The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184

Contents lists available at ScienceDirect

The Egyptian Journal of Radiology and Nuclear Medicine

journal homepage: www.sciencedirect .com/ locate /e j rnm

Original Article

Groove pancreatitis: Imaging features and management

http://dx.doi.org/10.1016/j.ejrnm.2016.09.0160378-603X/� 2016 The Egyptian Society of Radiology and Nuclear Medicine. Production and hosting by Elsevier.This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer review under responsibility of The Egyptian Society of Radiology andNuclear Medicine.⇑ Corresponding author at: Department of Diagnostic and Interven-

tional Radiology, Faculty of Medicine, Alexandria University Hospital,Champollion Street, El Azareeta, Egypt.

E-mail address: elnekidy@gmail.com (A. E.-A. M. El-Nekidy).

Abd El-Aziz Mohamed El-Nekidy a,⇑, Mohamed Eid Ibrahim a, Mohamed Saied Abdelgawad b,Rania A.M. Abouyoussef c, Ahmed Abdellatief Abdelkader d, Tarek Abdelhalim Elfaiomy d

aRadiodiagnosis Department, Alexandria Faculty of Medicine, EgyptbRadiodiagnosis Department, National Liver Institute, El-Menoufia University, Egyptc Tropical Medicine Department, Alexandria Faculty of Medicine, EgyptdDepartment of Surgery, Alexandria Faculty of Medicine, Egypt

a r t i c l e i n f o

Article history:Received 9 May 2016Revised 23 September 2016Accepted 26 September 2016Available online 29 September 2016

Keywords:Groove pancreatitisPancreatic cancerMRCP

a b s t r a c t

The aim of this retrospective study was to highlight the imaging findings of groove pancre-atitis (GP) as well as its management.Patients and methods: 16 patients diagnosed to have GP were enrolled in this work. Theincluded patients had complete records of the thorough clinical examination and labora-tory workup. All patients had been examined by multi-phase contrast enhanced MDCT tai-lored for pancreatic imaging. Six of these patients were additionally examined by MRIincluding MRCP.,Results: MDCT Multiple detector computed tomography of the 16 patients revealedthefollowing: (1) a hypodense sheet in the pancreaticoduodenal (PD) groove seen in 12patients with mild enhancement in the delayed phase seen in 6 of the them; (2)Duodenal wall thickening was seen in 10 patients while (3) associated cysts within theduodenal wall or in PD groove were seen in 6 patients; (4) and pancreatic head enlarge-ment in 8 patients. MRI of Six patients revealedthe following: (1) a T1 hypointense andT2 iso to hyperintense sheet at the PD groove in 4 patients with delayed enhancementin 3 of them; (2) Duodenal wall thickening with T2 high signal was seen in 6 patients whileassociated cysts within the duodenal wall were seen in 4 patients; (3) Pancreatic headenlargement seen in 4 patients; The MRCP of these patients showed dilated CBD with distaltapering and a distance separating its end from the duodenal wall.Conclusion: GP is a disease that should be considered in the list of differential diagnosis ofmasses implicating the pancreatic head and medial duodenal wall. Imaging findings thatare suggestive of GP include chronic inflammatory changes with fibrosis in the PD groovewith or without implication of the nearby head of the pancreas, duodenal medial muralthickening with luminal stenosis and cysts at the PD groove or within the duodenal wall.Vascular invasion is a sign against diagnosis of GP.� 2016 The Egyptian Society of Radiology and Nuclear Medicine. Production and hosting byElsevier. This is an open access article under the CC BY-NC-ND license (http://creativecom-

mons.org/licenses/by-nc-nd/4.0/).

1. Introduction

The pancreaticoduodenal (PD) groove is a potentialspace located between the head of the pancreas medially,second part of duodenum laterally, third part of duodenumand inferior vena cava posteriorly and the first part of the

1176 A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184

duodenum superiorly. This potential space contains lymphnodes, portion of the common bile duct, distal main pan-creatic duct, distal accessory pancreatic duct as well asthe major minor papilla. Small vessels are passing withinthis space, and the most important of these is the superiorPD artery [1].

Groove pancreatitis (GP) is a specific type of pancreati-tis originating in the PD groove. Sometime it involves thenearby head of the pancreas, second part of the duodenumand the common bile duct [2,3].

Several different terms have been used to describe thisinflammatory processes centered in the PD groove, includ-ing groove pancreatitis, periampullary duodenal wall cyst,myoadenomatosis, and cystic dystrophy of the duodenalwall. These disorders have clinically been grouped togetherand are termed ‘‘paraduodenal pancreatitis” [4]. Paraduo-denal pancreatitis was first described by Becker in 1973then Stolte et al. [2] introduced the term ‘‘groove pancre-atitis” for the same condition which is more used in theliterature.

The incidence of GP was reported in three different ser-ies of pancreaticoduodenectomy in patients with chronicpancreatitis, which was different. The percentage of occur-rence was 2.7%, 19.5% and 24.5% in the three different ser-ies respectively [2,5].

Patients affected by GP are usually adult males withalcohol abuse while GP has been described in females spo-radically [6].

2. Aim of the work

The purpose of this study was to describe the imagingfindings of GP as well as its management.

