COMPARISON OF ESOPHAGEAL SCREEN FINDINGS ON VIDEOFLUOROSCOPY WITH FULL ESOPHAGRAM RESULTS

Transcription

1 ORIGINAL ARTICLE COMPARISON OF ESOPHAGEAL SCREEN FINDINGS ON VIDEOFLUOROSCOPY WITH FULL ESOPHAGRAM RESULTS Jacqui E. Allen, MBChB, FRACS, Cheryl White, MA, CCC, Rebecca Leonard, MS, PhD, Peter C Belafsky, MD, PhD Centre for Voice and Swallowing, University of California, Davis, Sacramento, California. Jacqueline.Allen@ucdmc.ucdavis.edu Accepted 14 December 2010 Published online 5 April 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI: /hed Abstract: Background. Videofluoroscopic swallowing studies do not routinely obtain images of the esophagus. We incorporated a single esophageal screening swallow into our videofluoroscopic swallowing study protocol. The purpose of this study was to compare findings from esophageal screening with the results of full esophagram. Methods. Patients undergoing videofluoroscopic swallowing studies with an esophageal screen followed by full esophagram between January 1, 2009, and October 1, 2009, were retrospectively reviewed. Comparison of esophageal screening and esophagram results were undertaken, with esophagram used as the gold standard. Results. Seventy-four patients underwent esophageal screening and esophagram. Sensitivity of esophageal screening is 63% (CI 50%-73%); specificity is 100% (CI 39%-100%). Positive and negative predictive values of esophageal screening were 100% and 13%, respectively. Conclusions. Esophageal screening identified 44/70 (63%) patients with esophageal disease. Esophageal screening is a simple tool that may guide further esophageal investigation. The sensitivity of esophageal screening is limited (63%). If clinical suspicion is high, formal esophagram should be considered. VC 2011 Wiley Periodicals, Inc. Head Neck 34: , 2012 Keywords: Dynamic videofluoroscopic swallowing studies; esophageal screen; esophagram; dysphagia screening; deglutition; deglutition disorders Correspondence to: J. E. Allen This manuscript was presented at the Dysphagia Research Society Meeting, 3-6 March 2010, San Diego, CA. VC 2011 Wiley Periodicals, Inc. Videofluoroscopy is widely performed as a diagnostic procedure in patients with dysphagia or suspected aspiration. It evaluates the oral cavity, oropharynx, hypopharynx and posterior cricoid region, and the proximal cervical esophagus. Not uncommonly, esophageal disease may contribute to symptoms in the throat, upper cervical region, and suprasternal notch. 1 Up to one third of patients with the sensation of cervical dysphagia will have an esophageal cause for the symptom. 1 The comprehensive evaluation of dysphagia therefore should combine a fluoroscopic study of the oropharyngeal phase of deglutition, in our clinic referred to as a dynamic swallow study with esophagram. Addition of esophagram to the dynamic swallow study entails several additional barium boli, a barium tablet, and a watersiphon test. This adds to the overall fluoroscopy screening time and increases the radiation dosage to the patient. We hypothesize that limited esophageal screening may offer a useful alternative to comprehensive esophagram while reducing fluoroscopy screening time in patients without esophageal disease. This type of screening approach has not been previously reported. The purpose of this investigation was to evaluate the sensitivity, specificity, and predictive value of esophageal screening. MATERIALS AND METHODS The study was approved by the Institutional Review Board of University of California, Davis. The charts and fluoroscopic studies of all individuals undergoing a dynamic swallowing study with an esophageal screen followed immediately by comprehensive esophagram, between January 1, 2009, and October 1, 2009, were retrospectively reviewed. Information regarding patient demographics, indications for the procedure, esophageal screening, and esophagram results were recorded in an Excel spreadsheet (Microsoft Corp, Redmond, WA). Dynamic swallow studies and esophageal screening and esophagram results were recorded on a Sony MD-1000 DVD recorder (Sony Corp. America, New York, NY), and were played back with WinDVD7 (Intervideo, Corel Corp., Ottawa, Canada). Esophageal Screening. The esophageal screen was performed after completion of the oropharyngeal phase of the dynamic swallowing study. A single 20- ml liquid barium (EZpaque, Westbury, NJ) bolus was administered and screened in the anteroposterior view. To accommodate the speed of lowering the C-arm 264 Videofluoroscopic Esophageal Screening Versus Esophagram HEAD & NECK DOI /hed February 2012

