Neuroendocrine tumors of the small bowel: evaluation with MR-enterography

    Neuroendocrine Tumors of the Small Bowel: enterography

Evaluation with MR-

Anthony Dohan, Hassan El Fattach, Maxime Barat, Youcef Guerrache, Clarisse Eveno, Raphael Dautry, S´ebastien Mul´e, Mourad Boudiaf, Christine Hoeffel, Philippe Soyer PII: DOI: Reference:

S0899-7071(16)00012-7 doi: 10.1016/j.clinimag.2015.12.016 JCT 7986

To appear in:

Journal of Clinical Imaging

Received date: Accepted date:

15 October 2015 20 December 2015

Please cite this article as: Dohan Anthony, El Fattach Hassan, Barat Maxime, Guerrache Youcef, Eveno Clarisse, Dautry Raphael, Mul´e S´ebastien, Boudiaf Mourad, Hoeffel Christine, Soyer Philippe, Neuroendocrine Tumors of the Small Bowel: Evaluation with MR-enterography, Journal of Clinical Imaging (2016), doi: 10.1016/j.clinimag.2015.12.016

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

RI P SC

NU

Anthony Dohan1,2,3 ([email protected]) Hassan El Fattach1 ([email protected]) Maxime Barat1 ([email protected]) Youcef Guerrache1 ([email protected]) Clarisse Eveno2,3,4 ([email protected]) Raphael Dautry 1,2 ([email protected]) Sébastien Mulé 5 ([email protected]) Mourad Boudiaf1 ([email protected]) Christine Hoeffel5 ([email protected]) Philippe Soyer1,2,3 ([email protected])

T

Neuroendocrine Tumors of the Small Bowel: Evaluation with MR-enterography

of Body and Interventional Imaging, Hôpital Lariboisière, AP-HP, 2 rue Ambroise Paré, 75475 Paris Cedex 10, France. 2Université Paris-Diderot, Sorbonne-Paris Cité, 10 rue de Verdun, 75010 Paris, France 3UMR INSERM 965, Hôpital Lariboisière, 2 rue Amboise Paré, 75010 Paris, France 4Department of Digestive Surgery, Hôpital Lariboisière, AP-HP, 2 rue Ambroise Paré, 75475 Paris cedex 10, France 5Department of Radiology, Hôpital Robert Debré, 11 Boulevard Pasteur, 51092 Reims Cedex, France

PT

ED

MA

1Department

AC

CE

Correspondence: Philippe Soyer, MD, PhD Professor and Chairman Department of Body & Interventional Imaging Hôpital Lariboisière - APHP 2 Rue Ambroise Paré 75 475 Paris Cedex 10, France Short title: MR-enterography of Small Bowel Neuroendocrine Tumors

ACCEPTED MANUSCRIPT Full abstract Purpose. To determine the sensitivity of MR-enterography for the detection of neuroendocrine tumors of the small-bowel (NETSB) and analyze the imaging presentation of

T

NETSB on MR-enterography.

RI P

Patients & Methods. The MR-enterography studies (including HASTE, TruFISP, and 3D VIBE MR sequences before and after intravenous administration of a gadolinium-chelate) of

SC

19 patients with pathologically confirmed NETSB were blindly reviewed. Images were analyzed with respect to imaging presentation. Sensitivity of MR-enterography as well as

NU

that of each individual MR-enterography sequence for the diagnosis of NETSB was estimated with 95% confidence intervals (CIs). Comparisons between individual MR-

MA

enterography sequences were performed using the McNemar test. Results. Twenty-seven NETSBs were confirmed in 19 patients. Overall sensitivity of MRenterography for NETSB detection was 74% (20/27; 95%CI: 54-89%) on a per-lesion basis.

ED

On a per-patient basis, MR-enterography had a sensitivity of 95% (18/19; 95%CI: 74-100%) for the detection of NETSB. Best degrees of sensitivity were achieved with 3D VIBE MR-

PT

enterography sequences after intravenous administration of a gadolinium-chelate (Se=95%; 18/19) by comparison with HASTE (Se=26%; 5/19) and TruFISP (Se=26%; 5/19) sequences

CE

(p=0.00022). Visible focal small-bowel mass, mesenteric stranding, and mesenteric mass were found in 16/19 (84%), 17/19 (89%) and 15/19 (79%) patients, respectively.

AC

Conclusion. MR-enterography shows highly suggestive features for the diagnosis of NETSB and has high degrees of sensitivity for the diagnosis of NETSB on a per-patient basis.

Index terms: MR-enterography - Intestinal neoplasm - Neuroendocrine tumor - Small bowel tumor - Small bowel - Small bowel tumor detection

ACCEPTED MANUSCRIPT Introduction

Neuroendocrine tumors (NETs) represent almost one third of all primary neoplasms of the

T

small-bowel [1-4]. They originate from enterochromaffin cells and are predominantly located

RI P

in the distal ileum [1-3]. Most NET show an overexpression of certain subtypes of somatostatin receptors on the cell membrane [5]. To avoid potential confusion, NET is now

SC

being used instead of carcinoid tumors [1, 2]. NETs of the small bowel have been extensively studied with computed tomography (CT) [6-13]. It has been showed that CT used in

NU

conjunction with distension of the small bowel and intravenous administration of iodinated contrast material has high degrees of sensitivity for the detection of NETs of the small bowel

MA

[14-17]. In addition, NETs of the small-bowel usually induce a desmoplastic reaction within the mesentery and are often associated with mesenteric masses and tethering of adjacent small-bowel loops so that CT provides suggestive findings to the diagnosis of NET [13, 18-

ED

20]. However, because of concerns regarding radiation dose given to the patient and tolerance, the use of CT-enteroclysis or CT-enterography is now less favored at many

PT

facilities [18, 21, 22].