2.1. Patients

16 patients with a diagnosis of GP were enrolled in thisstudy. They included 12 males and 4 females, and theirages ranged between 35 and 73 years with the mean ageof 58 years. Inclusion criteria include those patients withthe final diagnosis of GP while those with the final diagno-sis of pancreatic groove carcinoma, peptic ulcer disease,duodenal cancer, ampullary cancer or pancreatic head can-cer were excluded.

2.2. Methods

This study is a retrospective study where patient con-sent was waived by the Research Ethics Board, assuringrespect of the confidentiality of the medical record. Wehave reviewed our medical records for the diagnosis ofGP during the period between January 2011 and October2015.

The included patients had complete records of the thor-ough clinical examination, laboratory workup includingroutine laboratory work as well as lipase, amylase CEAand CA19.9. All patients had been examined by multi-phase contrast enhanced MDCT tailored for pancreaticimaging including non-contrast, pancreatic, portal anddelayed phases. The machine used was Toshiba Aquilion

128-MDCT unit kV/effective mAs/rotation time(s):120 kV/225 eff. mAs/0.35 s; slice thickness 0.5 mm.

Non-ionic IV contrast was injected with a dose of1.5 ml/kg (maximum = 150 ml), with average rate of4 ml/s using automatic pump injector.

The pancreatic phase timing was fixed at 45 s, portalphase at 70 s, and delayed phase after 5 min from the startof contrast injection respectively, and examination wasdone using Siemens Emotion 6 and 64 MSCT. Scanningparameters were as follows: Volumetric High-spatial-frequency kernel algorithm; Slice thickness: 1–1.25 mm;Table speed for volumetric HRCT to enable the least cyclesof breath-holds as possible; Tube rotation: 0.6–0.9 s (mean0.75 s); Detector Collimation 1 mm; Helical mode (volu-metric HRCT); and kVp and mA per slice: 120–130 kVpand 200–400 mA, according to weight of the patient andclinical indication.

Six of these patients were additionally examined byMRI including MRCP, using a 1.5 T closed MRI imager(Avanto, Siemens, Erlangen, Germany). The pulsesequences used were transverse T2FSE with and withoutfat saturation, T1 chemical shift sequences (In/opposedphase), Dynamic pre- and post Gadolinium VolumetricInterpolated Breath-Hold Examination (VIBE) sequences,and MRCP sequences (thin slice 3D, as well as thick slabsingle shot).

Scanning parameters are as follows:

� Localizing T1-W gradient echo sequences were used.� Axial 2D T2-W turbo spin echo (HASTE/TSE) fat sup-pression sequence from the level of lower chest tomid abdomen level as finishing the whole liver span.TR1600, TE 70, flip angle 90, FOV 375, slice thickness7 mm, NSA 1 total scan time averaging 43 s.

� Axial 2D T2-W turbo spin echo (HASTE/TSE) fat sup-pression sequence with longer TE = 190. It was impor-tant to diagnose the degree of signal intensity of thelesion in long TE.

� In-phase and opposed phase (IP/OP) sequence TR 500,TE in-phase (2.2), TE opposed phase (4.4) FOV 375, flipangle 80, slice thickness 7 mm, NSA 2, average scanningtime 48 s. It was important in diagnosis of lesions con-taining intracellular fat as well as in diagnosis of focal orpatchy fatty infiltration or sparing.

� SSFP (BFFE in Philips) TR500, TE60, flip angle 60,FOV255, slice thickness 7 mm, NSA 1 with average scantime 22 s.

� DWI with variable b values 50–1000 s/mm2 with TR1000, TE 137, flip angle 90, FOV 370, slice thickness10 mm and average scan time 1.15 min. and automati-cally computer-generated ADC map.

� 3D T1 spoiled gradient fat-suppressed sequence with TR50, TE 500, FOV 355, flip angle 10, slice thickness 3 mmwith average scan time 19 s and this is repeated intriphasic study as HAP (15 s after the end of contrastinjection), PVP (70 s from the end of contrast injection)and delayed phase (about 3 min from the end of con-trast injection).

MRCP examinations were obtained with a single-shot,heavy T2W FSE sequence, HASTE (Siemens) by using respi-

A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184 1177

ratory gating and fat saturation which allows very rapidimage acquisition during brief breath hold and even it isso fast that allows imaging in patients who cannot holdbreath efficiently. This sequence includes the following:

� Thin section multisections moderate T2W FSEsequence: A TR of 1600 ms, an echo time (TE) of290 ms, a slice thickness = 2 mm, FOV = 380, aver-age = 2.0, flip angle, 140�, matrix size = 412 � 576 andturbo factor = 109.

� Thin section multisections heavy T2W FSE sequence: ATR of 2900 ms, an echo time (TE) of 700 ms, a slicethickness = 1 mm, FOV = 380, matrix size = 357 � 384average = 2.0, flip angle, 140�, and turbo factor = 135.

� A two dimensional thick single slab projectional image:A repetition time (TR) of 4500 ms, an echo time (TE) of700 ms, a slice thickness = 40 mm, distant factor = 50%,Field of view (FOV) = 350 mm, matrix size = 307 � 384,average = 1.0, flip angle, 180� and turbo factor = 256.The MIP images generated from the entire volume of athin section multisection data set resemble ERCPimages or thick slab images.

The management of patients with GP was either conser-vative in 6 patients or surgical with Whipple procedure in10 patients. Satisfactory improvement of the symptomswas noted in 4 out of 6 with conservative managementand 7 patients with surgical management. The diagnosiswas confirmed by histopathology in 12 patients (10 surgi-cal specimens and 2 FNAC), while clinical improvementand follow-up CT confirmed the diagnosis in the other 4patients.

