ORIGINAL ARTICLE

Perforated appendicitis: an underappreciated mimicof intussusception on ultrasound

Beverley Newman & Matthew Schmitz & Rakhee Gawande &Shreyas Vasanawala & Richard Barth

Received: 17 October 2013 /Accepted: 6 January 2014 /Published online: 26 January 2014# Springer-Verlag Berlin Heidelberg 2014

AbstractBackground We encountered multiple cases in which the USappearance of ruptured appendicitis mimicked intussuscep-tion, resulting in diagnostic and therapeutic delay and multipleadditional imaging studies.Objective To explore the clinical and imaging discriminatoryfeatures between the conditions.Materials and methods Initial US images in six children (age16 months to 8 years; 4 boys, 2 girls) were reviewed indepen-dently and by consensus by three pediatric radiologists. Thesefindings were compared and correlated with the original re-ports and subsequent US, fluoroscopic, and CT images andreports.Results All initial US studies demonstrated a multiple-ring-like appearance (target sign, most apparent on transverseviews) with diagnostic consensus supportive of intussuscep-tion. In three cases, US findings were somewhat discrepantwith clinical concerns. Subsequently, four of the six childrenhad contrast enemas; two were thought to have partial orcomplete intussusception reduction. Three had a repeat USexamination, with recognition of the correct diagnosis. Noneof the US examinations demonstrated definite intralesionallymph nodes or mesenteric fat, but central echogenicitycaused by debris/appendicolith was misinterpreted as fat. Allshowed perilesional hyperechogenicity that, in retrospect, rep-resented inflamed fat walling off of the perforated appendix.There were four CTs, all of which demonstrated a double-ringappearance that correlated with the US target appearance, with

inner and outer rings representing the dilated appendix andwalled-off appendiceal rupture, respectively. All six childrenhad surgical confirmation of perforated appendicitis.Conclusion Contained perforated appendicitis can produceUS findings closely mimicking intussusception. Clinical cor-relation and careful multiplanar evaluation should allow forsonographic suspicion of perforated appendicitis, which canbe confirmed on CT if necessary.

Keywords Children . Appendicitis . Intussusception .

Ultrasound . Computed tomography

Introduction

Acute appendicitis and ileocolic intussusception are commoncauses of acute abdominal pain in young children. Timely andaccurate diagnosis of these conditions is essential for prompt,appropriate treatment and minimization of patient morbidity.

Imaging evaluation of abdominal pain in children oftenbegins with a sonogram of the abdomen and pelvis [1]. Thetypical sonographic appearances of acute appendicitis andintussusception have been well illustrated [24]. Several po-tential mimics of intussusception on imaging have been de-scribed [1, 5, 6]. Little, however, has been written regardingthe confusion of the sonographic appearances of rupturedappendicitis and intussusception. We consider ruptured ap-pendicitis to be an important mimic of intussusception, withpotential for delayed diagnosis or misdiagnosis. We also sug-gest that careful examination of specific findings on a sono-gram can help in differentiating the conditions in a timelyfashion.

B. Newman (*) :M. Schmitz : R. Gawande : S. Vasanawala :R. BarthDepartment of Radiology, Stanford University,Lucile Packard Childrens Hospital, 725 Welch Road, MC 5913,Stanford, CA 94305, USAe-mail: bevn@stanford.edu

Pediatr Radiol (2014) 44:535541DOI 10.1007/s00247-014-2873-8

Materials and methods

The peer-review missed-case process identified six childrenover a 7-year period (20062013) who presented to the emer-gency room (ER) with abdominal symptoms including painand had an initial sonographic diagnosis of intussusceptionbut subsequent diagnosis of ruptured appendicitis. In all casesinitial US images were obtained in the emergency room; theseincluded gray-scale, cine and color Doppler images acquiredby US technologists trained in adult and pediatric sonography.In two cases, the children were also scanned by a pediatricfellow (n=1) or attending physician (n=1). Most US studieswere obtained on a GE Logiq E9 scanner (GE Healthcare,Milwaukee, WI), with one initial and several follow-up USstudies on a Siemens Sequoia unit (Siemens Healthcare,Mountainview, CA). Two of the three follow-up studies wereobtained in the pediatric sonography unit with dedicated pe-diatric technologists. All three repeat US studies were scannedby a faculty pediatric radiologist.

