clinical and family history, but without knowledge of diagnos-
tic results from the local centre reading or the gold standard
status. The readings of the CE-MRA sequence were per-
formed with an axial reformat from the skull base to the tho-
racic inlet and a sagittal oblique reformat with thin sections
and a maximal intensity projection (MIP) of 8–10 mm (Figs. 2
and 3). We evaluated sensitivity, specificity and likelihood
ratios for the short post-contrast MRI protocol (CE-MRA
and T1WIV), and compared them to the local centre reading
results and to the gold standard status from the PGL.EVA
study. A HNPGL was diagnosed when a hypervascular tumor
was found in a common location for HNPGL. Common loca-
tions for HNPGL included the carotid bifurcation, the region
of the vagus nerve (posterior region of the internal carotid
artery) and the jugulo-tympanic region. Each possible location
for HNPGLs in each patient was defined as normal (tumour-
free), positive (one or more paragangliomas detected), or
doubtful (hypervascular tumour in an uncommon location

Fig. 1 Flow chart of the PGL.EVA study and the HNMRI PGL.EVA sub-study
for paraganglioma or a non-hypervascular tumour in a com-
mon location for paraganglioma) for each sequence. The pa-
tient was considered positive if either one of the two se-
quences was positive for HNPGL.
Statistical analyses
Continuous variables are presented as means +/− 1 SD.
Categorical variables are presented as numbers and per-
centages. We obtained exact 95 % confidence intervals
(CIs) for sensitivity and specificity from the binomial dis-
tribution. We calculated likelihood ratios for a positive
test result as sensitivity divided by (1 – specificity) and
likelihood ratios for negative result as (1 – sensitivity)
divided by specificity. We calculated 95 % CIs for likeli-
hood ratios by using the normal distribution approxima-
tion. Because a doubtful status was a possible result for
all exams, we calculated performance characteristics in
two different ways: 1) calculation of the likelihood ratio
for a doubtful result and 2) exclusion of doubtful results
from calculations. Likelihood ratio for a doubtful result
was calculated as [(number of doubtful exams with a pos-
itive gold standard status/number of exams with a positive
gold standard status)/(number of doubtful exams with a
negative gold standard status/number of exams with a
negative gold standard status)]. Doubtful results were
Eur Radiol
reclassified as negative when the likelihood ratio was less
than or equal to 1, and as positive when the likelihood
ratio was more than 1.
Role of the funding source
The Programme Hospitalier de Recherche Clinique had no
role in study design, data collection and analysis, interpreta-
tion or the decision to submit the manuscript for publication.
Results
Patients and tumours
One hundred and fifty-seven subjects were included, 75
(47.8 %) patients were index cases and 82 (52.2 %) were
relatives. At the gold standard status, 62 (39.5 %) patients
had a HNPGL, 49 were index cases and 13 were relatives.
Thirty-two (20.4 %) patients had two or more HNPGLs
(Table 1).
Diagnostic performances of the two MRI protocols
The diagnostic performances of the two protocols are de-
scribed in Table 2. Two lesions were considered doubtful with
Eur Radiol

Fig. 2 CE-MRA Sagittal oblique

MIP reformats oriented along the
axis of right (a) and left (b) carotid
bifurcation in a patient showing
four HNPGLs [two carotid PGLs
(open arrows) and two vagal
PGLs (arrows)] as four tumoral
blushes in a 35-year-old patient
carrying a SDHD mutation.
Because of their sizes, the right
carotid body PGL displaces
anteriorly the ECA (open
arrowhead) and posteriorly the
ICA (arrowhead), and the right
vagal PGL displaces anteriorly
the ICA (small arrow). c and d,
Axial contrast-enhanced fat-
saturated T1-weighted images
showing the four PGLs

the simplified protocol versus 11 with the usual protocol. post-contrast MRI protocol and the readings of the full pre-
There was no statistical difference in the detection perfor- contrast and post-contrast MRI protocol with sensitivity and
mance of HNPGLs between the readings of the simplified specificity, respectively, of 88.7 % (95 % CI=78.1–95.3) and

Fig. 3 CE-MRA axial MPR

image (a) and sagittal oblique
MIP reformat (b) in a 65-year-old
patient with a left jugulo-
tympanic PGL (arrow)
 Eur Radiol