The MRI and CT scans were analyzed to highlight imag-ing features of groove pancreatitis.

3. Results

3.1. Clinical picture (Table 1)

The patients presented with variable clinical picturesincluding epigastric pain in 10 patients, obstructive jaun-dice in 8 patients, vomiting in 8 patients, weight loss in 6patients, and diarrhea in 4 patients. Some of the patientspresented with more than one of above. Mild elevation ofamylase and lipase was reported in 8 and 6 patientsrespectively while elevated indirect bilirubin was reportedin 8 patients. None of the patients showed significant riseof the CEA, CA19.9 or other tumor markers.

Table 1Summary of the different positive and negative CT findings among the 16patients.

Clinical picture Number of patients

Epigastric pain referred to the back 10Vomiting 8Jaundice 8Weight loss 6Diarrhea 4

NB some patients may have more than one of the above.

3.2. Multiple findings were seen on the MDCT examinations ofthe 16 patients (Table 2) including

(1) A hypodense sheet at the PD groove was seen in 12patients with modest enhancement identified in delayedphase seen in 6 of the them. (2) Duodenal wall thickeningwas seen in 10 patients while associated cysts within theduodenal wall or in PD groove were seen in 6 patients.(3) Pancreatic head enlargement with diffuse enhance-ment was seen in 8 patients. (4) Mild pancreatic ductdilatation was seen in 8 patients while dilatation of theCBD was seen in 10 patients with distal tapering andintra-hepatic biliary dilatation. (5) None of the patientsshowed peri-pancreatic fluid collections, vascular invasionor occlusion, ascites, locoregional suspicious nodes orother stigmata of intra abdominal metastatic disease. Noneof our cases showed distended GB (Figs. 1–3).

3.3. Six patients out of the above described 16 patients hadMRI as well (Table 2), showing the following

(1) A T1 low signal and T2 iso to hyperintense signalsheet at the PD groove was seen in 4 patients with delayedenhancement in 3 of them. (2) Duodenal wall thickeningwith T2 high signal was seen in 6 patients while associatedcysts of T2 fluid signal within the duodenal wall were seenin 4 patients. (3) Pancreatic head enlargement with low T1signal alteration was seen in 4 patients. (4) Mild pancreaticduct dilatation was seen in 4 patients with pancreatic bodyand tail atrophy in 2 of them while dilatation of the CBDwas seen in 4 patients with distal tapering and intra-hepatic biliary dilatation. (5) None of the patients showedperi-pancreatic fluid collections, ascites, loco-regional sus-picious nodes or other stigmata of intra abdominal meta-static disease (Figs. 2 and 4).

The MRCP of these patients showed dilated CBD withdistal tapering and a distance separating its end from theduodenal wall in addition to fluid filled cysts at the duode-nal wall (Figs. 2 and 4) seen in 4 patients while the other 2patients had almost unremarkable MRCP.

4. Discussion

GP has two types: pure type or form that affects exclu-sively the PD groove with sparing of the pancreatic headand segmental type which is epicentered in groove withextension medially into the head of the pancreas. The dif-ferentiation between these two forms is not usually clear[7,2].

In our study pure form of the disease was seen in CTexaminations in 8 cases (50%) presenting with, the classicfindings of a hypo-dense sheet like soft tissue densitywithin the PD groove. Four of these showed retained con-trast represented by enhancement in the delayed phaserepresenting fibrous tissue. This delayed enhancement isexplained by hindered blood flow caused by fibrous tissuegrowth impeding the arterial flow due to arterial constric-tions [8].

The GP segmental type was seen in the other 8 patients(50%) where the sheet like focal hypodense lesions

Fig. 1. Multiphase MDCT; Axial (A and B) in the porto-venous phase,showing ill defined sheet of hypodensity at the duodeno-pancreaticgroove which showed delayed enhancement in delayed phase (C)(arrows). The medial wall of the second part duodenum is thickened(asterisk). Pure form of groove pancreatitis.

1178 A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184

extended into the pancreatic head in vicinity of the duode-nal wall with pancreatic head enlargement.

Similar findings were reported in previous studies withemphasis on the coronal reconstructions of MDCT data that

can allow better identification of ill-defined fat strandingand inflammatory changes in the PD groove accompaniedwith increasing delayed enhancement as a result of a sig-nificant fibrotic component [9].

The pure form is rather easy to identify. On the otherhand, the segmental form can be difficult to diagnose,because involvement of the groove is often obscured bymass like involvement of the pancreatic head. The segmen-tal type of GP is confused for a pancreatic head mass, anddifferentiating the two entities is not easy on the MRIand CT [10,11].

The main pancreatic ducts showed mild dilatation inseen at the body and tail of the pancreas in the 8 cases withsegmental type, while in the pure form of GP the main pan-creatic ducts were not dilated. It is also reported that pan-creatic duct can also be narrowed toward the head of thepancreas in a smooth gradual pattern. In more chronicstage, pancreatic parenchymal changes resembling thoseof ordinary chronic pancreatitis can develop including pan-creatic calcifications, ductal dilatation, and ductal beadingor irregularity [9,11].