Three pediatric radiologists (2 faculty, each with over30 years of experience in pediatric radiology, 1 fellow) inde-pendently and with subsequent consensus evaluated the initialUS studies in a non-blinded fashion. These evaluations werecompared with the formal reports of the initial and follow-upUS studies. Reviewers, aware of the final diagnosis of rup-tured appendicitis, were asked whether they would have ac-cepted the initial US diagnosis of intussusception and whatUS findings were most suggestive of the correct diagnosis.

The images and reports of subsequent contrast enemas andCT scans were reviewed and correlated with the US images.Demographic data, clinical presentation andmanagement, andsurgical findings were obtained from a review of the patientscharts; IRB permission was obtained for image and chartreview.

Results

The number of cases scanned in the same hospital ER forappendicitis or intussusception over the exact time period isnot known. However a report from the same hospital notedthat approximately 1,200 children were evaluated in the ERfor suspected appendicitis during a 5-year period between2003 and 2008 [7]. The six children described in the currentreport (Table 1) (20062013) ranged in age from 16months to8 years (mean age 4.4 years); there were 4 boys and 2 girls(Table 1). All six children presented with abdominal pain; fourof the six had a history of fever or were febrile in the ER; threeof six had an elevated white blood cell count (Table 1). Threechildren (ages 68 years) were older than the typical age forpediatric ileocolic intussusception, and in two cases clinicalsymptoms of periumbilical- to right-lower-quadrant pain weremore typical of appendicitis than intussusception.

The initial US images were obtained in the ER and werepreliminarily interpreted by a radiology resident (1) or pedi-atric radiology fellow (2), with final interpretation (the sameday in four and the next day in two cases) by various facultypediatric radiologists with a wide range of experience. In allsix cases sonographic diagnoses of right mid- to lower-quadrant mass consistent with intussusception were sug-gested. This was thought to be ileocolic and idiopathic in fourcases (Figs. 1 and 2) (Table 1). In one child, the lesion wasnoted to contain a large central shadowing calcification thatwas interpreted as a probable pathological calcified mass andthe lead point of an intussusception (Fig. 3). In a second child,the interpreting pediatric radiologist recommended a CT scanbecause of the patients age (8 years) and clinical symptoms ofpain and fever, although the US findings were thought to beconsistent with ileocolitis and an ileoileal intussusception. Athird child, 23 months old, with an emergency-room USdiagnosis of intussusception had repeated US imaging in theERwith real-time scanning by the pediatric radiologist prior toa planned air-reduction enema. Recognition of markedperilesional inflammatory changes and fluid led to a changein the suggested diagnosis and cancellation of the enema.

Independent and consensus review of the six initial USexaminations produced diagnostic agreement among the threereviewing pediatric radiologists that the initial diagnosis ofintussusception was reasonably supported by the imagingappearance. Each case exhibited a right lower abdominal masswith an alternating hypoechoic and echogenic ring-like ap-pearance, suggesting bowel within bowel (Figs. 1, 2 and 3).

Table 1 Ruptured appendicitis mimicking intussusception

Demographic datan=6

Age: 16 months8 years, mean 4.4 years4 boys, 2 girls

Clinical findingsn=6

Abdominal pain6Fever4Elevated WBC3

Initial US diagnosisn=6

Idiopathic ileocolic intussusception4Intussuscepting calcified ileocolic mass1Ileoileal intussusception1

Repeat USn=3

Suggested ruptured appendicitis3(additional clinical information and consultationavailable, all three scanned directly byradiologist, one case after negative enema

Contrast enemainterpretation

n=4

Normal1Pericolic inflammation, loculated free air1(suggestive of ruptured appendicitis)

Intussusception reduced2 (partial in 1)

Contrast-enhancedCT

n=4

All diagnostic of ruptured appendicitis

Surgeryn=6

Acute laparoscopic appendectomy3Percutaneous abscess drainage1Antibiotics and interval appendectomy2

WBCwhite blood cell count

536 Pediatr Radiol (2014) 44:535541

This US appearance was considered too complex to simplyrepresent thickened cecum.

