2012 SSAT QUICK SHOT PRESENTATION

Anatomophysiology of the Pharyngo-Upper Esophageal Areain Light of High-Resolution Manometry

Luciana C. Silva & Fernando A. M. Herbella &

Luciano R. Neves & Fernando P. P. Vicentine &

Sebastião P. Neto & Marco G. Patti

Received: 24 April 2013 /Accepted: 9 September 2013 /Published online: 28 September 2013# 2013 The Society for Surgery of the Alimentary Tract

AbstractIntroduction The motility of the pharynx, upper esophageal sphincter (UES), and proximal esophagus is still poorly understood.These structures have anatomical and functional peculiarities that hinder the accurate study of their motility with the technologytraditionally available. High-resolution manometry (HRM) has characteristics that make it more suitable for the study of the upperdigestive tract. This study aims to evaluate in healthy volunteers, using HRM and transnasal pharyngoscopy, (1) the correlationbetween anatomical landmarks and HRM plots and (2) the normal values for manometric parameters of the pharynx, UES, andproximal esophagus.Methods We studied 40 asymptomatic volunteers with HRM (50 % male; median age, 27 years). Fourteen of those alsounderwent transnasal pharyngoscopy.Results and discussion Pharyngeal peak pressure, rise time, recovery time, and duration of contraction were 128 mmHg (range,100–164 mmHg), 197 ms (range, 169–268 ms), 385 ms (range, 285–465 ms), and 604 ms (range, 544–626 ms) at the velum and116mmHg (range, 97–139mmHg), 128 ms (range, 100–156ms), 194ms (range, 148–219ms), and 336ms (range, 267–386ms)at the epiglottis, respectively. UES extension, basal pressure, residual pressure, and duration of relaxation were 3 cm (range, 2.6–3.6 cm), 76 mmHg (range, 58–109 mmHg), 4.4 mmHg (range, 1.2–6.9 mmHg), and 678 ms (range, 636–757 ms), respectively.In the proximal esophagus, wave amplitudes at 2, 4, and 6 cm below the UES were 72 mmHg (range, 53–97 mmHg), 56 mmHg(range, 42–76 mmHg), and 48 mmHg (range, 35–59 mmHg), respectively.Conclusions In conclusion, normal values were established. These values may prove clinically useful and could contribute tofuture studies with dysphagic patients.

Keywords High-resolutionmanometry . Pharynx . Upperesophageal sphincter . Proximal esophagus . Striatedmuscle

Introduction

The motility of the pharynx, upper esophageal sphincter(UES), and proximal esophagus is a topic of digestive phys-iology still not entirely understood even though their proximalanatomic position facilitates access to diagnostic tests.

The pharynx is an anatomically and functionally complexorgan. It is composed of intrinsic and extrinsic muscularlayers. The pharyngeal constrictors overlap posteriorly in themedian raphe. Anteriorly, however, the muscles are attachedto bony and cartilaginous structures and their muscular fibersare scarce.1–3 The longitudinal muscular layer is also asym-metric, since most of its fibers are situated in the lateral andposterior pharyngeal walls. These anatomic characteristicslead to functional asymmetry as well. Sears at al. showed theinequality of longitudinal and radial pharyngeal pressureswith the aid of catheters with circumferential sensors.3 Themuscular fibers of the pharynx, the UES, and the proximal

Quick shot presentation at the 53rd SSAT Meeting at DDW, San Diego,CA, May 2012.

L. C. Silva : F. A. M. Herbella : F. P. P. Vicentine : S. P. NetoDepartment of Surgery, Escola Paulista de Medicina, FederalUniversity of São Paulo, São Paulo, Brazil

L. R. NevesDepartment of Otorhinolaryngology, Escola Paulista de Medicina,Federal University of São Paulo, São Paulo, Brazil

M. G. PattiDepartment of Surgery, University of Chicago, Chicago, IL, USA

F. A. M. Herbella (*)Surgical Gastroenterology, Division of Esophagus and Stomach,Hospital São Paulo, Rua Diogo de Faria 1087 cj 301, São Paulo,SP 04037-003, Brazile-mail: herbella.dcir@epm.br

J Gastrointest Surg (2013) 17:2033–2038DOI 10.1007/s11605-013-2358-3

esophagus are striated, unlike the distal digestive tract, showingquick motor response. The pharyngeal contractions in normalindividuals generate waves whose amplitude can be ashigh as 200 mmHg, speed up to 25 cm/s, and duration asshort as 0.3 s.4

Usually, the conventional manometry apparatus is a water-perfused system with unidirectional and widely spaced sensorswith slow response to pressure variations. These characteristicsmake the conventional system unable to accurately analyze themotor function of the striated muscle.5 High-resolution ma-nometry (HRM) seems to gather technological advances thatallow a better study of the pharynx, the UES, and the proximalesophagus due to the circumferential disposition of the sensorsand the ability to detect rapid pressure changes.