To address the issue of radiation while keeping optimal small-bowel distension, some

CE

authors have advocated the use of MR-enteroclysis for the detection of small bowel tumors [23-26]. However, this technique still requires administration of enteral contrast using a

AC

nasojejunal tube that is placed under fluoroscopic guidance, so that patient tolerance remains limited and the issue of radiation is partially addressed [24, 24]. More recently, some researchers have investigated MR-enterography in the field of small-bowel tumor detection [27, 28]. In this regard, these researchers have reported sensitivity of 90% for the detection of small-bowel tumors using MR-enterography [27, 28]. However, the issue of NET detection with MR-enterography has not been extensively addressed. One reason is that patients with NET were included in a more general population [27, 28] or because only a few studies with a limited number of patients with NETs have specifically addressed this issue [29-31]. The goals of this study were to determine the sensitivity of MR-enterography for the detection of NETs of the small-bowel and analyze the imaging presentation of these tumors on MR-enterography.

ACCEPTED MANUSCRIPT MATERIALS AND METHODS

RI P

T

Study Population

From February 2009 to January 2015 inclusively, the MR imaging databases of two

SC

institutions were retrospectively queried to identify all patients who had MR-enterography for suspicion of NET of the small bowel. This study was approved by our institutional review

NU

boards and informed consent was waived. The study coordinator initially identified 34 patients for whom clinical, histopathological and imaging data were analyzed to ascertain that they actually had NET of the small bowel.

MA

Fifteen patients were thus excluded from the study because the lesion seen at MRenterography corresponded to an alternate diagnosis (n=9) NETs were located on the

ED

duodenum (n=2), or no definite confirmation of the diagnosis of NET was available (n=4) according to our reference standard.

PT

The final cohort comprised 19 patients with histopathologically confirmed NET of the small-bowel. There were 10 men and 9 women, with a mean age of 59.3 years ± 17.8 (SD)

CE

(range: 48-87 years). The patients had MR-enterography because of abnormal videocapsule endoscopy findings (n=6), obscure gastrointestinal bleeding (n=3), carcinoid syndrome

AC

(n=2), hypervascular hepatic metastases on CT or MR imaging suggestive for NET (n=6) or tumor mass of the terminal ileum at video ileocolonoscopy (n=2).

MR-enterography protocol

All MR examinations were performed with a MR unit working at 1.5 T (Magnetom Avanto, BV17 version, Siemens Healthcare, Erlangen, Germany) with two six-channel abdominal phased-array coils. Patients fasted for at least six hours before MR-enterography and were asked to drink 1.5 L of an iso-osmotic solution of water mixed with polyethylene glycol (PEG) and electrolytes during 45 min before the examination. The PEG-water solution was prepared by dissolving a granular powder containing 59 g of Macrogol 3350 (Colopeg,

ACCEPTED MANUSCRIPT Bayer, Paris, France), 1.461 g of anhydrous sodium sulfate, 1.680 g of sodium bicarbonate, 0.746 g of sodium chloride in 1.5 L of tap water [27, 28].

T

Images were acquired with the patient placed in prone position. All patients had T2-

RI P

weighted single shot half-Fourier turbo spin echo (HASTE) and true free-induction with steady state free precession (TrueFISP) MR sequences in the transverse and coronal planes providing complete abdominal and pelvic coverage. Fat-saturated three-dimensional low

SC

angle volumetric interpolated breath hold (3D VIBE) T1-weighted gradient echo MR sequences were acquired before and after intravenous administration of 0.2 mg/kg gadoterate

NU

meglumine (Dotarem®, Guerbet, Roissy-Charle de Gaulle, France) at a rate of 3 mL/sec. Arterial and portal venous phases were acquired in the coronal plane and followed by three

MA

acquisitions in the transverse plane at 120, 180 and 240 sec. Imaging parameters of the various MR sequences are described in Table 1.

To reduce small-bowel peristalsis, 10 patients received 0.5 mg of glucagon

ED

(Glucagen®; Novo Nordisk, Bagsvaerd, Denmark) that was intravenously administered one min before the start of 3D VIBE T1-weighted MR sequences [27, 28] and 9 patients received

CE

Image Analysis

PT

tiemonium methylsulphate (Viscéralgine®, Laboratoires Organon, Puteaux, France) [32].

AC

MR-enterography examinations were reviewed using a picture archiving and communication system (PACS) workstation (Directview, 11.3 version, Carestream Health Inc, Rochester, NY, USA) by two radiologists with an experience of 15 and 2 years in reading MRenterography examination working in consensus, blinded to any information relative to the patient, including the fact that the patients had NET of the small bowel. Four individual reading sessions were performed during which individual MR sequences of all patients were analyzed separately for presence or absence of NET of the small bowel. During these reading sessions, images were analyzed with respect to NET of the small-bowel detection only. To minimize recall bias, all reading sessions were performed at least two weeks from each other. In addition, MR-enterography sequences were evaluated for global image quality and small bowel distension.

ACCEPTED MANUSCRIPT After the four individual reading sessions were completed, a fifth reading session was performed during which all MR enterography sequences of an individual patient were reviewed as a single set. The results obtained using each single MR sequence were compared

T

to those obtained with the whole MR-enterography examination with respect to sensitivity,

RI P

specificity and accuracy for the diagnosis of NET of the small-bowel. During this session, several MR-enterography findings were evaluated by using a standardized data collection

SC

form. They included the presence of visible small-bowel masses, focal small-bowel wall thickening, mesenteric masses [14, 15, 27, 28], mesenteric stranding [23, 24], presence of

NU

any visible lymph nodes other than the mesenteric mass in the abdomen or retroperitoneum, presence of enlarged lymph nodes [33, 34], vascular encasement [7, 14], peritoneal nodules

MA

other than the mesenteric mass, and peritoneal free-fluid effusion. MR-enterography examinations were also analyzed for the presence of luminal narrowing of the small-bowel, small bowel dilatation proximal to luminal narrowing [35], dilatation of small bowel loops

ED

involved by NET and hepatic metastasis [36]. When present, mesenteric mass was further scrutinized for internal calcifications [6, 10, 37].