Subsequently four of the six children underwent contrastenemas for intussusception reduction. Experienced attendingpediatric radiologists performed these studies. One enema in a3-year-old was interpreted as normal; a repeat US demonstrat-ed a persistent mass; the possibility of perforated appendicitiswas considered and a CT scan was recommended (Fig. 1)(Table 1). A water-soluble enema in a 16-month-old childrevealed a small bubble of air adjacent to the cecum in theright lower quadrant as well as pericecal inflammatory chang-es, suggesting ruptured appendicitis. However, in two 6-year-old children, the pediatric radiologists performing the enemaexaminations were convinced that intussusception had beenpresent; one was thought to have been rapidly reduced(Fig. 2), and in the other case it was thought to have beenincompletely reduced with a residual right lower quadrant

mass (Fig. 3). The child who was thought to have had anintussusception reduced on enema had progressive clinicalsymptoms of right lower quadrant pain and fever. Heunderwent a repeat US with direct scanning by an attendingpediatric radiologist and discussion with the pediatric surgeon.Although the US appearance was similar to that of the priorstudy, marked echogenicity was recognized around the mass,suggesting inflammation, leading to the conclusion that rup-tured appendicitis was probably a more likely diagnosis. Thisresulted in a confirmatory CT scan and prevented a plannedrepeat enema for intussusception reduction (Fig. 2).

A total of four children had contrast-enhanced CTs dem-onstrating ruptured appendicitis. All the CT examinationsdemonstrated a double-ring appearance correlating with theUS findings of the target sign, with inner and outer ringsrepresenting the dilated appendix and walled-off appendicealrupture, respectively (Figs. 1, 2 and 3), with variable

Fig. 1 Ruptured appendicitis in a3-year-old girl who presentedwith cramping abdominal pain,was afebrile and had normal whiteblood cell count. aAnteroposterior abdominalradiograph shows multiple dilatedbowel loops and a paucity of gasin the right lower quadrant. Theproperitoneal fat appears normal,and the cecum is not visualized.Clinical and imaging findings aremost suggestive ofintussusception. b Transverse USimage demonstrates a mass in theright lower quadrant withmultiple alternating layerssuggestive of intussusception.Eccentric internal echogenicity(arrows) was thought to besuggestive of mesenteric fat. Inretrospect, moderate perilesionalechogenicity (inflammation) canbe seen. c, dAxial contrastenhanced CT (c) and coronalreconstruction (d). The CTwasobtained after a normal watersoluble contrast enema followedby repeat US which showed themass to be unchanged. The CT(c, d) demonstrated a complexmass in the right lower quadrantwith central enhancing appendix(arrows) containing anappendicolith and surroundingfluid and inflammation, consistentwith perforated appendicitis

Pediatr Radiol (2014) 44:535541 537

surrounding phlegmon or fluid. Reaching the correct diagno-sis of ruptured appendicitis took between 3 h and 24 h in thesesix children, with a mean of 10.5 h.

All six children eventually had surgical confirmation ofruptured appendicitis with no evidence of appendiceal or otherintussusception; in three cases laparoscopic appendectomyand drainage were performed acutely. One of these childrendeveloped subsequent adhesive small-bowel obstruction

requiring additional surgery. One child underwent percutane-ous abscess drainage by interventional radiology, developed acolocutaneous fistula and had a subsequent interval appendec-tomy and drain removal. Two other children were treated withantibiotics acutely with interval appendectomy (Table 1).

Upon further review of the initial US examinations, all sixchildren demonstrated an appearance suggesting multipleconcentric rings of bowel (target sign, most apparent on

Fig. 2 Abdominal pain andperforated appendicitismimicking ileocolicintussusception in a 6-year-oldboy. a, b Transverse (with colorDoppler) (a) and longitudinal (b)US images of the right lowerquadrant demonstrate a 3.53-cmmass with sonolucent andhyperechoic layers thought torepresent an intussusception. Acentral partially shadowingechogenic focus (arrows) wasmisinterpreted as central fat (fatshould not be expected to shadowto this extent). c Fluoroscopicimage from a water-solubleenema at which anintussusception was thought tohave been rapidly reduced. Thereare mild right-side pericolicinflammatory changes, but nomass is evident. The appendix isnot filled. d, eCT scan with IVcontrast the next day, after repeatUS suggested possible rupturedappendicitis. Axial (d) andcoronal reconstruction (e) imagesshow a central dilated fluid-filledappendix (arrows) with wallenhancement and peri-appendiceal fluid collection orabscess, suggestive of acuteperforated appendicitis. There isresidual bowel contrast from theprior enema. An appendicolithcan be seen at the base of theappendix (arrow in e) and there isan additional fluid collection inthe pelvis