Other important findings were also noted in our seriesincluding focal duodenal wall thickening seen in 10 cases(62.5%) and cysts in the duodenal wall itself or in groovebetween the pancreatic head and the duodenum in 6 cases(37.5%). The cysts were variable in size and number rang-ing from tiny to large even multi-locular cystic mass likelesion was seen in a single case. Appreciating medial duo-denal wall thickening is easier also on the coronal images[8].

Other authors found that these multiple cysts suggestcystic dystrophy in heterotopic pancreatic islands withinthe duodenal wall. They also claimed that heterotopic pan-creatic tissues could not be identified on CT until they gotinflamed with cystic changes [12].

The pathology of GP in the literature addressed that theduodenal mucosa between the major and minor papillae ismarkedly thickened. The involved areas show gelatinouscontents, edema and fibrosis with possible cyst formation.The cysts may contain small calculi. Microscopic evalua-tion reveals duodenal wall thickening, with glandular andmuscular inflammation and hyperplasia. Sometimes pan-creatic islands heterotopia can be seen. The scarring impli-cates the lower portion of the common bile duct in the PDgroove [13–16].

The origin of the cysts in GP is controversial, and themost popular theory claims that they are cystic dystrophyof the pancreatic heterotopic islands in the duodenal wall[17]. Other authors suggested that these cysts may bedilated Santorini duct branches [18].

The cause of GP is still controversial including coexist-ing biliary disorder, gastric ulcers, and heterotopic pancre-atic tissues or disturbed pancreatic fluid flow through theduct of Santorini [2,5]. The specific location of the lesionsaround the minor papilla suggests possible anatomical orfunctional disorder related to this area as GP may occurin cases with pancreas divisum, absent or narrow duct ofSantorini [19], or may be due to obstruction of the acces-sory pancreatic duct [17,19].

An important point noticed in our series and reportedpreviously, is that even in severe GP, the surrounding ves-

Fig. 2. Multiphase MDCT Axial (A and B) in late arterial phase, showing ill defined sheet of hypodensity at the duodeno-pancreatic groove with partialextension into the pancreatic head. The corresponding Axial T2 WI (C&D) done two weeks later showed similar sheet of tissues expressing mild T2hyperintense signal as well as mild duodenal wall thickening and enlarged pancreatic head. Coronal single shot MIP (E), showing the intra and extra-hepaticbiliary dilatation with distal tapering of the CBD as well as a relatively gapping distance between its end and the 2nd part of the duodenum. Tiny cysticchanges in the duodenal wall seen (C&E). Segmental form of groove pancreatitis.

A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184 1179

sels are spared without thrombosis or infiltration [8,10,20].Also in our series we have not detected CT signs of acutepancreatitis or loco regional metastasis or adenopathies.The previous studies reported rarity to visualize fluid inthe para-renal spaces or surrounding the pancreas [10,11].

In our study CBD dilatation and distal smooth taperingwere seen in 10 patients (62.5%) including all the segmen-tal types and 2 of the pure type leading to intra- and extra-hepatic biliary system dilatation.

Previous studies also found that distal common bileduct can appear attenuated and narrowed in both pureand segmental types of GP. This was better assessed onthe coronal multiplanar reconstruction (MPR).

In most cases, this narrowing was relatively smooth,tapered, and regular, without evidence of shouldering orirregularity [9,11].

As a fact, CT provides superior spatial resolution, butwith less contrast resolution to discriminate pancreaticcancer from inflammation. The high soft-tissue resolutionof MRI provides more accurate evaluation of the pancreatictissues, specifically for tissue characterization in inflamma-tory and neoplastic processes and analysis of contents ofcysts [21–24].

MRI and MRCP were available in 6 patients in our study.There was a CT similarity regarding the sheet of tissueswithin the pancreaticoduodenal groove. These were seen

Fig. 3. Multiphase MDCT; Axial (A) and Coronal (B) porto-venous phase,showing ill defined sheet of hypodensity; coupled with cystic changes atthe duodeno-pancreatic groove. The medial wall of the second partduodenum is thickened. Partial extension to the pancreatic head. Mildsegmental form of groove pancreatitis.

1180 A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184

expressing T1 hypo-intense and T2 slightly hyperintensesignal in 3 patients with depiction of mild enhancementin the delayed phases in three of them (50%). These casesrepresented the pure type of GP, while in the segmentalform there was associated pancreatic head enlargementin the other 3 patients.

Involvement of the pancreas was reported to be wellvisualized on MRI compared to CT, with progressive loss

Table 2Summary of the different positive and negative CT and MRI findings among the p

CT MRI

Finding No. patients % Findi

Hypodense sheet 12 75 HypoDuodenal wall thickening 10 62.5 DuodCBD dilatation and distal tapering 10 62.5 CBDPancreatic head enlargement 8 50 PancrPancreatic duct dilatation 8 50 PancrDelayed enhancement 6 37.5 DelayDuodenal cysts 6 37.5 DuodCollections and stranded mesentery 0 0 ColleLocoregional lymph nodes 0 0 LocorIntra-abdominal metastasis 0 0 IntraVascular invasion 0 0 VascuGB distension 0 0 GB d

of T1 signal intensity in the head of the pancreas as a resultof parenchymal atrophy and fibrosis [25,26].

Irie et al. [7] and Blasblag et al. [26] also reported sim-ilar MRI findings in GP explaining the T2 iso to hyperin-tense signal variation of the lesion in the PD groove or inthe head of the pancreas head to match with stage of thedisease, as the subacute phase disease shows higher T2 sig-nal due to edema, and chronic phase has darker T2 signaldue to evolution of fibrosis.