Table 1 General characteristics

of patients according to their SDHB SDHC SDHD All
SDHx mutation
n 81 13 63 157
Mean age +/- sd (yr) 42.0 +/- 15.3 39.0 +/- 14.8 45.0 +/- 15.3 42.9 +/- 15.3
Relatives [n (%)] 57 (70.4) 7 (53.9) 18 (28.6) 82 (52.2)
Index cases [n (%)] 24 (29.6) 6 (46.2) 45 (71.4 75 (47.8)
0 HNPGL [n (%)] 69 (85.2) 8 (61.4) 18 (28.6) 95 (60.5)
1 HNPGL [n (%)] 11 (13.6) 4 ( 30.7) 15 (23.8) 30 (19.1)
2 HNPGLs [n (%)] 1 (1.2) 1 (7.7) 20 (31.8) 22 (14.0)
3 HNPGLs [n (%)] 0 (0.0) 0 (0.0) 4 (6.4) 4 (2.6)
4 HNPGLs [n (%)] 0 (0.0) 0 (0.0) 5 (7.9) 5 (3.2)
5 HNPGLs [n (%)] 0 (0.0) 0 (0.0) 1 (1.6) 1 (0.6)
Positive at GS [n (%)] 12 (14.8) 5 (38.5) 45 (71.4) 62 (39.5)
Negative at GS [n (%)] 69 (85.2) 8 (61.5) 18 (28.6) 95 (60.5)

HNPGL=head and neck paraganglioma; GS=gold standard

93.7 % (95 % CI=86.8–97.7) versus 80.7 % (95 % CI=68.6–
89.6) and 94.7 % (95 % CI=88.1–98.3). With the short proto-
col, seven positive patients were missed (false negatives) versus
12 with the usual full protocol. There were six false positives
with the short protocol versus five with the usual full protocol.
Discussion
Early imaging detection of HNPGLs can prove valuable to
clinical management. MRI has proven superior to CT for the
detection of HNPGLs, especially for smaller paragangliomas
[5, 26]. The prior PGL.EVA study demonstrated improved de-
tection rates with MRI compared to mIBG scans and SRS.
Moreover, some authors refer to MRI as a standard for the
diagnosis of HNPGLs compared to 18F-fluorodeoxyglucose
positron emission tomography with computed tomography
(18F-FDG PET/CT) and to mIBG scans [27]. CE-MRA is high-
ly specific for the detection of HNPGLs. Neves et al. [28] indi-
cated that the addition of a CE-MRA to a conventional MRI
protocol significantly improves the specificity, from 41 to 94 %.
To our knowledge, no study has ever tested the diagnostic
performance of a short MRI protocol comprising a CE-MRA
for the detection of HNPGLs in SDHx mutation carriers. In
2004, van den Berg et al. did test MR angiography techniques
for the detection of HNPGLs, but those techniques were only
unenhanced sequences, such as 3D phase-contrast angiogra-
phies, 2D and 3D time-of-flight angiographies and proton
density weighted sequences [29]. In order to establish a
shorter screening protocol and to limit radiation exposure over
lifetime, the diagnostic performance of a rapid unenhanced
sequence whole-body MRI for the detection of
paragangliomas (including HNPGLs) in SDHx mutation car-
riers was tested with a sensitivity of 87.5 % [30]. Although
these results are promising, the significance of this study is
limited by the small size of the cohort (six tumours found in 37
SDHx mutation carriers) and the lack of a clear gold standard
status to reference.
18
F-FDOPA PET could be an interesting imaging modality
for the detection of HNPGLs. Hoegerle et al. [31] evaluated
the diagnostic performance of 18F-FDOPA PET for the detec-
tion of HNPGLs in ten consecutive SDHD mutation carriers
and compared it to MRI. In this work, 18F-FDOPA PET
seemed more sensitive than MRI, with three tumours detected
by 18F-FDOPA PET that were missed by MRI, and further
confirmed by the correlation of 18F-FDOPA PET and MRI.
In 2011, King et al. [32] also evaluated the diagnostic perfor-
mance of five different functional imaging techniques, includ-
ing 18F-FDOPA PET. 18F-FDOPA PET localized 26 of the 26
tumours when five were missed by MRI. However, MRI pro-
tocols in those two studies were not optimal because they
lacked an arterial phase post-contrast sequence. Therefore,
MRI wasn’t able to highlight the typical arterial phase contrast
enhancement of PGLs, which is essential to differentiate small
HNPGLs in atypical locations from lymph nodes and also
improves the screening performance. Efficiency of 18F-
FDOPA PET should be confirmed in larger cohort and com-
pared to an appropriate MRI protocol with arterial contrast
enhancement. In routinely practice, 18F-FDOPA PET is still
restricted by the radiopharmaceutical cost which ranges be-
tween 1000 and 1500€ in our country. Furthermore, nuclear
medicine imaging modalities should be reserved to further
characterize detected tumours, and are not indicated for period-
ic imaging of patients with a germline mutation in a suscepti-
bility gene for whom radiation exposure should be limited [19].
In our experience, the post-contrast 3D angio-MR at arte-
rial phase is sufficient for the detection of HNPGLs. We also
think that the addition of a post-contrast T1-weighted se-
quence in the protocol provides additional diagnostic detail
that assists in localizing HNPGLs (Fig. 4), which can prove
useful for the pre-surgical or pre-radiotherapy assessment. We
believe that using a MIP of 8–10 mm is essential to depict sub-
Eur Radiol