2 while tracking the bolus, the patient was positioned standing with knees slightly flexed and asked to hold the bolus in the mouth, then swallow all at once and straighten the knees to achieve full standing height. Screening during this maneuver followed the bolus from the oral cavity, down the full length of the esophagus to the stomach. Screening was timed to 15 seconds to allow reasonable time for full clearance of the bolus. Formal Esophagram. Esophagram was performed after completion of the esophageal screening. Our protocol follows: The patient is first evaluated in the upright position. The patient is administered a capful of effervescent agent (EZ-Gas II; E-Z-EM, Lake Success, NY) and is then given a large quantity of high-density liquid barium (E-Z-EM) to consume via a cup. The patient is instructed to take the largest sip comfortable from the cup and consume the barium with one swallow. Initiating a second swallow will arrest and reinitiate esophageal peristalsis and will affect the evaluation of esophageal body motility. Once the barium has cleared the entire esophagus, a second cup sip with identical instructions, is administered. Collapsed and partially collapsed mucosal relief views are obtained. Once the barium has cleared the esophagus, the patient is given a cup of water and asked to swallow a 13-mm barium tablet. The tablet helps to identify a site of obstruction narrower than 13 mm. The examination is recorded under real-time fluoroscopy until the tablet has entered the stomach. If there is significant delay in the passage of the tablet, the fluoroscopy is intermittently turned off to limit fluoroscopy exposure. The patient is then placed prone in the right anterior oblique (RAO) position. The patient is administered a large sip of low-density barium from a straw with similar instructions regarding the importance of just a single swallow. A second sip is then administered. When the barium has cleared the esophagus, the patient is asked to gulp from the cup with multiple swallows to maximally distend the esophagus. If there is barium stasis in the transition zone, the patient is asked to perform 2 dry swallows to relax the upper esophageal sphincter and evaluate for esophagopharyngeal reflux. The patient is then placed supine with their head on a pillow to simulate nocturnal positioning and provocative maneuvers are performed to evaluate for gastroesophageal reflux. The patient is asked to raise and hold their legs 6 inches off of the fluoroscopy table for 10 seconds. While in this position the patient is asked to perform a Valsalva maneuver and is subsequently given water 80 ml to perform the water siphon test. Screening of the gastroesophageal junction is performed, to identify gastroesophageal reflux. This completes the study. The patient is advised to drink three 16-oz glasses of water throughout the remainder of the day to help clear the barium and prevent constipation. Analyses. Each dynamic swallowing study was evaluated according to our unit protocol (previously published). 2 In brief, the protocol includes frame-byframe analysis of images taken from the study data recorded on DVD. Displacement, timing and distension measures are obtained from images and input into an Access database (Microsoft Corp, Redmond, WA). Measures obtained in the lateral view include, maximal opening of the pharyngoesophageal segment (centimeters), duration of opening of the pharyngoesophageal segment (seconds), maximal hyoid approximation to the mandible (centimeters), duration of hyolaryngeal elevation (seconds), pharyngeal area at rest (with a 1-cc bolus held in the oral cavity) and pharyngeal area at maximal constriction (during swallow of a ml bolus) (centimeters squared), and pharyngeal transit time. Measures obtained in the anteroposterior view included maximal pharyngoesophageal segment opening (centimeters), and laterality of bolus flow. Additionally, structural abnormalities such as hiatal hernia, narrowing of the barium column, webs or cricopharyngeal bars were recorded, as was dysfunctional motility such as intraesophageal, gastroesophageal or laryngopharyngeal reflux. Data was recorded and analyzed with SPSS 17.0 for Macintosh (Chicago, IL). The esophageal screen recordings were reviewed by an otolaryngologist blinded to the esophagram results and patient history. The esophagrams were reviewed by a second otolaryngologist blinded to the esophageal screening results and patient history. The esophageal screen was reviewed for bolus transit times and any structural or functional abnormalities. Findings were graded as normal, minor abnormality (finding not requiring further