Duodenal wall thickening was seen in the MRI of all the6 patients with 4 of them showing mural cysts. The medialwall of duodenum is involved in the pure as well as thesegmental forms of GP, with multiple T2 hyperintensecysts in both the duodenal wall and PD groove.

Blasblag et al. [26] described similar changes in theirstudy. Also they described peripheral enhancement in thepancreatic and portal phases with progressive fill in thedelayed phases. This delayed enhancement was reportedalso in the related thickened medial wall of the duodenum.This delayed enhancement reflects presence of fibrosis[25].

In our study pancreatic head enlargement with low T1signal alteration was seen in 4 patients with pancreaticduct dilatation in addition to atrophy of the rest of pan-creas. This reflects chronic inflammatory disease withfibrous tissues replacing the glandular tissues of the pan-creas. In the pure type of GP, the pancreas appears normaland shows relatively high T1 signal [26].

In our study duodenal wall thickening with T2 high sig-nal was seen in 6 patients while associated small cysts ofT2 fluid signal within the duodenal wall were seen in 4patients. MRCP facilitates determination of relationshipbetween these cysts and CBD and pancreatic ducts [26].

Marked duodenal wall thickening is usually not associ-ated with pancreatic neoplastic processes while it is com-mon in GP [26]. This sign can help in differentiating GPfrom pancreatic cancers.

In our study MRCP was available in 6 patients, showingdilated CBD with distal tapering with a distance separatingdistal ends from the duodenal walls, fluid filled cysts at theduodenal wall or in the groove with ectatic pancreaticducts in 4 patients, while the other 2 patients had almostunremarkable MRCP.

atients with GP.

ng No. patients %

intense sheet 4 66.6enal wall thickening with high T2 signal 6 100dilatation and distal tapering 4 66.6eatic head enlargement 3 50eatic duct dilatation 4 66.6ed enhancement 3 50enal cysts 4 66.6ctions and stranded mesentery 0 0egional lymph nodes 0 0-abdominal metastasis 0 0lar invasion 0 0istension 0 0

Fig. 4. Axial T1VIBE late arterial phase (A) showing ill defined enhancing sheet with cystic changes at the duodeno-pancreatic groove (green arrows).Delayed phase (B) shows retained contrast enhancement (green arrows). Axial T2W (C) and coronal single shot MRCP (D), showing duodenal wallthickening and cystic changes (green arrows C&D) at the medial wall of the second part duodenum (hence the name: Cystic Dystrophy of Ectopic Pancreas).The pancreatic head is enlarged. Segmental form of groove pancreatitis.

A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184 1181

Blasblag et al. [26] described similar smooth distaltapering which was regular without shoulder sign orabrupt interruption of the ducts in cancers.

On MRCP, the distance separating distal pancreatic andCBD from duodenal lumen is due to inflammatory lesion inthe PD groove and the duodenal wall marked thickening.This is common in GP and not in pancreatic neoplasticlesions [25,26].

The main pancreatic duct is usually not dilated in thepure form of GP, while the segmental form shows stricturewithin the pancreatic head that is usually longer thanthose associated with pancreatic neoplasms. Also GPshowed milder upstream dilatation of the pancreatic ductsin the rest of the pancreatic body and tail [27].

None of our cases showed abnormal dilatation of theGB. The GB tends to be normally distended in GP. Previousstudies described Banana like gallbladder in cases of GPsimulating those seen in traditional chronic pancreatitis[26].

MRCP is valuable in diagnosis of GP as it yields diagnos-tic data more or less similar to ERCP. However, in GP, duo-denal stenosis often hinders ERCP [25]. ERCP is limited tovisualization of a tapered lower bile duct, which can some-

times be difficult to differentiate GP smooth long stricturefrom irregular strictures in malignancies [13].

The appearance of GP with both transabdominal andendoscopic ultrasound is not well detailed in the literature.In the early stages with more inflammatory component, U/S may show hypoechoic band like thickening of the PDgroove and thickening of the adjacent duodenum with orwithout hypoechoic heterogeneous pancreatic head. Inthe chronic stages, the echogenicities of all these lesionsbecome hyper as fibrosis dominates over inflammation[28]. It is common to visualize regular narrowing of theCBD and the Santorini duct on endoscopic ultrasound [29].

Fine needle biopsy even those guided by endoscopic USare challenging to pathologists. The relatively small samplevolume may not be adequate to exclude the presence ofmalignant cells [3,16,30]. Even Fibrosis diagnosed bypathologists does not exclude neoplastic changes. as pan-creatic adenocarcinoma may show desmoplastic reactionsimulating the fibrotic changes in chronic inflammatoryabnormality such as GP putting down the diagnostic meritof the Fine needle biopsy in such cases [31,32].

Similar to our results, previous studies reported limitedvalue of laboratory markers in diagnosis of GP because

1182 A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184

bilirubin, alkaline phosphatase, lipase and amylase areoften normal or just minimally elevated. Pancreatic tumormarkers (CEA and CA-19-9) are usually negative in GP.These negative biomarkers may suggest GP rather thanneoplastic process [6,33].