Table 2 Diagnostic performances of the two MRI protocols

CE-MRA+T1WIV Full MRI

n 157 157
Doubtful exams [n (%)] 2 (1.3) 11 (7)
Doubtful likelihood ratio 1.5 0.9
Sensitivity (95 % CI) 88.7 (78.1–95.3) 80.7 (68.6–89.6)
Specificity (95 % CI) 93.7 (86.8–97.7) 94.7 (88.1–98.3)
Positive predictive value (95 % CI) 90.2 (79.8–96.3) 90.9 (80.1–97.0)
Negative predictive value (95 % CI) 92.7 (85.6–97.0) 88.2 (80.4–93.8)
Positive likelihood ratio (95 % CI) 14.0 (6.4–30.6) 15.3 (6.5–36.3)
Negative likelihood ratio (95 % CI) 0.1 (0.1–0.2) 0.2 (0.1–0.3)
Sensitivity (without doubtful cases) (95 % CI) 88.5 (77.8–95.3) 86.2 (74.6–93.9)
Specificity (without doubtful cases) (95 % CI) 94.7 (88.0–98.3) 94.3 (87.2–98.1)

CE-MRA=contrast-–enhanced MR angiography; T1WIV=post-contrast T1-weighted sequence with fat saturation

centimetric tumours by differentiating them from small vas- Although the sensitivities of the two MRI protocols are not
cular branches. Likewise, we advise the use of a sagittal statistically different, we suspect that the detection perfor-
oblique reformat, oriented along the axis of carotid bifurca- mance is slightly greater with the short protocol and further
tion, to detect small carotid or vagal paragangliomas. studies may pursue this hypothesis. The short protocol, in

Fig. 4 A 49-year-old patient with

a laryngeal PGL (arrow) carrying
a SDHD mutation. CE-MRA
axial MPR image (a) and CE-
MRA coronal MIP reformat (b)
show the tumor blush at arterial
phase in the laryngeal region. c
The axial contrast-enhanced fat-
saturated T1-weighted sequence
is optimal for the definitive spatial
localization