This study aims to establish (1) the anatomofunctionalcorrelation between HRM plots and anatomical landmarksand (2) the normal values for the manometric parameters ofthe pharynx, the UES, and the proximal esophagus.

Material and Methods

Population

Forty asymptomatic volunteers (median age, 27 years (range,23–34 years); 50 % men) were studied. Individuals withprevious operation on the studied organs, with digestivesymptomsmore than once amonth, or who refuse to participatewere excluded.

High-Resolution Manometry

HRM data of all volunteers were acquired using a solid-stateHRMassembly with 36 sensors spaced at 1 cm intervals (GivenImaging, Los Angeles, CA, USA). All studies were performedwith patients in the left lateral position, after a minimum fastingperiod of 8 h. The manometric protocol included a 20-s periodto assess UES basal pressure and ten 5-mL water swallows.Manometric data were analyzed using the ManoView analysissoftware (Given Imaging, Los Angeles, CA, USA).

The following manometric parameters were studied(Fig. 1):

At the topography of the velopharynx and epiglottis:

& Peak pressure: measured at the point of highest pressureaccording to the transnasal pharyngoscopy parameters (inmillimeters of mercury);

& Rise time: time interval from the onset of pharyngealcontraction to peak pressure (in milliseconds);

& Upstroke: peak pressure/rise time rate (in millimeters ofmercury per second);

& Recovery time: time interval from peak pressure to the endof the contraction (in milliseconds);

& Contraction duration (in milliseconds);& UES relaxation latency: time interval between the onset of

pharyngeal contraction and the start of UES relaxation (inmilliseconds);

& UES maximum relaxation latency: time interval betweenthe onset of pharyngeal contraction and the UES relaxa-tion nadir (in milliseconds);

& UES end relaxation latency: time interval between theonset of pharyngeal contraction and the end of UES re-laxation (in milliseconds).

At the topography of the UES:

& Extension (in centimeters);& Basal pressure: measured at the midpoint, during the 20-s

period before the first swallow (in millimeters of mercury);& Residual pressure: nadir pressure at the midpoint (in mil-

limeters of mercury);& Relaxation duration (in milliseconds);& Relaxation time to nadir: time between the beginning of

the relaxation and the nadir pressure (in milliseconds);& Recovery time: time between the nadir and the end of the

relaxation (in milliseconds).

Proximal esophagus parameters:

& Wave amplitude at 2, 4, and 6 cm below the UESmidpoint(in millimeters of mercury);

& Wave duration at 2, 4, and 6 cm below the UES midpoint(in seconds);

& Wave velocity between 2–4, 4–6, and 2–6 cm below theUES midpoint (in centimeters per second).

Transnasal Pharyngoscopy

Fourteen (median age, 26 years (range, 23–29 years); 85 %men) of the 40 volunteers underwent transnasal pharyngoscopy

Fig. 1 Manometric parameters studied. 1 velopharynx, 2 epiglottis, 3UES, 4 2 cm proximal esophagus, 5 4 cm proximal esophagus, 6 6 cmproximal esophagus. White targets denote time measurements. Blacktargets denote pressure measurements

2034 J Gastrointest Surg (2013) 17:2033–2038

immediately after HRM. The endoscopy was introducedthrough the same nostril previously used to insert the manom-etry catheter.

The distance of the following structures was measured incentimeters, from the nostril: (a) the velopharynx, (b) theepiglottis, and (c) the upper border of the cricopharyngeus.The obtained distances were correlated to the correspondingHRM images.

Statistics

Data are shown as the median (interquartile range). All data,except basal pressure, correspond to a mean of 10 swallows.

Ethics

The study protocol was approved by the local ethics committee,and written informed consent was obtained from each subject.