PT

When a small-bowel mass was visible, it was further analyzed for location (jejunum, proximal ileum or distal ileum), size (largest dimension), morphological presentation

CE

(intraluminal, extraluminal or both), and patterns of enhancement after intravenous administration of gadolinium-chelate relative to adjacent small-bowel wall [27, 28].

AC

Enhancement was classified as hyperenhancement when NET enhancement was greater than that of adjacent small-bowel wall or mild when similar to that of adjacent small-bowel wall. Luminal narrowing was classified as severe (> 50%) or moderate (< 50%). Liver metastasis was considered when heterogeneous focal liver lesions were visible [29, 36]. Mesenteric stranding was considered in the presence of radiating infiltration of the mesenteric fat [7, 15, 19, 29]. The presence of lymph nodes was evaluated visually and searched in the mesenteric and ileocecocolic areas. When present, lymph nodes were analyzed in terms of size (i.e., shortest axial diameter in mm) and degree of enhancement (hyper, mild or absent). Mesenteric lymph node enlargement was considered when visible mesenteric lymph nodes had a short-axis diameter > 10-mm [34, 38]. Vascular encasement was defined as either occlusion or narrowing of a vessel by a soft-tissue mass surrounding the area of involvement [7, 39].

ACCEPTED MANUSCRIPT Small-bowel dilatation was considered present when the maximal lumen diameter of the small-bowel loop was > 30 mm [32, 35]. Fluid effusion was searched in the perihepatic and perisplenic areas, the Douglas pouch, the parietocolic gutters and the perirenal spaces.

RI P

T

The presence of peritoneal fluid effusion was binary evaluated as present or absent.

SC

Standard of Reference

The standard of reference consisted of histopathological confirmation of NET of the small-

NU

bowel after surgical removal using open laparotomy and the results of intraoperative exploration, including manual palpation of the small bowel in all patients. Before planning

MA

the surgical approach, the surgeon was aware of the results of MR-enterography and endoscopic examinations. Hepatic metastases were histopathologically confirmed either after percutaneous (n = 6) or intraoperative biopsy (n = 2).

ED

For per-lesion analysis, the study coordinator established the standard of reference for determining the actual number of NETs present in each patient using surgical and

PT

histopathological reports for all patients, and the results of video ileocolonoscopy and videocapsule endoscopy in 12 and 6 patients, respectively. Findings were matched using

CE

tumor-by-tumor correlation on the basis of location, dimensions and tumor morphology. Dimensions of NETs missed at MR-enterography were determined on the basis of

AC

histopathological and surgical reports.

Statistical Analysis

Statistical analysis was performed with commercially available software (SAS, version 9.2, SAS Institute, Cary, NC, USA). Descriptive statistics were calculated for all variables evaluated at MR-enterography. Qualitative (binary) data were expressed as raw numbers, proportions and frequencies. The number of missed NETs and that of visible ones at MR-enterography were counted and further analyzed according to their dimensions. The capabilities of each individual MR-enterography sequence and those of the full set for the detection of NET were

ACCEPTED MANUSCRIPT evaluated in terms of sensitivity. Sensitivity was defined as the true-positive rate and calculated on per-patient and per-lesion basis. On a per patient basis, sensitivity was defined for each MR-enterography sequence as

T

the number of patients correctly depicted as having NET with the MR-enterography sequence

RI P

divided by the number of patients in whom NETs were actually present as defined with the standard of reference. On a per-lesion basis, sensitivity was defined for each MR-

SC

enterography sequence as the number of NETs correctly depicted with the MR-enterography sequence divided by the number of NETs actually present as defined with the standard of

NU

reference. Sensitivities and corresponding 95% confidence intervals (CIs) were thus calculated for each individual MR-enterography sequence and the full set of MR-

MA

enterography sequences.

The difference in sensitivity between two given MR-enterography sequences was the sensitivity of one sequence minus the sensitivity of the other sequence. Ninety-five p. cent CIs

ED

were calculated for the differences in sensitivity between each sequence. With a confidence of 95%, the calculation of the CI provided a range of plausible differences in sensitivities [40].

PT

Differences in sensitivity between the different MR-enterography sequences were analyzed with the Mc Nemar test for paired data [40]. A two-tailed P value of less than 0.05

AC

RESULTS

CE

was considered statistically significant.

Histopathological results

On the basis of the standard of reference, a total of 27 NETs of the small-bowel were confirmed in the 19 patients. Of these, 16 patients (16/19; 84%) had a single NET only, and three patients (3/19; 16%) had multiple NETs (i. e., two patients with 3 NETs each and one patient with 5 NETs). On the basis of pathological reports, all NETs were well differentiated tumors. They had a mean dimension of 13.1 mm ± 7.9 mm (SD) [range: 3 mm-40 mm]. Nine NETs were < 10 mm in size and 16 were ≥ 10 mm. All NETs developed in the ileum and 25/27 (93%) more specifically in the distal ileum. According to the TNM classification, NETs were classified T2 in 15/19 (79%) patients, T3 in 2/19 (11%) patients and T4 in 2/19

ACCEPTED MANUSCRIPT (11%) patients. All NETs were classified N1 for the lymph node status and M1 in 8/19 (42%) patients.

RI P

T

Sensitivity of individual MR sequences

All MR-enterography sequences were considered as having optimal image quality and

SC

adequate small bowel distension. No sequences were excluded from image analysis due to poor image quality.