which only two sequences are analysed, may be more focused,
as opposed to the full eight sequences protocol, which may
contribute to information overload. Finally, although T1- and
T2-weighted images seem to have no impact for screening,
they are surely useful for the characterization of neck masses
in patients without a HNPGL history or an SDHx mutation.
We acknowledge that our retrospective study has weak-
nesses. First, the readings of the original full protocol MRIs
were not performed by the reviewers of the short MRIs, but
we think that if both protocols had been read by the three
reviewers, then the results could have been biased by the first
reading. Secondly, the study design differentiated patients into
two groups, those with and without paragangliomas. Thus, an
exam finding one paraganglioma in a patient who had two or
more lesions was considered a positive exam, even though the
other lesions might have been missed. We believed that this
has little impact on the comparison of the protocols, because
this study design applies to the two protocols tested. Thirdly,
the gold standard was not based on histopathology, while all
patients didn’t undergo surgical resection. However, we be-
lieve that our gold standard remained robust while it was de-
pending on multiple imaging techniques (MRI, SRS and
mIBG scans) as well as clinical, genetic and biological data
through an expert committee.
The decreased scan time using our short protocol raises the
possibility of combining the head and neck screening MRI
with the thoracic, abdominal and pelvic (TAP) MRI screening
recommended by the clinical practice guideline from the en-
docrine society [19]. The patient could achieve a whole body
screening within the same day, rather than by the two-day
protocol currently employed, further reducing logistical re-
dundancies. The feasibility and sensitivity of two consecutive
contrast exams would have to be confirmed by further studies.
In conclusion, our simplified post-contrast angio-MR with
exam duration of 5 to 10 minutes showed equal performance
in the detection of HNPGLs in screening exams compared to
the typically lengthy full pre-contrast and post-contrast MRI
with an average duration of 40 minutes. Screening of
HNPGLs in SDHx mutations carriers could be performed with
only a post-contrast 3D angio-MR at arterial phase and a post-
contrast T1-weighted with fat saturation sequence. This
shorter MRI screening protocol would drastically optimize
equipment and personnel resources, due to decreased exam
duration and diminished reader fatigue and information over-
load from sequences not necessary for screening exams. In
addition, this protocol could further optimize resources by
possibly allowing the option to combine with a same-setting,
whole body MRI screening in a single two-part exam.
Acknowledgments Authors would like to thank Steven M. Yevich for
editing the English text. The scientific guarantor of this publication is
Prof. Philippe Halimi. The authors of this manuscript declare no relation-
ships with any companies, whose products or services may be related to
Eur Radiol
the subject matter of the article. This study has received funding by
Program Hospitalier National de Recherche Clinique 2004 (PCR05007).
One of the authors has significant statistical expertise: Aurore Caumont-
Prim from Assistance Publique-Hôpitaux de Paris, Hôpital Européen
Georges Pompidou, Unité d’Épidémiologie et de Recherche Clinique,
F-75015 Paris, France. Institutional Review Board approval was obtained.
Written informed consent was obtained from all subjects (patients) in this
study. Some study subjects or cohorts have been previously reported in:
Gimenez-Roqueplo A-P, Caumont-Prim A, Houzard C, et al. (2013) Im-
aging work-up for screening of paraganglioma and pheochromocytoma in
SDHx mutation carriers: a multicenter prospective study from the
PGL.EVA Investigators. J Clin Endocrinol Metab 98:E162–173. Method-
ology: retrospective, diagnostic study, multicenter study.
References
1. Mariman EC, van Beersum SE, Cremers CW et al (1993) Analysis
of a second family with hereditary non-chromaffin paragangliomas
locates the underlying gene at the proximal region of chromosome
11q. Hum Genet 91:357–361
2. Lack EE, Cubilla AL, Woodruff JM, Farr HW (1977)
Paragangliomas of the head and neck region. A clinical study of
69 patients. Cancer 39:397–409
3. Castelblanco E, Gallel P, Ros S et al (2012) Thyroid paraganglioma.
Report of 3 cases and description of an immunohistochemical pro-
file useful in the differential diagnosis with medullary thyroid car-
cinoma, based on complementary DNA array results. Hum Pathol
43:1103–1112
4. Alkadhi H, Schuknecht B, Stoeckli SJ, Valavanis A (2002)
Evaluation of topography and vascularization of cervical
paragangliomas by magnetic resonance imaging and color duplex
sonography. Neuroradiology 44:83–90
5. Olsen WL, Dillon WP, Kelly WM et al (1987) MR imaging of
paragangliomas. AJR Am J Roentgenol 148:201–204
6. van Gils AP, van den Berg R, Falke TH et al (1994) MR diagnosis
of paraganglioma of the head and neck: value of contrast enhance-