investigation or treatment), or major abnormality (disease requiring intervention, further investigation or ongoing treatment). Minor abnormalities included scant esophageal residue or stasis, isolated tertiary contractions, and nonobstructing cricopharyngeal bar. Major abnormalities included obstructing cricopharyngeal bar, Zenker s diverticulum, marked intraesophageal stasis or reflux, ineffective esophageal motility, delayed esophageal clearing, rings, obstructing webs, pill stasis, dilation of the esophagus, and gastroesophageal or esophagopharyngeal reflux. The findings from the esophageal screening and the esophagram were then compared. Sensitivity and specificity were calculated by use of the formal esophagram results as the gold standard. Differences in detection rates between the esophagram and esophageal screening result were calculated for individual findings, and compared by use of Fisher s exact test. A p value of <.05 was considered significant and all confidence intervals shown are 95% (Table 1). Videofluoroscopic Esophageal Screening Versus Esophagram HEAD & NECK DOI /hed February

3 Table 1. Comparison of esophagram and esophageal screen detection rates for individual fluoroscopic findings. No. (%) Finding Detection rate on esophagram Detection rate on esophageal screen p value Cricopharyngeal bar 14 (19%) 6 (8%).09 Zenker diverticulum 2 (3%) 2 (3%) 1 Intraesophageal reflux or stasis 57 (77%) 29 (39%) <.0001 Hiatal hernia 15 (20%) 5 (7%).03 Ring, web or stricture 9 (12%) 5 (7%).4 Gastroesophageal reflux 16 (22%) 0 (0%) <.0001 Laryngoesophagopharyngeal reflux 6 (8%) 0 (0%).03 Ineffective esophageal motility 19 (26%) 14 (19%).43 The p values in italics show statistical significance. RESULTS Seventy-four patients were identified and included in the study. The mean age (SD) of the cohort was 61 years (15.5 years). Sixty-one percent were female. Referral diagnoses were varied with patients often complaining of more than 1 symptom. The main indications for the procedure were solid food dysphagia (62%, 46/74), cough, or aspiration (38%, 28/74), globus sensation (6.7% 5/74), and cricopharyngeal dysfunction or Zenker diverticulum (6.7%, 5/74). Five patients had been treated for head and neck cancer (7%), three patients (4%) had neurologic disease (Parkinson s disease), and 3 had autoimmune disease (scleroderma and mixed connective tissue disorder). Almost one third (23/74) were being treated for gastroesophageal reflux disease. Esophageal Screening. Forty percent of esophageal screens were normal (30/74), 11% (8/74) revealed minor abnormalities, and 49% (36/74) demonstrated major abnormalities. Findings identified on the esophageal screen included esophageal residue or stasis, isolated tertiary contractions, nonobstructing and obstructing cricopharyngeal bars, Zenker s diverticuli, intraesophageal reflux, ineffective esophageal motility, delayed esophageal clearing, rings, obstructing webs, and dilation of the esophagus (Table 1). Findings that were not identified on the esophageal screen included pill stasis (pill not given during esophageal screening) (Figures 1 and 2), gastroesophageal reflux, and esophagopharyngeal reflux (Table 1). With esophagram used as the gold standard, the sensitivity of the esophageal screening was 62.8% (95% CI ¼ ), and the specificity was 100% (95% CI ¼ ). Positive predictive value is 100%; that is, all abnormal esophageal screening results demonstrated abnormal esophagram results. The negative predictive value was 13%; that is, a negative esophageal screen was a poor indicator of a normal esophagram result. In almost three fourths of subjects (55/74, 74%) the esophageal screening result was either in full agreement, missed only a minor finding that would not have changed management, or directed investigations such that all findings would have become apparent. In 19/ 74 (26%) subjects the esophageal screening result failed to identify a significant finding that would have FIGURE 1. Anteroposterior videofluoroscopic esophageal screen 10 cc bolus swallow, showing no abnormality of barium flow through the upper esophageal sphincter and upper chest. FIGURE 2. Anteroposterior videofluoroscopic view of 13 mm barium tablet stuck at level of Aortic Arch, in patient with a normal esophageal screen swallow in Figure Videofluoroscopic Esophageal Screening Versus Esophagram HEAD & NECK DOI /hed February 2012