NU

Individual MR-enterography sequence had a sensitivity ranging from 21% (4/19) to 95% (18/19) on a per-patient basis (Table 2). T1-weighted 3D VIBE MR-enterography

MA

sequence obtained after intravenous administration of gadolinium chelate had a sensitivity of 95% that was greater than those of HASTE (p = 0.00022), TrueFISP (p = 0.00022) and unenhanced T1-weighted 3D VIBE (p = 0.00012) MR-enterography sequences (Figs. 1, 2)

ED

(Table 2).

On a per-lesion basis, each individual MR-enterography sequence had a sensitivity

PT

ranging from 48 % (13/27) to 74% (20/27) (Table 2). T1-weighted 3D VIBE MRenterography sequence obtained after intravenous administration of gadolinium chelate

CE

allowed depiction of significantly more NETs than did HASTE (p = 0.03125), TrueFISP (p = 0.03125) and unenhanced T1-weighted 3D VIBE (p = 0.01562) MR-enterography sequences

AC

(Table 2).

Overall sensitivity of MR-enterography

On a per-patient basis, MR-enterography had an overall sensitivity of 95% (18/19; 95%CI: 74-100%) for the detection of NET similar to that of the T1-weighted 3D VIBE MRenterography sequence obtained after intravenous administration of gadolinium chelate alone (Table 2). On a per-lesion basis, overall sensitivity for NET detection was 74% (20/27; 95%CI: 54-89%). Regarding detection of NET ≥10 mm, the sensitivity was 94% (15/16; 95%CI: 70%-100%). Regarding detection of NET < 10 mm, the sensitivity was 45% (5/11: 95%CI:

ACCEPTED MANUSCRIPT 17%-77%). Seven NETs in three patients were not visible on MR-enterography; they had a mean diameter of 5.2 mm ± 2.5 (SD) [range: 3 - 15 mm]. Patient-by-patient and lesion-by-lesion comparison between the results obtained with

T

each individual unenhanced MR-enterography sequence and those obtained with the full

RI P

MR-enterography imaging set showed that all NETs visible on each individual unenhanced MR-enterography sequence were also visible on T1-weighted 3D VIBE sequence obtained

SC

after intravenous administration of gadolinium chelate.

NU

Imaging Presentation

MA

All visible NETs were located in the ileum. In 16/19 patients (84%) NETs presented as a well individualized intraluminal small bowel mass with intraluminal growth only that showed marked enhancement after intravenous administration of gadolinium chelate (Figs. 1,

ED

2). In 2/19 patients (11%) NETs presented as a focal thickening of the mesenteric aspect of the ileal wall with marked enhancement (Figs. 3, 4). The various MR-enterography findings

PT

observed in the 19 patients are reported in Table 3. Of these, mesenteric stranding (Figs. 3, 4, 5) and mesenteric mass were found in 15/19 (79%) and 15/19 (79%) patients (Figs. 2-5),

(Fig. 2).

CE

respectively. Vascular encasement by mesenteric mass was observed in 10/19 (53%) patients

AC

Mesenteric lymph nodes were observed in 18/19 patients (95%) and all showed enhancement after intravenous administration of gadolinium-chelate (Fig. 2). Enlarged retroperitoneal lymph nodes were observed in 1/19 patient (5%) (Fig. 5). Small bowel dilatation corresponding to the small bowel loops involved by NETs was observed in 2/19 patients (11%) (Fig. 4). No peritoneal nodules indicating peritoneal carcinomatosis were observed on MR-enterography, and this was further confirmed intraoperatively. Similarly, in no patient were internal calcifications within mesenteric masses. Hepatic metastases were visible in 8/19 patients (42%) (Fig. 4). Peritoneal fluid effusion was present in 1/19 patients (5%) (Fig. 4).

ACCEPTED MANUSCRIPT

DISCUSSION

T

In our retrospective study, we analyzed the MR-enterography features of 19 patients

RI P

with a total of 27 histopathologically confirmed NETs of the small-bowel. We found that MR-enterography shows highly suggestive features for the diagnosis of NET of the small-

SC

bowel and achieves high degrees of sensitivity for tumor detection on a per-patient basis. However, on a per-lesion basis, the sensitivity is lower, because NET may present as

NU

multifocal, tiny tumors in a subset of patients that are beyond the reach of MR imaging [27, 28, 41, 42]. Another result of our study is that gadolinium chelate enhanced fat-suppressed

MA

T1-weighted MR imaging provided best detection of NET of the small bowel by comparison with unenhanced sequences.

Previous comparative studies have reported that MR imaging used in conjunction

ED

with intravenous administration of gadolinium chelate achieved highest degrees of sensitivity for the depiction of NETs of the small bowel by comparison with unenhanced sequences,

PT

including diffusion-weighted MR imaging [29, 30, 27, 28]. One study that used MRenteroclysis reported that gadolinium chelate enhanced fat-suppressed T1-weighted gradient

CE

echo yielded best results for the detection of NET of the small bowel with only one missed tumor in 1/15 patient [30]. In the study by Bader et al., MR imaging failed to depict NET of

AC

the small bowel in 4/12 patients (33%) [29]. However, this study used MR imaging without small bowel distension so that direct comparison with our study may be difficult [29]. CT-enteroclysis is an important imaging modality for the detection of small bowel tumors, which conveys high degrees of sensitivity [14-16, 43]. In the elective area of NETs, CT-enteroclysis has demonstrated sensitivity that ranges between 86% and 100% [14, 15]. However, CT-enteroclysis is a radiating technique, so that a replacement nonradiating method is now a favored option. Our results suggest that MR-enterography can be considered as an alternate option in patients with suspected NET of the small bowel. Macroscopically, NETs of the small bowel generate small-bowel wall thickening and mesenteric stranding [3, 7, 18]. Presence of radiating strand-like lines originating from mesenteric nodule is an almost pathognomonic finding [7, 15, 19, 20]. They correlate to the

ACCEPTED MANUSCRIPT degree of fibrosis seen at histopathologic analysis [10]. Stranding was reported with variable frequencies ranging from 40% [11] to 82% [14]. Focal small bowel wall thickening may be the initial manifestation of NET [32, 44].