There is no conflict of interest. All authors contributedsufficiently to be named as authors and are responsible forthe manuscript; no professional or ghost writer was hired.

Results

Anatomofunctional Correlation

The distance between velum, epiglottis, and UES and thenares were, respectively, 11.5 cm (range, 10.7–11.9 cm),15.3 cm (range, 14.2–15.8 cm), and 19 cm (range, 17.6–19.3 cm). Figure 2 shows the correlation between HRM plotsand anatomical landmarks for volunteer no. 3. Identical findingswere obtained in all 14 volunteers.

Manometric Parameters

The normal values for pharyngeal manometric parameters atthe topography of the velopharynx and epiglottis are shown inTable 1. The normal values for UES manometric parametersare depicted in Table 2. The normal values for proximalesophagus manometric parameters are shown in Table 3.

Discussion

High-Resolution Manometry of the Pharyngo-UpperEsophageal Area

Since the advent of HRM, only few studies have emergedwiththe objective of exploring the potentials of this new technol-ogy for better comprehension of the digestive striated musclephysiology. Certain authors analyzed manometric parametersin situations where “normal” swallowing was modified byexternal factors, such as variation in bolus volume,6 changesin the position of the head and neck,7,8 cough,9 and ma-neuvers to improve swallowing in dysphagic patients, suchas effortful swallow,10 the Mendelshon,10 and the tongue-holding maneuver.11 The UES has also been studied in

Fig. 2 Correlation between HRM plots and anatomic landmarksidentified by transnasal pharyngoscopy

Table 1 Normal values forpharyngeal manometric parametersat the topography of thevelopharynx and the epiglottis

Data are shown as the median(interquartile range)

Velopharynx Epiglottis

Peak pressure (mmHg) 128.7 (100.3–164.3) 116.4 (97.5–139.3)

Rise time (ms) 197.5 (169.2–268.5) 128.0 (100.0–156.0)

Upstroke (mmHg/s) 514.8 (345.4–664.5) 714.3 (534.6–970.1)

Recovery time (ms) 385.5 (285.5–465.2) 194.0 (148.7–219.0)

Contraction duration (ms) 604.5 (544.5–626.7) 336.0 (267.0–386.0)

UES relaxation latency 59.5 (−43.0–126.0) −237.0 (−325.0 to −239.0)UES maximum relaxation latency 259.0 (193.5–312.0) −73.0 (−129.0–65.0)UES end relaxation latency 754.5 (638.5–787.75) 416.0 (331.0–472.0)

J Gastrointest Surg (2013) 17:2033–2038 2035

symptomatic patients with gastroesophageal reflux diseaseand chronic cough12 and globus.13 Despite these initialstudies, the determination of normal values for manometricparameters of the pharyngo-upper esophageal area wasrarely the aim of recent reports, especially concerning theanatomofunctional correlations and the temporal associationbetween swallowing events.

In 2008, Takasaki et al.14 studied, with the aid of HRM, 33asymptomatic Japanese volunteers aiming to establish normalvalues for manometric parameters from the pharynx to theproximal esophagus. The authors analyzed peak pressure inthe regions of the velopharynx, mesopharynx, and the UES aswell as the distance between these structures and the nostril.There has been no assessment of the temporal correlationsbetween swallow events, and the authors limited their study tothe analysis of pharyngeal and UES pressures during swallow.In an effort to establish a correlation between the HRM andthe pharyngeal anatomy, the authors utilized functional ma-neuvers. The velopharynx was considered as the high-pressure zone observed with vocalization of “kakakaka,”and the mesopharynx corresponded to the area between thevelopharynx and the UES. Despite the different methodolo-gies for anatomofunctional correlation, our study confirms thevelopharynx position in the HRM plots with the aid oftransnasal pharyngoscopy.