538 Pediatr Radiol (2014) 44:535541

transverse views), but none demonstrated typical echogeniccrescentic central mesenteric fat or intralesional lymph nodes.Focal central echogenicity on US was misinterpreted in somecases as possible fat but proved to be debris or appendicoliths.All six of our cases had surrounding thick hyperechoicechogenicity that, in retrospect, likely represented inflamedfat walling off the perforated appendix. In five of six childrenthe size of the mass on US ranged from 3 cm to 4.7 cm, with amean size of 3.7 cm by 3 cm, similar to typical ileocolicintussusception (35 cm). Mass size was 2 cm by 2 cm inone child thought to have ileoileal intussusception.

Discussion

Acute abdominal pain in children is a common diagnosticdilemma, with acute appendicitis and intussusceptionrepresenting two of the most important diagnostic consider-ations [1, 6]. The classic symptoms of acute appendicitisinclude anorexia, periumbilical pain followed by right lowerquadrant pain, and vomiting, which can be present in manyother causes of acute abdominal pain in children.

Classic symptoms of intussusception include intermittentabdominal pain and irritability with later diarrhea, bloodystools and lethargy, although approximately half of patientsdo not present with typical symptoms [1, 5].

The clinical presentations of children with acute appendi-citis (including ruptured appendicitis) or intussusception canoverlap, with both conditions capable of resulting in leukocy-tosis, fever and acute abdominal pain. A palpable mass in theright flank or lower abdomen is also common to both entities.

Idiopathic intussusception is most often seen in children,mostly boys, between 6 months and 3 years of age and isconsidered to be related to hypertrophy of bowel lymphoidtissue (Peyer patches), often triggered by a preceding gastro-intestinal illness. Intussusception in neonates, as well as olderchildren, is less likely to be idiopathic and is frequentlyassociated with a pathological lead point such as a Meckeldiverticulum, duplication cyst, appendicitis, inspissated me-conium (in cystic fibrosis), or a neoplasm such as a polyp orlymphoma [8]. Small-bowel intussusception can be a transientincidental finding on imaging, but symptomatic cases mightrequire operative management, especially in postoperativepatients, in association with small-bowel obstruction and witha longer segment of involvement [9, 10].

US examination is often the first imaging study performedin children with an acute abdominal presentation. Because ofconcerns for radiation exposure, especially in young children,CT is reserved for cases in which US findings are equivocal ornon-diagnostic [1, 7]. The classic sonographic appearance ofileocolic intussusception is well-established [13, 5]. US isconsidered a very good imaging study for the detection ofintussusception, with reported sensitivity of 100% and

Fig. 3 Ruptured appendicitis in a 6-year-old girl. a Initial US, transverseview, shows a mass (4.74 cm) consisting of multiple concentric ringswith a central echogenic shadowing focus (arrow) that was initiallyinterpreted as an intussusception, with a calcified mass as the likely leadpoint. bAn air enema later the same day demonstrates reflux of air intothe small bowel with mass effect on the cecum (arrows) thought to

represent a residual unreduced intussusception with air having escapedinto small bowel past the mass. c Axial post-contrast CT the same daydemonstrates ruptured appendicitis with a central fluid-filled appendixand appendicolith (arrow) and surrounding fluid and inflammation cor-responding to the multiple layers seen on ultrasonography

Pediatr Radiol (2014) 44:535541 539

specificity of 88%. The target, bulls-eye, or doughnut sign ofmultiple hypoechoic concentric rings caused by edematousbowel walls, lymph nodes, and blood vessels with a crescentichyperechoic focus of mesenteric fat between the two layers ofbowel are well-known sonographic findings (Fig. 4) [2, 3].The transverse appearance of a crescent in a doughnut isconsidered almost pathognomonic [3, 5] for intussusception.On longitudinal scans, the intussusception lesion is ovoid inshape, with different tissues appearing layered longitudinallyand often referred to as a sandwich or pseudokidney sign(Fig. 4) [2]. Intralesional lymph nodes are commonly seen inileocolic intussusception [10]. Other conditions can mimic theappearance of intussusception on US, including enterocolitis,volvulus, inflammatory bowel disease and other causes ofedematous or hemorrhagic bowel, as well as feces-filled colonand even psoas hematoma [2, 5, 11]. Small-bowel intussus-ception has the same appearance but is usually more central inlocation, smaller than the typical 3- to 5-cm size of the mass inileocolic intussusception and containing less mesenteric fatand fewer lymph nodes [1, 10].