The most challenging differential diagnosis of GP (espe-cially its segmental form) is from pancreatic head adeno-carcinoma and malignancies which arise adjacent to thePD groove and do not show the typical pancreatic doubleduct cutoff and upstream atrophy. The management plansof the GP and pancreatic cancer two are significantly differ-ent [6]. This differentiation can be impossible, and manypatients may undergo Whipple procedure because of lackof this preoperative discrimination [5]. Even more this dif-ferentiation is important in optimizing therapeutic deci-sions, including the decision of whether or not to usepreoperative chemotherapy if GP is excluded [34].

Unlike GP, most pancreatic adenocarcinomas do notshow internal cystic change and are much more likely toinfiltrate posteriorly into the retroperitoneum and encasethe vasculature. Moreover, thickening of the medial duode-nal wall, a common finding with GP, is uncommon withpancreatic adenocarcinoma. Enhancement pattern for GPtends to be more patchy and heterogeneous with delayedcontrast retained while pancreatic adenocarcinoma usuallyshows more homogeneous hypodensity [25,26].

Kalb et al. [34] reported that contrast-enhanced MRimaging may help accurately differentiating GP from pan-creatic cancer when using suggested 3 diagnostic signsfor GP including mural duodenal thickening, delayedenhancement of the second part of the duodenum; andcysts seen within duodenal wall or PD groove. They foundcorrect diagnosis of GP was achieved with accuracy of87.2% while exclusion of cancer can had a negative predic-tive value of 92.9%.

The differentiation between GP and scirrhous adenocar-cinoma invading the groove is difficult on CT and MR imag-ing. Both produce similar T1 low signal with or withoutdelayed enhancement [35–37]. However, carcinomas oftenhave more discrete and round configuration while fibrosisis more diffuse and ill defined [26]. The presence of vascu-lar invasion is highly suggestive of pancreatic carcinomaand not reported in GP [38,35].

Furthermore, it is more difficult to differentiatebetween GP and carcinomas arising in the PD groove,because they may be associated with duodenal muralthickening and stenosis. Duodenal biopsies may correctlyreach a diagnosis [35].

Differentiating GP from acute pancreatitis is rather easydue to fluid collections and inflammation in the PD groovethat evolves rapidly on serial follow-up imaging andshould usually resolve later on, whereas the imaging find-ings associated with GP often persist. Acute edematouspancreatitis, involves a large portion of the pancreatic par-enchyma, and not only epicentered in the groove withperipancreatic fluid and inflammation tracking into thepararenal spaces, while GP typically shows little retroperi-toneal inflammation or fluid, and even in the segmentalform, involvement of the pancreas is usually limited tothe pancreatic head. Elevated lipase level is also an impor-tant differentiating marker [39].

Differential diagnosis of GP from chronic pancreatitiswith acute with pseudocysts within the duodenal wall israther easy because the later has no mural duodenal thick-ening or luminal stenosis [40].

Pure form of GP should be differentiated from otherconditions including duodenal cancer as well as distalCBD and ampullary carcinomas. The later produces focalmalignant lesions at the ampulla, while GP is more ill-defined crescentic soft tissue, still larger ampullary carci-nomas may not be easily differentiated from GP on imag-ing basis [40].

MRCP can help in differentiating GP from cholangiocar-cinomas involving the CBD, since GP shows a longersmooth CBD stenosis or tapering compared to the irregularnarrowing, shouldering and abrupt termination of the CBDin cholangiocarcinomas [40].

Carcinoids and gastrinomas may rarely originate withinthe PD groove. These tumors show early hyper-enhancement due to high vascularity and hyperintenseT2 signal compared to the delayed centripetal enhance-ment of GP. Duodenal gastrointestinal stromal tumor(GIST) is more hypodense lesion and still more hyper-vascular and should not be easily confused with GP [11].

Acute phase of GP is conservatively treated using bedrest, analgesics and intra-venous nutrition. Most patientsimprove with this conservative treatment, but somepatients suffer from relapses of acute pancreatitis espe-cially those with an anatomic or functional disturbanceof pancreatic duct system [41].

Sometimes GP resists medical treatment and surgicaltreatment may be inevitable at late stages of the diseaseas the patient may develop marked duodenal stenosisendocrine and exocrine pancreatic failure, or extensivefibrotic changes causing severe pain. Those patientsusually benefit from surgery with symptomatic relief[42,43]. A pancreaticoduodenectomy using the Whippleprocedure or less frequently a pylorus-preserving pancrea-ticoduodenectomy are the surgical procedures of choice ingroove pancreatitis [6]. Both treatment plans were presentin our cases with satisfactory improvement of the symp-toms that were noted in 4 out of 6 with conservative man-agement and 7 out of 10 patients with surgicalmanagement.

Another option was reported as a promising treatmentfor GP including ERCP drainage of the duct of Santorini[44].

Radiological suggestion of the suspicion or suggestionthat a pancreatic head lesion may represent GP shoulddirect the surgeons for further workup before the decisionof radical surgery. On the other hand, the diagnosis ofgroove pancreatitis should not be confirmed until the otherpossibility of adenocarcinoma is carefully excluded.

Limitations of our study include retrospective design,and we limited the scope of our study to describe the imag-ing features of the documented cases of GP with lack ofcomparison with cases of pancreatic head cancer thatmay simulate GP. Another limitation is that not all caseshave MRI and MRCP examinations and we did not providecomparison between CT and MRI.