T

This finding was present in 11% of our patients and may be a clue when the tumor itself is

RI P

not visible [4, 29]. More rarely, wall thickening involving small bowel folds may be present and it is assumed that it indicates small bowel ischemia. This finding is more frequently

SC

observed in case of mesenteric vessel encasement by tumor [14, 45]. Additional findings including kinking of the small-bowel wall, luminal narrowing, fixation and angulation of the

NU

involved small bowel loop can be observed in association with large NETs [14, 18, 37]. In our study, no patients exhibited calcifications within mesenteric mass on MR-

MA

enterography. In theory, calcification may be seen as areas of signal void on MRenterography [37]. However, calcifications in NETs are often tiny so that they are better depicted with CT. In this regard, the reported prevalence of calcifications on CT ranges

ED

between 12.5% [8] and 70% [10], supporting the fact that calcifications are missed when using MR-enterography.

PT

All visible NETs of our study showed marked enhancement after intravenous administration of gadolinium chelate. This is consistent with the results of the study by

CE

Kamaoui et al. using CT-enteroclysis [15], those of Maselli et al. using MR-enteroclysis [23] and those of Amzallag et al. using MR-enterography [27, 28].

AC

NET of the small bowel can present as multifocal tumors in a subset of patients [41, 42]. In this regard, in our study three patients had multiple NETs of the small-bowel. Of interest, patients with multiple NETs often have tiny tumors, so that a number of them are not visible on MR-enterography [27, 28]; In our study, vascular involvement defined as either occlusion or narrowing of a vessel, usually with an associated soft-tissue mass surrounding the area of involvement was observed in 10/19 patients. It has been reported that collateral vessels can be observed in patients with NET and large mesenteric masses [7]. However, such collateral vessels were not observed in our series In our study, hepatic metastases were found in 8/19 (42%) of patients, which is a prevalence close to that reported by Kamaoui et al. who found hepatic metastases in 41% of their patients using CT-enteroclysis [15]. However, the prevalence of hepatic metastases may

ACCEPTED MANUSCRIPT reach up to 68% [9]. Hepatic metastases from NETs are often hypervascular but they usually show moderately intense enhancement [15, 36]. Large metastases may show heterogeneous enhancement due to central necrosis [6].

T

Our study has several limitations. The first is due to the fact that all our patients had

RI P

surgery so that only patients with resectable NETs were included and that patients with unresectable NETs were excluded. It is thus assumable that the MR imaging presentation we

SC

observed may be different in a more general population. A second limitation was the inclusion of patients with confirmed NET, so that the issue of specificity and accuracy was

NU

not addressed because of the absence of control subjects without NET. A third limitation is the retrospective design of the study, thus having introducing inclusion bias. A fourth

MA

limitation was the absence of comparison between MR-enterography with other imaging techniques. However, the goal of our study was to determine the sensitivity of MRenterography in the detection of NET of the small bowel using a strong standard of reference

ED

and investigate the respective sensitivities of each MR-enterography sequences. In conclusion, MR-enterography shows highly suggestive features for the diagnosis of

PT

NET of the small-bowel and yields high degrees of sensitivity for NET detection using a single enteric phase. However, further studies are needed to determine the added value of an

CE

arterial phase by comparison with a single enteric phase. Moreover prospective trials should be done to investigate the relative merits of MR-enterography for the identification of

AC

patients with suspected NET of the small bowel by comparison with other non-invasive imaging tests including videocapsule endoscopy [46].

ACCEPTED MANUSCRIPT

References

T

Scherübl H, Jensen RT, Cadiot G, Stölzel U, Klöppel G. Neuroendocrine tumors of the small

RI P

1.

bowels are on the rise: early aspects and management. World J Gastrointest Endosc 2010;2:325-334. Klöppel G, Perren A, Heitz PU. The gastroenteropancreatic neuroendocrine cell system and its

SC

2.

tumors: the WHO classification. Ann N Y Acad Sci 2004;1014:13-27.

Gore RM, Mehta UK, Berlin JW, Rao V, Newmark GM. Diagnosis and staging of

NU

3.

small bowel tumours. Cancer Imaging 2006;6:209-212.

Wong M, Kong A, Constantine S, Pathi R, Parrish FJ, Verma R, Lim C, Steer C.

MA

4.

Radiopathological review of small bowel carcinoid tumours. J Med Imaging Radiat Oncol. 2009;53:1-12.

Öksüz MÖ, Winter L, Pfannenberg C, Reischl G, Müssig K, Bares R et al. Peptide

ED

5.

receptor radionuclide therapy of neuroendocrine tumors with (90)Y-DOTATOC: is treatment

PT

response predictable by pre-therapeutic uptake of (68)Ga-DOTATOC? Diagn Interv Imaging 2014;95:289-300.

Pelage JP, Soyer P, Boudiaf M, Brocheriou-Spelle I, Dufresne AC, Coumbaras J,

CE

6.

Rymer R. Carcinoid tumors of the abdomen: CT features. Abdom Imaging 1999;24:240-245. Horton KM, Kamel I, Hofmann L, Fishman EK. Carcinoid tumors of the small bowel:

AC

7.

a multitechnique imaging approach. AJR Am J Roentgenol 2004;182:559-567. 8.

Picus D, Glazer HS, Levitt RG, Husband JE. Computed tomography of abdominal

carcinoid tumors. AJR Am J Roentgenol 1984;143:581-584. 9.

Sugimoto E, Lörelius LE, Eriksson B, Oberg K. Midgut carcinoid tumours: CT

appearance. Acta Radiol 1995;36:367-371. 10.