In 2006, Ghosh et al.15 analyzed UES relaxation in 75asymptomatic volunteers with HRM. The authors determinedmanometric parameters such as minimal relaxation pressure,relaxation interval, median intrabolus pressure during therelaxation interval, and deglutitive sphincter resistance. Thedetermination of such parameters was achieved by using a

mathematical software. Nevertheless, the reference valuesestablished by this group are currently the most used ones,although in clinical practice, manometric values are usuallysimply calculated by the commercial built-in software. Simi-larly, Mielens at al.16 also used the same software to developan algorithm for automated extraction and analysis of HRMdata from 12 normal subjects and 3 with impaired swallowing.The authors consider that the analysis of the area under thepressure curve of the regions of interest would provide moreaccurate information than the analysis of the data that originat-ed from a single sensor, since the structures move up to 4 cm inthe longitudinal axis during swallow. In our study, we used thecommercial dedicated software since it is intuitive and easy tolearn, without the need of complex mathematical skills. Thereis little interobserver variability in the results, even when thesoftware is operated by inexperienced physicians.17 Our goalwas to determine normal values for manometric parameters ofthe proximal digestive tract, using tools that are available inclinical practice and amethodology easily reproducible in otherresearch centers or in nonacademic environment.

McCulloch et al.7 tried to manometrically define the regionsof interest based on intuitive interpretation of HRM plots. Thevelopharynx was defined as the region of swallow-relatedpressure change just proximal to the area of continuous nasalcavity quiescence and extending 2 cm. The tongue base regionwas defined as the area of swallow-related pressure changewith a high-pressure zone identified approximately midwaybetween the nasopharynx and the UES, with its epicenter atthe high-pressure point and extending 2 cm proximal and distalto that point. The UES region was defined as the midpoint ofstable high pressure just distal to the baseline low esophagealpressure zone, extending to a point of low esophageal pressuredistally and low baseline pharyngeal pressure proximally. Withthe use of transnasal pharyngoscopy and measurement of thedistance between the structure and the nostril, we could deter-mine the velopharynx region more precisely in the HRM plotand slightly distal to that indicated by the previously mentionedauthors. Furthermore, we could precisely identify the epiglottis'manometric impression within the region considered by theseauthors as the tongue base. Obviously, the tongue base is thecontractile structure that moves the epiglottis posteriorly.

Table 2 Normal values for UES manometric parameters

Extension (cm) 3.0 (2.6–3.6)

Basal pressure (mmHg) 76.3 (58.2–109.1)

Residual pressure (mmHg) 4.4 (1.2–6.9)

Relaxation time to nadir (ms) 201.0 (144.0–241.0)

Relaxation duration (ms) 678.0 (636.0–757.0)

Recovery time (ms) 501.0 (394.0–549.0)

Data are shown as the median (interquartile range)

Table 3 Normal values formanometric parameters of theproximal esophagus

Data are shown as the median(interquartile range)

Wave amplitude at 2 cm below UES midpoint (mmHg) 72.1 (53.8–97.5)

Wave amplitude at 4 cm below UES midpoint (mmHg) 56.9 (42.3–76.5)

Wave amplitude at 6 cm below UES midpoint (mmHg) 48.4 (34.9–59.1)

Wave duration at 2 cm below UES midpoint (s) 1.4 (1.25–1.75)

Wave duration at 4 cm below UES midpoint (s) 1.4 (1.1–1.7)

Wave duration at 4 cm below UES midpoint (s) 1.4 (1.2–1.7)

Wave velocity between 2 and 4 cm below UES midpoint (cm/s) 4.9 (4.4–5.8)

Wave velocity between 4 and 6 cm below UES midpoint (cm/s) 7.6 (6.0–9.0)

Wave velocity between 2 and 6 cm below UES midpoint (cm/s) 4.1 (3.5–5.1)

2036 J Gastrointest Surg (2013) 17:2033–2038

Clinical Application of Pharyngo-Upper Esophageal AreaManometry

Conventional esophageal manometry had its clinical applica-tion on the study of the pharyngo-upper esophageal area limitedto the evaluation of the UES due to technical limitations aspreviously discussed. At that period, most studies conductedwere interested in the pathogenesis and therapy of Zenker'sdiverticulum and other indications for cricopharyngealmyotomy. Very interestingly, there are no studies showingHRM findings in patients with Zenker's diverticulum. In theHRM era, a renewed interest in the study of this area againfocused studies not only on the cricopharyngeal myotomy18

but also on new applications of the methodology, such as sleepapnea19 and globus20 as examples. After the determination ofnormal values and future studies to allow the identification ofpathologic findings, HRM may help manage patients withoropharyngeal dysphagia.

Study Limitations

This paper represents a modest contribution to the study of themotility of the upper digestive tract. This is the first studyattempting to establish temporal relations between swallowingevents using HRM. In the future, this information may help toidentify swallowing disorders that are caused by the lack ofcoordination between events as well as abnormalities of wavecontraction amplitude or duration. Another point is that nor-mative data were obtained using HRM built-in software,without the need of exporting data to other softwares. Thischaracteristic may render our results especially applicable innonacademic centers.