A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184 1183

5. Conclusion

GP is a disease that should be considered in the list ofdifferential diagnosis of masses implicating the pancreatichead and medial duodenal wall. GP has two types: puretype that affects exclusively the PD groove and segmentaltype which is epicentered in groove with extension medi-ally into the head of the pancreas. Imaging findings that aresuggestive of GP include chronic inflammatory changeswith fibrosis in the PD groove with or without implicationof the nearby head of the pancreas, duodenal medial muralthickening with luminal stenosis and cysts at the PDgroove or within the duodenal wall. Vascular invasion isa sign against diagnosis of GP. Sometime the differentia-tion between the two forms of GP is not usually clear.

Conflict of interest

The authors declared that there are no conflict ofinterests.

References

[1] Hernandez-Jover D, Pernas JC, Gonzalez-Ceballos S, Lupu I, MonillJM, Pérez C. Pancreatoduodenal junction: review of anatomy andpathologic conditions. J Gastrointest Surg 2011;15(7):1269–81.

[2] Stolte M, Weiss W, Volkholz H, Rösch W. A special form of segmentalpancreatitis: groove pancreatitis. Hepatogastroenterology 1982;29(5):198–208.

[3] Adsay N, Zamboni G, editors. Paraduodenal pancreatitis: a clinico-pathologically distinct entity unifying ‘‘cystic dystrophy ofheterotopic pancreas,”‘‘para-duodenal wall cyst,” and ‘‘groovepancreatitis. Seminars in diagnostic pathology, vol. 21. Elsevier;2004. p. 247–54.

[4] Pezzilli R, Santini D, Calculli L, Casadei R, Morselli-Labate AM,Imbrogno A, et al. Cystic dystrophy of the duodenal wall is notalways associated with chronic pancreatitis. World J Gastroenterol:WJG 2011;17(39):4349.

[5] Yamaguchi K, Tanaka M. Groove pancreatitis masquerading aspancreatic carcinoma. Am J Surg 1992;163(3):312–6.

[6] Balakrishnan V, Chatni S, Radhakrishnan L, Narayanan VA, Nair P.Groove pancreatitis: a case report and review of literature. JOP2007;8(5):592–7.

[7] Irie H, Honda H, Kuroiwa T, Hanada K, Yoshimitsu K, Tajima T, et al.MRI of groove pancreatitis. J Comput Assist Tomogr 1998;22(4):651–5.

[8] Itoh S, Yamakawa K, Shimamoto K, Endo T, Ishigaki T. CT findings ingroove pancreatitis: correlation with histopathological findings. JComput Assist Tomogr 1994;18(6):911–5.

[9] Kwak SW, Kim S, Lee JW, Lee NK, Kim CW, Yi MS, et al. Evaluation ofunusual causes of pancreatitis: role of cross-sectional imaging. Eur JRadiol 2009;71(2):296–312.

[10] Perez-Johnston R, Sainani NI, Sahani DV. Imaging of chronicpancreatitis (including groove and autoimmune pancreatitis).Radiol Clin North Am 2012;50(3):447–66.

[11] Shanbhogue AK, Fasih N, Surabhi VR, Doherty GP, Shanbhogue DK,Sethi SK. A clinical and radiologic review of uncommon types andcauses of pancreatitis. Radiographics: a review publication of theRadiological Society of North America, Inc. 2009 Jul-Aug;29(4):1003–26.

[12] Vullierme M-P, Vilgrain V, Fléjou J-F, Zins M, O’Toole D, RuszniewskiP, et al. Cystic dystrophy of the duodenal wall in the heterotopicpancreas: radiopathological correlations. J Comput Assist Tomogr2000;24(4):635–43.

[13] Malde DJ, Oliveira-Cunha M, Smith AM. Pancreatic carcinomamasquerading as groove pancreatitis: case report and review ofliterature. JOP: J Pancreas 2011;12(6):598–602.

[14] Kim JD, Han YS, Choi DL. Characteristic clinical and pathologicfeatures for preoperative diagnosed groove pancreatitis. J KoreanSurg Soc 2011;80(5):342–7.

[15] Levenick JM, Gordon SR, Sutton JE, Suriawinata A, Gardner TB. Acomprehensive, case-based review of groove pancreatitis. Pancreas2009;38(6):e169–75.

[16] Adsay NV, Bandyopadhyay S, Basturk O, Othman M, Cheng JD,Klöppel G, et al., editors. Chronic pancreatitis or pancreatic ductaladenocarcinoma? In: Seminars in diagnostic pathology, vol. 21;2004. p. 268–76.

[17] Chatelain D, Vibert E, Yzet T, Geslin G, Bartoli E, Manaouil D, et al.Groove pancreatitis and pancreatic heterotopia in the minorduodenal papilla. Pancreas 2005;30(4):e92–5.

[18] Holstege A, Barner S, Brambs H, Wenz W, Gerok W, Farthmann E.Relapsing pancreatitis associated with duodenal wall cysts.Gastroenterology 1985;88(3):814–9.

[19] Shudo R, Obara T, Tanno S, Fujii T, Nishino N, Sagawa M, et al.Segmental groove pancreatitis accompanied by protein plugs inSantorini’s duct. J Gastroenterol 1998;33(2):289–94.

[20] Raman SP, Salaria SN, Hruban RH, Fishman EK. Groove pancreatitis:spectrum of imaging findings and radiology-pathology correlation.AJR: Am J Roentgenol 2013 Jul;201(1):W29–39.