Pantongrag-Brown L, Buetow PC, Carr NJ, Lichtenstein JE, Buck JL. Calcification

and fibrosis in mesenteric carcinoid tumor: CT findings and pathologic correlation. AJR Am J Roentgenol 1995;164:387-391. 11.

Cockey BM, Fishman EK, Jones B, Siegelman SS. Computed tomography of

abdominal carcinoid tumor. J Comput Assist Tomogr 1985;9:38-42.

ACCEPTED MANUSCRIPT 12.

Woodard PK, Feldman JM, Paine SS, Baker ME. Midgut carcinoid tumors: CT

findings and biochemical profiles. J Comput Assist Tomogr 1995;19:400-405. 13.

Coulier B, Pringot J, Gielen I, Maldague P, Broze B, Ramboux A, Clausse M.

T

Carcinoid tumor of the small intestine: MDCT findings with pathologic correlation. JBR-

14.

RI P

BTR 2007;90:507-515.

Soyer P, Dohan A, Eveno C, Dray X, Hamzi L, Hoeffel C, et al. Carcinoid tumors of

15.

SC

the small-bowel: evaluation with 64-section CT-enteroclysis. Eur J Radiol 2013;82:943-950. Kamaoui I, De-Luca V, Ficarelli S, Mennesson N, Lombard-Bohas C, Pilleul F.

NU

Value of CT enteroclysis in suspected small-bowel carcinoid tumors. AJR Am J Roentgenol 2010;194:629-633.

Soyer P, Aout M, Hoeffel C, Vicaut E, Placé V, Boudiaf M. Helical CT-enteroclysis

MA

16.

in the detection of small-bowel tumours: a meta-analysis. Eur Radiol 2013;23:388-399. 17.

Theillac M, Jouvet JC, Boussel L. Meckel's diverticulum revealed by microcytic

18.

ED

anemia: the contribution of CT enteroclysis. Diagn Interv Imaging 2014;95:625-627. Soyer P, Boudiaf M, Fishman EK, Hoeffel C, Dray X, Manfredi R, et al. Imaging of

PT

malignant neoplasms of the mesenteric small bowel: new trends and perspectives. Crit Rev Oncol Hematol 2011;80:10-30.

Hristova L, Placé V, Nemeth J, Boudiaf M, Laurent V, Soyer P. Small bowel tumors:

CE

19.

spectrum of findings on 64-section CT enteroclysis with pathologic correlation. Clin Imaging

20.

AC

2012;36:104-112.

Levy AD, Sobin LH. From the archives of the AFIP: Gastrointestinal carcinoids:

imaging features with clinicopathologic comparison. Radiographics 2007;27:237-257. 21.

Fidler JL, Guimaraes L, Einstein DM. MR imaging of the small bowel.

RadioGraphics 2009;29: 1811-1825. 22.

Masselli G, Gualdi G. MR imaging of the small bowel. Radiology 2012;264:333-348.

23.

Masselli G, Polettini E, Casciani E, Bertini L, Vecchioli A, Gualdi G. Small-bowel

neoplasms: prospective evaluation of MR enteroclysis. Radiology 2009;251:743-750. 24.

Van Weyenberg SJ, Meijerink MR, Jacobs MA, Van der Peet DL, Van Kuijk C,

Mulder CJ, et al. MR enteroclysis in the diagnosis of small-bowel neoplasms. Radiology 2010;254:765-773.

ACCEPTED MANUSCRIPT 25.

Gourtsoyiannis N, Papanikolaou N, Grammatikakis J, Pras- sopoulos P. MR

enteroclysis: technical considerations and clinical applications. Eur Radiol 2002;12:26512658. Umschaden HW, Szolar D, Gasser J, Umschaden M, Haselbach H. Small-bowel

T

26.

RI P

disease: comparison of MR enteroclysis images with conventional enteroclysis and surgical findings. Radiology 2000;215: 717-725.

Amzallag-Bellenger E, Soyer P, Barbe C, Diebold MD, Cadiot G, Hoeffel C.

SC

27.

Prospective evaluation of magnetic resonance enterography for the detection of mesenteric

28.

NU

small bowel tumours. Eur Radiol 2013;23:1901-1910.

Amzallag-Bellenger E, Soyer P, Barbe C, Nguyen TL, Amara N, Hoeffel C.

MA

Diffusion-weighted imaging for the detection of mesenteric small bowel tumours with magnetic resonance- enterography. Eur Radiol 2014;24:2916-2926. 29.

Bader TR, Semelka RC, Chiu VC, Armao DM, Woosley JT. MRI of carcinoid

ED

tumors: spectrum of appearances in the gastrointestinal tract and liver. J Magn Reson Imaging 2001;14:261-269.

Schmid-Tannwald C, Zech CJ, Panteleon A, Sommer WH, Auernhammer C,

PT

30.

Herrmann KA. Characteristic imaging features of carcinoid tumors of the small bowel in MR

31.

CE

enteroclysis. Radiologe 2009;49:242-245. Azoulay R, Boudiaf M, Soyer P, Hamzi L, Abitbol M, Najmeh N, Rymer R.

AC

Carcinoid tumor of the small bowel: value of hydro-MR imaging for diagnosis. J Radiol. 2003;84(12 Pt 1):1982-1985. 32.

Boudiaf M, Jaff A, Soyer P, Bouhnik Y, Hamzi L, Rymer R. Small bowel diseases:

prospective evaluation of multidetector row helical CT enteroclysis in 107 consecutive patients. Radiology 2004;203:338-344. 33.

Soyer P, Boudiaf M, Dray X, Fargeaudou Y, Vahedi K, Aout M, et al. CT

enteroclysis features of uncomplicated celiac disease: retrospective analysis of 44 patients. Radiology 2009;253:416-424. 34.