The study is limited by the small number of individualsstudied due to the difficulty in recruiting asymptomatic vol-unteers and a large proportion of young subjects since volunteerswere randomly recruited and not selected on a specific popula-tion or based on gender or age. This makes it less representativeof the older population usually affected by swallowing problemseven though it is elusive if aging really leads to the deteriorationof proximal digestive tract motility.21,22

We did not carry out a comparison between HRM andconventional manometry since plenty of previous literaturedata showed that conventional manometry is unsuitable for acomplex study of the pharyngo-upper esophageal area dueto limitations of the technology. Furthermore, it is virtuallyimpossible to locate the parameters of interest with conventionalmanometry.

We also did not perform a simultaneous videofluoroscopy.First, we did not want to submit volunteers to radiation forethical reasons. Second, the act of swallowing and airwayprotection is managed by the coordinated motility and tonusof the upper pharyngo-upper esophageal area, all able to bemeasured by HRM.We tested HRM as a tool that does not use

radiation, fast, intuitive, relatively simple, and possible to beperformed at bedside to evaluate the physiology of thepharyngo-upper esophageal area. We agree that the combina-tion of HRM and videofluoroscopy may be helpful in thealtered swallowing status, identifying possible manometricfindings in the presence of aspiration and residue as elegantlyreviewed by Nativ-Zeltzer et al.23 Furthermore, in the normalstatus, the combination of HRM and videofluoroscopy wouldallow the correct correlation between anatomic structuresand manometric tracings that was possible in our study bytransnasal pharyngoscopy. However, some recent papersshowed that HRM may bring the exact information providedby videofluoroscopy.24

Conclusion

In conclusion, normal values were established. These valuesmay prove clinically useful and could contribute to futurestudies with dysphagic patients.

Acknowledgments Dr. Luciana C. Silva was funded by a governmentalfunding agency, Coordenação de Aperfeiçoamento de Pessoal de NívelSuperior.

Conflict of Interest There is no conflict of interest.

References

1. Gardner E, Gray DJ, O'Rahilly R. 1960. Anatomia. Philadelphia:W.B. Saunders.

2. Mittal RL. Motor function of the pharynx, esophagus and its sphinc-ters. San Rafael: Morgan & Claypoll Life Sciences; 2011.

3. Sears VW, Castel JA, Castell DO. Radial and longitudinal asymmetryof human pharyngeal pressures during swallowing. Gastroenterolo-gy. 1991 Dec; 101(6): 1559–63.

4. DoddsWJ, HoganWJ, Lydon SB, Stewart ET, Stef JJ, Arndorfer RC.Quantitation of pharyngeal motor function in normal subjects. J ApplPhysiol. 1975 Oct; 39(4): 692–6.

5. Ayazi S, Crookes PF. High-resolution esophageal manometry: usingtechnical advances for clinical advantages. J Gastrointest Surh. 2010Feb; 14 Suppl 1: S24-32. Epub 2009 Sep 18.

6. Hoffmann MR, Ciucci MR, Mielens JD, Jiang JJ, McCulloch TM.Pharyngeal swallow adaptions to bolus volume measured with high-resolution manometry. Laryngoscope. 2010 December; 120 (12):2367–2373.

7. McCulloch TM, HoffmanMR, Ciucci MR. High-resolution manom-etry of pharyngeal swallow pressure events associated with head turnand chin tuck. Ann Otol Rhinol Laryngol. 2010; 119(6): 369–76.

8. Takasaki K, Umeki H, Kumigami H, Takasaki H. Influence of headrotation on upper esophageal sphincter pressure evaluated by high-resolution manometry system. Otolaryngol Head Neck Surg. 2010Feb, 142(2):214–7.

9. Amaris M, Dua K, Naini SR, Samuel E, Shaker R. Characterizationof the upper esophageal sphincter response during cough. Chest2012; 142 (5):1229–1236

10. Hoffmann MR, Mielens JD, Ciucci MR, Jones CA, Jiang JJ,McCulloch TM. High-resolution manometry of pharyngeal swallow

J Gastrointest Surg (2013) 17:2033–2038 2037

pressure events associated with effortful swallow and theMendelshon maneuver. Dysphagia. 2012; 27:418–426.