Appendicitis can occur at any age but is relatively uncom-mon in younger infants. Appendiceal intussusception is veryunusual, occurring either as an isolated entity or very rarely aspart of an ileocolic intussusception [12, 13]. Appendicealintussusception in children is associated with appendicitis oran appendiceal mass. None of our cases had surgical evidenceof appendiceal or other intussusception.

The classic sonographic appearance of acute appendicitishas been well-described. A blind-ending tubular structurearising from the base of the cecum measuring greater than6 mm in diameter and having hyperemic walls as well assurrounding hyperechoic fat, free fluid or enlarged lymphnodes while being non-compressible is diagnostic of acuteappendicitis [1, 4, 7]. Appendiceal rupture is particularlycommon in children [1]. Ruptured appendicitis can prove tobe a difficult diagnosis, both clinically and sonographically.Sonographic signs of appendiceal perforation include loss ofthe echogenic mucosa, increased periappendiceal

echogenicity from surrounding inflammation, and a complexmass or focal fluid collection [4]. The appendix itself mightnot be seen at all or be difficult to distinguish from surround-ing inflammation, fluid and gas.

We found that sonographic appearances of ruptured appen-dicitis can very closely mimic the sonographic findings ofintussusception; we recognize that our sample size of six casesis small and that the unblinded methodology used to re-evaluate the US studies is biased, but it still confirms that thismisinterpretation can readily occur. Careful imaging, clinicalcorrelation and an astute radiologist are required to distinguishthe entities. The correct diagnosis is most likely to be consid-ered when there is real-time visualization of the lesion inmultiple planes by a pediatric radiologist, with recognitionof surrounding echogenicity caused by inflamed fat, and theabsence of central mesenteric fat or intralesional nodes. Le-sion size was not helpful in our cases because the mass sizeswere close to those expected in typical ileocolic intussuscep-tion, with the exception of one case, which was thought to beileoileal intussusception. Clinical presentation and featuressuch as fever and leukocytosis that might raise suspicion forappendicitis were not present in several of our cases.

Also, the detailed clinical information was usually notavailable to the radiologist at the time of initial US examina-tion. Because the US imaging of intussusception is typicallyhighly reliable, the diagnosis was readily accepted and thenacted on by clinicians. One result was a delay in diagnosis andtreatment; mean diagnostic delay was 10.5 h but was as longas 24 h in one case. Another consequence was the addition ofmultiple imaging studies, including four contrast enemas,three repeat US studies and four CT scans.

Very few prior reports of this misdiagnosis are in theliterature, although we have encountered and been confound-ed by this problem on multiple occasions, as have otherexperienced pediatric radiologists (personal communications).A published review of emergency pediatric US imaging illus-trates a case of ruptured appendicitis in a 7-year-old thatclosely mimicked an intussusception on US; he underwent

Fig. 4 Imaging in a 17-month-old boy who presented withcramping abdominal pain andileocolic intussusception.Transverse (a) and sagittal (b) USimages show the typicalappearance of ieocolicintussusception, with amultilayered multiple-ringappearance with centralechogenic fat and intralesionalhypoechoic nodes (arrows). Notethe absence of surroundingechogenic inflamed fat. The childunderwent a successful air-reduction enema

540 Pediatr Radiol (2014) 44:535541

what was deemed an unsuccessful air-enema reduction andwas only correctly diagnosed at surgery, similar to our cases[1]. Another case report detailed misdiagnosis of rupturedappendicitis on both US and CT as probable appendicealintussusception in an 11-year-old girl [12]. In another casereport an appendiceal hematoma in an 8-year-old girl with vonWillebrand disease was originally thought to be an intussus-ception on US but was later identified as a hematoma on CT[11]. These individual reports suggest that the most helpfuldifferential features were older patient age and smaller size ofthe lesion compared to that of typical idiopathicintussusception.