Dorfman RE, Alpern MB, Gross BH, Sandler MA. Upper abdominal lymph nodes:

criteria for normal size determined with CT. Radiology1991;180:319-322. 35.

Boudiaf M, Soyer P, Terem C, Pelage JP, Maissiat E, Rymer R.. CT evaluation of

small bowel obstruction. Radiographics 2001;21:613-624.

ACCEPTED MANUSCRIPT 36.

Dromain C, de Baere T, Baudin E, Galline J, Ducreux M, Boige V et al. MR imaging

of hepatic metastases caused by neuroendocrine tumors: comparing four techniques. AJR Am J Roentgenol 2003;180:121-128. Masselli G, Colaiacomo MC, Marcelli G, Bertini L, Casciani E, Laghi F, et al. MRI

T

37.

RI P

of the small-bowel: how to differentiate primary neoplasms and mimickers. Br J Radiol 2012;85:824-837.

Soyer P, Boudiaf M, Sirol M, Dray X, Aout M, Duchat F, et al. Suspected

SC

38.

anastomotic recurrence of Crohn disease after ileocolic resection: evaluation with CT

39.

NU

enteroclysis. Radiology 2010;254:755-764.

Horton KM, Fishman EK. Volume-rendered 3D CT of the mesenteric vasculature:

MA

normal anatomy, anatomic variants, and pathologic conditions. Radiographics 2002;22:161172. 40.

Dwyer AJ. Matchmaking and McNemar in the comparison of diagnostic modalities.

41.

ED

Radiology 1991;178:328-330.

Nathan SR, Biernat L, Tang D. Multifocal small-bowel carcinoid tumor causing

PT

obscure recurrent gastrointestinal bleeding diagnosed by capsule endoscopy. Endoscopy 2007;39:E251-252.

Butte JM, Escobar C, O'Brien A, Zúñiga A. Multiple carcinoid tumors of the small

CE

42.

bowel report of one case. Rev Med Chil 2006;134:1306-1309. Khalife S, Soyer P, Alatawi A, Vahedi K, Hamzi L, Dray X, et al. Obscure

AC

43.

gastrointestinal bleeding: preliminary comparison of 64-section CT enteroclysis with video capsule endoscopy. Eur Radiol 2011;21:79-86. 44.

Anzidei M, Napoli A, Zini C, Kirchin MA, Catalano C, Passariello R. Malignant

tumours of the small intestine: a review of histopathology, multidetector CT and MRI aspects. Br J Radiol 2011;84:677-690. 45.

Mbengue A, Diallo I, Ndiaye AR, Keita I, Soko TO, Fall A, Diouf CT, Diakhate IC.

Acute intestinal ischaemia revealing a metastatic ileal endocrine tumour. Diagn Interv Imaging 2014;95:1109-1110. 46.

Soyer P. Obscure gastrointestinal bleeding: difficulties in comparing CT enterography

and video capsule endoscopy. Eur Radiol 2012;22:1167-1171.

ACCEPTED MANUSCRIPT Legends for Figures

Figure 1.

67-year-old woman who had MR-enterography because of tumor mass of the

T

terminal ileum found at video ileocolonoscopy. T1-weighted fat-suppressed three-

RI P

dimensional gradient echo (3D VIBE) MR-enterography image in the coronal plane obtained during the arterial phase of enhancement after intravenous administration of a gadolinium

SC

chelate shows hyperenhancing tumor mass (arrow) of the distal ileal loop. The lesion was not visible on the unenhanced MR sequences nor on HASTE and TruFISP sequences. There is

NU

no desmoplastic reaction. The lesion was surgically removed and histopathological analysis

Figure 2.

MA

confirmed a T2N1M0 well differentiated neuroendocrine carcinoma.

54-year-old woman who had MR-enterography because of tumor mass of the

terminal ileum found at video ileocolonoscopy. The lesion was surgically removed and

ED

histopathological analysis confirmed a T2N1M0 well differentiated neuroendocrine carcinoma.

PT

A, T1-weighted fat-suppressed three-dimensional gradient echo (3D VIBE) MRenterography image in the coronal plane obtained during the arterial phase of enhancement

CE

after intravenous administration of a gadolinium chelate shows hyperenhancing tumor mass (arrow) of the distal ileal loop. The lesion was not visible on the unenhanced MR sequences

AC

nor on HASTE and TruFISP sequences. B, At a different level, T1-weighted fat-suppressed 3D VIBE image in the coronal plane obtained during the arterial phase of enhancement after intravenous administration of a gadolinium chelate shows hyperenhancing mesenteric lymph nodes (arrowheads) and hyperenhancing mesenteric mass (arrow). Vessel encasement (curved arrow) by mesenteric mass is noted. C, TruFISP image in the coronal plane shows mesenteric lymph node (arrowhead) and mesenteric mass (arrow).

ACCEPTED MANUSCRIPT Figure 3.

77-year-old

man

who

had

MR-enterography

because

of

abnormal

videocapsule endoscopy findings raising suspicion for small bowel tumor. T1-weighted fatsuppressed three-dimensional gradient echo (3D VIBE) MR-enterography image in the

T

coronal plane obtained during the arterial phase of enhancement after intravenous

RI P

administration of a gadolinium chelate shows thickening of the mesenteric aspect of an ileal loop with marked enhancement (arrowheads) in association with mesenteric stranding

SC

(arrow). Histopathological analysis of resected portion of the ileum confirmed presence of

Figure 4.

NU

well differentiated T2N1M0 neuroendocrine carcinoma.