[21] Kalb B, Sarmiento JM, Kooby DA, Adsay NV, Martin DR. MR Imagingof Cystic Lesions of the Pancreas 1. Radiographics: a reviewpublication of the Radiological Society of North America, Inc.2009;29(6):1749–65.

[22] Morgan DE, Baron TH, Smith JK, Robbin ML, Kenney PJ. Pancreaticfluid collections prior to intervention: evaluation with MR imagingcompared with CT and US. Radiology 1997;203(3):773–8.

[23] Song SJ, Lee JM, Kim YJ, Kim SH, Lee JY, Han JK, et al. Differentiationof intraductal papillary mucinous neoplasms from other pancreaticcystic masses: Comparison of multirow-detector CT and MR imagingusing ROC analysis. J Magn Reson Imaging 2007;26(1):86–93.

[24] Waters JA, Schmidt CM, Pinchot JW, White PB, Cummings OW, PittHA, et al. CT vs MRCP: optimal classification of IPMN type andextent. J Gastrointest Surg 2008;12(1):101–9.

[25] Castell-Monsalve FJ, Sousa-Martin JM, Carranza-Carranza A. Groovepancreatitis: MRI and pathologic findings. Abdom Imaging 2008;33(3):342–8.

[26] Blasbalg R, Baroni RH, Costa DN, Machado MC. MRI features ofgroove pancreatitis. AJR: Am J Roentgenol 2007 Jul;189(1):73–80.

[27] Manfredi R, Costamagna G, Brizi MG, Spina S, Maresca G, Vecchioli A,et al. Pancreas divisum and ‘‘Santorinicele”: diagnosis with dynamicMR cholangiopancreatography with secretin stimulation 1.Radiology 2000;217(2):403–8.

[28] Wronski M, Karkocha D, Slodkowski M, Cebulski W, KrasnodebskiIW. Sonographic findings in groove pancreatitis. J Ultrasound Med2011;30(1):111–5.

[29] Tio T, Luiken G, Tytgat G. Endosonography of groove pancreatitis.Endoscopy 1991;23(5):291–3.

[30] Klöppel G. Chronic pancreatitis and the differential diagnosis versuspancreatic cancer. Arch Pathol Lab Med 2009;133(3):382.

[31] Brandt K, Charboneau J, Stephens D, Welch T, Goellner J. CT-and US-guided biopsy of the pancreas. Radiology 1993;187(1):99–104.

[32] Chute DJ, Stelow EB. Fine-needle aspiration features of paraduodenalpancreatitis (groove pancreatitis): a report of three cases. DiagnCytopathol 2012;40(12):1116–21.

[33] Triantopoulou C, Dervenis C, Giannakou N, Papailiou J, PrassopoulosP. Groove pancreatitis: a diagnostic challenge. Eur Radiol 2009 Jul;19(7):1736–43.

[34] Kalb B, Martin DR, Sarmiento JM, Erickson SH, Gober D, Tapper EB,et al. Paraduodenal pancreatitis: clinical performance of MR imagingin distinguishing from carcinoma. Radiology 2013 Nov;269(2):475–81.

[35] Gabata T, Kadoya M, Terayama N, Sanada J, Kobayashi S, Matsui O.Groove pancreatic carcinomas: radiological and pathologicalfindings. Eur Radiol 2003;13(7):1679–84.

[36] Furukawa H, Takayasu K, Mukai K, Kanai Y, Inoue K, Kosuge T, et al.Late contrast-enhanced CT for small pancreatic carcinoma: delayedenhanced area on CT with histopathological correlation.Hepatogastroenterology 1995;43(11):1230–7.

[37] Johnson PT, Outwater EK. Pancreatic carcinoma versus chronicpancreatitis: dynamic MR imaging 1. Radiology 1999;212(1):213–8.

[38] Uchida E, Aimoto T, Nakamura Y, Katsuno A, Chou K, Kawamoto M,et al. Clinicopathologic study on pancreatic groove carcinoma.Pancreas 2006;33(4):504.

[39] Yu J, Fulcher AS, Turner MA, Halvorsen RA. Normal anatomy anddisease processes of the pancreatoduodenal groove: imagingfeatures. Am J Roentgenol 2004;183(3):839–46.

[40] Robinson P, Sheridan M. Pancreatitis: computed tomography andmagnetic resonance imaging. Eur Radiol 2000;10(3):401–8.

1184 A. E.-A. M. El-Nekidy et al. / The Egyptian Journal of Radiology and Nuclear Medicine 47 (2016) 1175–1184

[41] Shudo R, Yazaki Y, Sakurai S, Uenishi H, Yamada H, Sugawara K, et al.Groove pancreatitis: report of a case and review of the clinical andradiologic features of groove pancreatitis reported in Japan. InternMed 2002;41(7):537–42.

[42] Toshihide A, Hidehiko Y, Satoshi I. A case of groove pancreatitisresisted conservative therapy and demanded a pylorus preservingpancreatoduodenectomy. J Surg 2001;46(2):35–9.

[43] Rahman SH, Verbeke CS, Gomez D, McMahon MJ, Menon KV.Pancreatico-duodenectomy for complicated groove pancreatitis.Hpb 2007;9(3):229–34.

[44] Isayama H, Kawabe T, Komatsu Y, Sasahira N, Toda N, Tada M, et al.Successful treatment for groove pancreatitis by endoscopic drainagevia the minor papilla. Gastrointest Endosc 2005;61(1):175–8.