The CT scans obtained in our cases were diagnostic ofruptured appendicitis with no confusion with intussusception(Figs. 1, 2 and 3). In situations where US or enema findingsare unusual or at odds with the clinical picture, radiationconcerns should not prevent the appropriate move to CT forclarification. MR is being used more commonly for the diag-nosis of acute appendicitis and might also be an appropriatealternative in suitable patients.

Our cases illustrate that ruptured appendicitis can appear asa rounded, mass-like structure with multiple rings that can beeasily mistaken for intussusception. Awareness of the possi-bility of this confusion should encourage more detailed,thoughtful scrutiny of the initial images, with a greater con-sideration of the possibility of complicated appendicitis.

Conclusion

Ruptured appendicitis is an important mimic of intussuscep-tion on US examination. Being mindful of ruptured appendi-citis as a diagnostic possibility in children with atypical age orclinical features and US findings suggestive of intussuscep-tion should allow for careful attention to subtle differences inthe sonographic appearances between the entities. Particularattention should be paid to evaluating the presence of sur-rounding inflamed fat in appendicitis and intralesionalhypoechoic lymph nodes in ileocolic intussusception. Aninner crescent of echogenic mesenteric fat between bowelloops is a virtually pathognomonic feature of intussusceptionthat was not definitely present in our cases of ruptured appen-dicitis, although it was misinterpreted as such when central

echogenic debris or appendicolith was present. Real-time USimaging by a radiologist in multiple planes and direct com-munication with clinicians were useful in suggesting the cor-rect diagnosis or correcting an erroneous one. CT was diag-nostic in all four cases in which it was employed and should beconsidered promptly when there are equivocal US features orthere is a discrepancy between US and clinical findings.

Conflict of interest None

References

1. Cogley JR, OConnor SC, Houshyar R et al (2012) Emergent pedi-atric US: what every radiologist should know. Radiographics 32:651665

2. Anderson DR (1999) The pseudokidney sign. Radiology 211:395397

3. Del-Pozo G, Albillos JC, Tejedor D (1996) Intussusception: USfindings with pathologic correlationthe crescent-in-doughnut sign.Radiology 199:688692

4. Sivit CJ (1993) Diagnosis of acute appendicitis in children: spectrumof sonographic findings. AJR Am J Roentgenol 161:147152

5. Daneman A, Navarro O (2003) Intussusception. Part 1: a review ofdiagnostic approaches. Pediatr Radiol 33:7985

6. Van Breda Vriesman AC, Puylaert JB (2006) Mimics of appendicitis:alternative nonsurgical diagnoses with sonography and CT. AJR AmJ Roentgenol 186:11031112

7. Krishnamoorthi R, Ramarajan N,WangNE et al (2011) Effectivenessof a staged US and CT protocol for the diagnosis of pediatricappendicitis: reducing radiation exposure in the age of ALARA.Radiology 259:231239

8. Williams H (2008) Imaging and intussusception. Arch Dis ChildEduc Pract Ed 93:3036

9. Ko SF, Lee TY, Ng SH et al (2002) Small bowel intussusception insymptomatic pediatric patients: experiences with 19 surgically prov-en cases. World J Surg 26:438443

10. Lioubashevsky N, Hiller N, Rozovsky K et al (2013) Ileocolic versussmall-bowel intussusception in children: can US enable reliabledifferentiation? Radiology 269:266271

11. Tejani C, Phatak T, Sivitz A (2012) Right lower-quadrant painmore than one diagnosis. Pediatr Emerg Care 28:12241226

12. Christianakis E, Sakelaropoulos A, Papantzimas C et al (2008) Pelvicplastron secondary to acute appendicitis in a child presented asappendiceal intussusception. A case report. Cases J 1:135

13. Dietz KR, Merrow AC, Podberesky DJ et al (2013) Beyond acuteappendicitis: imaging of additional pathologies of the pediatric ap-pendix. Pediatr Radiol 43:232242, quiz 259

Pediatr Radiol (2014) 44:535541 541

Copyright of Pediatric Radiology is the property of Springer Science & Business Media B.V.and its content may not be copied or emailed to multiple sites or posted to a listserv withoutthe copyright holder's express written permission. However, users may print, download, oremail articles for individual use.

Perforated appendicitis: an underappreciated mimic of intussusception on ultrasoundAbstractAbstractAbstractAbstractAbstractAbstractIntroductionMaterials and methodsResultsDiscussionConclusionReferences