75-year-old man who had MR-enterography because of the detection of

MA

hypervascular hepatic metastases on computed tomography. TruFISP MR-enterography image in the coronal plane shows irregular thickening of the mesenteric aspect of a small bowel (arrowheads) in association with mesenteric stranding (curved arrows) and mesenteric

ED

mass (arrow). Intrahepatic metastases are visible (stars) along with perihepatic fluid effusion. The lesion was surgically removed and histopathological analysis confirmed a T3N1M0 well

78-year-old woman who had MR-enterography because of abnormal

CE

Figure 5.

PT

differentiated neuroendocrine carcinoma.

videocapsule endoscopy findings raising suspicion for small bowel tumor.

AC

A, T1-weighted fat-suppressed three-dimensional gradient echo (3D VIBE) MRenterography image in the transverse plane obtained after intravenous administration of a gadolinium chelate shows mesenteric stranding (arrows) adjacent to ileal loop and small mesenteric mass (arrowhead). No tumor is directly visible. B, HASTE image in the transverse plane shows enlarged retroperitoneal lymph node (arrow). C, T1-weighted fat-suppressed 3D VIBE image in the coronal plane obtained after intravenous administration of a gadolinium chelate shows mildly enhancing enlarged retroperitoneal lymph node (arrow). The lesion was surgically removed and histopathological analysis

confirmed

a

T2N1M0

well

differentiated

neuroendocrine

carcinoma.

Histopathological analysis of lymph node confirmed presence of metastatic involvement by neuroendocrine carcinoma.

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

Fig. 1

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Fig. 2A

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Fig. 2B

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Fig. 2C

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Fig. 3

AC

Fig. 4

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Fig. 5A

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Fig. 5B

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

Fig. 5C

ACCEPTED MANUSCRIPT Table 1. Imaging parameters used for MR-enterography in 19 patients with neuroendocrine tumors of the small bowel

(sec)

Sequence

Flip TR angle (°)

(ms)

TE

Matrix Section

Acceleration

T

TA

ETL

factor (ms)

size

Receiver bandwidth

RI P

MR

Thicknes s

(KHz/pixel)

T2 HASTE coronal

40

180 180

0 100 0

96

0

5

216x32 4

2

202

401

2

216

401

2

N.A.

601

3

N.A.

601

2

N.A.

490

2

N.A.

490

0

307x32

TrueFISP 70

14

chelate-enhanced

AC

T1-weighted 3D

Transverse Gd-

chelate enhanced

70

16.5

CE

22

Coronal Gd-

16

0

15

4

307x32

PT

TrueFISP coronal

3.59 1.8

ED

20

transverse

VIBE

96

MA

60

transverse

202x32

NU

100

T2 HASTE

SC

(mm)

3.65 1.83 0

4

4.68 2.38 240x32 3 0

3.41 1.5

291x44 2.4 4

T1-weighted 3D VIBE

Note. TA indicates time of acquisition; TR indicates repetition time; TE indicates time of excitation; ETL indicates echo train length; N.A. indicates not applicable

ACCEPTED MANUSCRIPT Table 2. Sensitivity of individual MR-enterography sequences in 19 patients with 27 histopathologically proven neuroendocrine tumors of the small bowel

Proportions (%)

HASTE sequence

5

TrueFISP sequence

5

Unenhanced 3D VIBE

4

95%CI (%)

RI P

T

Raw numbers

9 - 51

5/19 (26)

9 - 51

4/19 (21)

6 - 46

18/19 (95)

74 - 100

18

18/19 (95)

74 - 100

HASTE sequence

14

14/27 (52)

32 - 71

TrueFISP sequence

14

14/27 (52)

32 - 71

Unenhanced 3D VIBE

13

13/27 (48)

29 - 68

Gd-chelate enhanced 3D VIBE

20

20/27 (74)

54 - 89

Full set

20

20/27 (74)

54 - 89

Per-patient basis

AC

CE

PT

SC

NU

ED

Per-lesion basis

18

MA

Gd-chelate enhanced 3D VIBE Full set

5/19 (26)

Note Data are raw numbers, proportions, numbers in parenthesis are percentages, followed by 95% CIs. Gd indicates gadolinium; MRE indicates MR-enterography; 3D indicates three-dimensional.

ACCEPTED MANUSCRIPT Table 3. MR-enterography features in 19 patients with histopathologically proven neuroendocrine tumors of the small bowel

Visible small-bowel mass at MRE

16

T

Proportions (%)

RI P

Raw numbers

95%CI

16/19 (84)

60 - 97

2/19 (11)

1 - 33

0/19 (0)

0 - 18

16

16/19 (84)

60 - 97

16

16/19 (84)

60 - 97

15

15/19 (79)

54 - 94

3

3/19 (16)

3 - 40

0

0/19 (0)

0 - 18

15

15/19 (79)

54 - 94

18

18/19 (95)

74 - 100

Enlarged mesenteric lymph nodes

7

7/19 (37)

16 - 62

Retroperitoneal lymph nodes

1

1/19 (5)

0 - 26

10

10/19 (53)

29 - 76

Peritoneal nodules

0

0/19 (0)

0 - 18

Free-fluid effusion

1

1/19 (5)

0 - 26

15

15/19 (79)

54 - 94

Small-bowel dilatation

2

2/19 (11)

1 - 33

Hepatic metastases

8

8/19 (42)

20 - 67

2

Small-bowel mass with calcification

0

NU

SC

Small-bowel wall thickening (> 3 mm)

Hyperenhancing small-bowel mass

MA

Intraluminal growth

Hyperenhancing mesenteric mass

PT

Mesenteric mass with calcifications

ED

Mesenteric mass

CE

Mesenteric stranding

AC

Visible mesenteric lymph nodes

Vascular encasement

Small-bowel luminal narrowing (> 50%)

ACCEPTED MANUSCRIPT Note. - Numbers in parenthesis are percentages. CI indicates confidence interval. Results are given on a per-

AC

CE

PT

ED

MA

NU

SC

RI P

T

patient basis. MRE indicates MR-enterography.