11. Umeki H, Takasaki K, Enatsu K, Tanaka F, Kumigami H, TakhashiH. Effects of a tongue-holding maneuver during swallowing evalu-ated by high-resolution manometry. Otolaryngol Head Neck Surg.2009 Jul; 141(1):119:22.

12. Vadar R, Sweis R, AnggiansahA,Wong T, FoxMR.Upper esophagealsphincter and esophageal motility in patients with chronic cough andreflux: assessment by high-resolution manometry. Diseases of theEsophagus. 2012 doi:10.1111/j.1442.2050.2012.01354.x

13. Kwiatek MA, Mirza F, Kahrilas PJ, Pandolfino JE. Hyperdynamicupper esophageal sphincter pressure: a manometric observation inpatients reporting globus sensation. Am J Gastroenterol 2009 Feb;104(2): 289–98.

14. Takasaki K, Umeki H, Enatsu K, Tanaka F, SakihamaN, Kumagami H,Takahashi H. Investigation of pharyngeal swallowing function usinghigh-resolution manometry. Laryngoscope. 2008; 118:1729–1732.

15. Ghosh SK, Pandolfino JE, Zhang Q, Jarosz A, Kahrilas PJ.Deglutitive upper esophageal sphincter relaxation: a study of 75volunteer subjects using solid-state high-resolution manometry. AmJ Physiol Gastrointest Liver Physiol 2006: 291:G525-G531.

16. Mielens JD, Hoffmann MR, Ciucci MR. Automated analysis ofpharyngeal pressure data obtained with high-resolution manometry.Dysphagia (2011) 26:3–12.

17. Singh ER, Rife C, Clayton S, Naas P, Nietert P, Castell DO.Interobserver variability in esophageal body among new physicianusers. J Clin Gastroenterol. 2013 Feb; 47(2):e12-6.

18. Takasaki K, Umeki H, Enatsu K, Kumagami H, Takahashi H.Evaluation of swallowing pressure in a patient with amyotrophic

lateral sclerosis before and after cricopharyngeal myotomy usinghigh-resolution manometry system. Auris Nasus Larynx. 2010Oct;37(5):644–7. doi:10.1016/j.anl.2010.02.003. Epub 2010 Feb 20.

19. Kuribayashi S, Massey BT, Hafeezullah M, Perera L, Hussaini SQ,Tatro L, Darling RJ, Franco R, Shaker R. Upper esophageal sphincterand gastroesophageal junction pressure changes act to prevent gas-troesophageal and esophagopharyngeal reflux during apneic episodesin patients with obstructive sleep apnea. Chest. 2010 Apr;137(4):769–76. doi:10.1378/chest.09-0913. Epub 2009 Nov 13.

20. Kwiatek MA, Mirza F, Kahrilas PJ, Pandolfino JE. Hyperdynamicupper esophageal sphincter pressure: a manometric observation inpatients reporting globus sensation. Am J Gastroenterol. 2009Feb;104(2):289–98. doi:10.1038/ajg.2008.150. Epub 2009 Jan 20.)

21. Shaker R, Ren J, Podvrsan B, Dodds WJ, Hogan WJ, Kern M,Hoffmann R, Hintz J. Effect of aging and bolus variables on pharyn-geal and upper esophageal sphincter motor function. Am J Physiol.1993 Mar; 264 (3 Pt 1):G427-32.

22. Dantas RO, Alves LM, Dalmazo J, Santos CM, Cassiani Rde A,Nascimento WV. Effect of age on proximal esophageal response toswallowing. Arq Gastroenterol. 2010 Oct-Dec; 47(4):339–43.

23. Nativ-Zeltzer N, Kahrilas PJ, Logemann JA. Manofluorography inthe evaluation of oropharyngeal dysphagia. Dysphagia. 2012Jun;27(2):151–61.

24. Hoffman MR, Jones CA, Geng Z, Abelhalim SM, Walczak CC,Mitchell AR, Jiang JJ, McCulloch TM. Classification of high-resolution manometry data according to videofluoroscopic pa-rameters using pattern recognition. Otolaryngol Head Neck Surg.2013 Jul;149(1):126–33. doi:10.1177/0194599813489506. Epub2013 Jun 1.

2038 J Gastrointest Surg (2013) 17:2033–2038