ment. AJR Am J Roentgenol 162:147–153
7. Barber B, Ingram M, Khan S et al (2011) Clinicoradiological man-
ifestations of paraganglioma syndromes associated with succinyl
dehydrogenase enzyme mutation. Insights Imaging 2:431–438
8. Gimenez-Roqueplo A-P, Caumont-Prim A, Houzard C et al (2013)
Imaging work-up for screening of paraganglioma and pheochromo-
cytoma in SDHx mutation carriers: a multicenter prospective study
from the PGL.EVA Investigators. J Clin Endocrinol Metab 98:
E162–E173
9. Baysal BE, Willett-Brozick JE, Lawrence EC et al (2002)
Prevalence of SDHB, SDHC, and SDHD germline mutations in
clinic patients with head and neck paragangliomas. J Med Genet
39:178–183
10. Boedeker CC, Neumann HPH, Maier W et al (2007) Malignant
head and neck paragangliomas in SDHB mutation carriers. Off J
Am Acad Otolaryngol-Head Neck Surg 137:126–129
11. Drovdlic CM, Myers EN, Peters JA et al (2001) Proportion of
heritable paraganglioma cases and associated clinical characteris-
tics. Laryngoscope 111:1822–1827
12. Badenhop RF, Jansen JC, Fagan PA et al (2004) The prevalence of
SDHB, SDHC, and SDHD mutations in patients with head and
neck paraganglioma and association of mutations with clinical fea-
tures. J Med Genet 41:e99
13. Favier J, Amar L, Gimenez-Roqueplo A-P (2014) Paraganglioma
and phaeochromocytoma: from genetics to personalized medicine.
Nat Rev Endocrinol. doi:10.1038/nrendo.2014.188
Eur Radiol
14. Favier J, Gimenez-Roqueplo A-P (2010) Pheochromocytomas: the
(pseudo)-hypoxia hypothesis. Best Pract Res Clin Endocrinol
Metab 24:957–968
15. Burnichon N, Rohmer V, Amar L et al (2009) The succinate dehy-
drogenase genetic testing in a large prospective series of patients
with paragangliomas. J Clin Endocrinol Metab 94:2817–2827
16. Neumann HPH, Pawlu C, Peczkowska M et al (2004) Distinct
clinical features of paraganglioma syndromes associated with
SDHB and SDHD gene mutations. JAMA J Am Med Assoc 292:
943–951
17. Timmers HJLM, Kozupa A, Eisenhofer G et al (2007) Clinical
presentations, biochemical phenotypes, and genotype-phenotype
correlations in patients with succinate dehydrogenase subunit B-
associated pheochromocytomas and paragangliomas. J Clin
Endocrinol Metab 92:779–786
18. Brewis C, Bottrill ID, Wharton SB, Moffat DA (2000) Metastases
from glomus jugulare tumours. J Laryngol Otol 114:17–23
19. Lenders JWM, Duh Q-Y, Eisenhofer G et al (2014)
Pheochromocytoma and paraganglioma: an endocrine society clin-
ical practice guideline. J Clin Endocrinol Metab 99:1915–1942
20. Martin TPC, Irving RM, Maher ER (2007) The genetics of
paragangliomas: a review. Clin Otolaryngol Off J ENT-UK Off J
Neth Soc Oto-Rhino-Laryngol Cervico-Facial Surg 32:7–11
21. Makeieff M, Raingeard I, Alric P et al (2008) Surgical management
of carotid body tumors. Ann Surg Oncol 15:2180–2186
22. Paris J, Facon F, Thomassin JM, Zanaret M (2006) Cervical
paragangliomas: neurovascular surgical risk and therapeutic man-
agement. Eur Arch Oto-Rhino-Laryngol Off J Eur Fed Oto-Rhino-
Laryngol Soc EUFOS Affil Ger Soc Oto-Rhino-Laryngol - Head
Neck Surg 263:860–865
23. Huy PTB, Kania R, Duet M et al (2009) Evolving concepts in the
management of jugular paraganglioma: a comparison of
View publication stats
radiotherapy and surgery in 88 cases. Skull Base Off J North Am
Skull Base Soc Al 19:83–91
24. Dupin C, Lang P, Dessard-Diana B et al (2014) Treatment of head
and neck paragangliomas with external beam radiation therapy. Int J
Radiat Oncol Biol Phys 89:353–359
25. Jansen JC, van den Berg R, Kuiper A et al (2000) Estimation of
growth rate in patients with head and neck paragangliomas influ-
ences the treatment proposal. Cancer 88:2811–2816
26. Vogl T, Brüning R, Schedel H et al (1989) Paragangliomas of the
jugular bulb and carotid body: MR imaging with short sequences
and Gd-DTPA enhancement. AJR Am J Roentgenol 153:583–587
27. Timmers HJLM, Chen CC, Carrasquillo JA et al (2012) Staging
and functional characterization of pheochromocytoma and
paraganglioma by 18F-fluorodeoxyglucose (18F-FDG) positron
emission tomography. J Natl Cancer Inst 104:700–708
28. Neves F, Huwart L, Jourdan G et al (2008) Head and neck
paragangliomas: value of contrast-enhanced 3D MR angiography.
AJNR Am J Neuroradiol 29:883–889
29. van den Berg R, Schepers A, de Bruïne FT et al (2004) The value of
MR angiography techniques in the detection of head and neck
paragangliomas. Eur J Radiol 52:240–245
30. Jasperson KW, Kohlmann W, Gammon A et al (2014) Role of rapid
sequence whole-body MRI screening in SDH-associated hereditary
paraganglioma families. Fam Cancer 13:257–265
31. Hoegerle S, Ghanem N, Altehoefer C et al (2003) 18F-DOPA pos-
itron emission tomography for the detection of glomus tumours.
Eur J Nucl Med Mol Imaging 30:689–694
32. King KS, Chen CC, Alexopoulos DK et al (2011) Functional im-
aging of SDHx-related head and neck paragangliomas: comparison
of 18F-fluorodihydroxyphenylalanine, 18F-fluorodopamine, 18F-
fluoro-2-deoxy-D-glucose PET, 123I-metaiodobenzylguanidine
scintigraphy, and 111In-pentetreotide scintigraphy. J Clin
Endocrinol Metab 96:2779–2785