4 either required treatment (eg, medication for gastroesophageal reflux) or further investigation (eg, ph/manometry studies or esophagoscopy). It is important to note that in 6 of the 19 subjects where the esophageal screening result was negative, if a 13-mm barium tablet had been given, it would have identified an abnormality that required investigation (Figures 1 and 2). This would have increased the sensitivity of the esophageal screening result to 71%. Esophagram. Esophagram results in this study demonstrated a variety of findings. The most common were esophageal residue or stasis, intraesophageal reflux, ineffective esophageal motility, and delayed esophageal clearing. Other findings included isolated tertiary contractions, nonobstructing and obstructing cricopharyngeal bars, Zenker s diverticuli, rings, obstructing webs, pill stasis, dilation of the esophagus, and gastroesophageal or esophagopharyngeal reflux (Table 1). Six percent (4/74) of esophagram results were completely normal, 20% (15/74) revealed minor abnormalities, and 74% (55/74) demonstrated major abnormalities. DISCUSSION Fluoroscopy is a widely used diagnostic tool for evaluation of deglutition from lips to stomach. Very little has been published about single-swallow esophageal videofluoroscopic screening, as an adjunct to comprehensive pharyngeal videofluoroscopic swallowing studies. Sharitzer et al 3 reported on more than 3000 patients undergoing fluoroscopy with esophageal views that included at least 4 additional barium boluses in 2 different positions. They found 434 (14%) primary esophageal abnormalities and 210 (7%) upper esophageal sphincter abnormalities in patients with complaints of dysphagia, globus, possible aspiration and noncardiac chest pain. 3 Smith et al 1 described multiphasic radiographic examination of the esophagus with a marshmallow bolus for the identification of esophageal rings. At least 10 swallows in different positions were used. Grishaw et al 4 report a barium esophagram consisting of six different boli and positions for evaluation of functional abnormalities of the esophagus in 139 subjects with dysphagia. Although these protocols are very accurate in detecting esophageal problems, a proportion of patients will have normal findings, and some with disease identified on esophageal screening may be better candidates for manometry or esophagoscopy rather than esophagram. Cumulative radiation dosage for esophagram compared with our esophageal screening is at least 6 to 10 times greater and is coupled with the dosage incurred by preceding dynamic swallow study or other imaging studies undertaken by the patient. Radiation Risk. Because lifetime cumulative radiation dose is now a concern when imaging patients, because of mounting evidence suggesting increased risk of malignancy, 5 12 it is incumbent on the clinician to minimize testing that exposes the patient to ionizing radiation. Recent publications confirm the unprecedented rise in radiologic investigations in the United States (estimated at 70 million CT scans in 2007 alone) and the wide range of dosimetry that occurs with these procedures Berrington de González et al 5,6,11 estimate an additional 29,000 cancers will be produced because of medical radiation usage in 2007 alone, whereas other groups have suggested that 1.5% to 2% of all cancers may be due to iatrogenic irradiation. Risks are related linearly to dose exposure and to the organs exposed Lymphoid tissue, the thyroid gland, lung, thymus, breast, spinal cord, and esophagus are all susceptible organs exposed within the screening field when conducting upper aerodigestive tract and esophageal fluoroscopy. Esophageal screening may offer the opportunity to reduce radiation exposure while still being able to evaluate the esophagus for disease in an effective manner. Unlike full esophagram, esophageal screening is short and can be performed easily at the conclusion of a normal dynamic swallowing study. It allows visualization of the pharyngoesophageal segment, esophagus, and esophagogastric junction. Previous work has defined normal transit time through the esophagus as less than 13 seconds. 13,14 Bolus content remaining within the esophagus after this time is considered abnormal and may warrant referral for further investigation such as manometry. Thus screening can be terminated after 13 to 15 seconds, ensuring that additional screening time added to the dynamic swallowing study is negligible and that the overall exposure time of the patient is increased by only 15 seconds. Less than 0.1 msv radiation would be delivered during this length of screening. Compared with background radiation estimates per year (3 msv) or dose received from CT of the abdomen (10 msv), this is a very small incremental increase. 6,7,9,10 Full esophagram, although offering additional information, entails considerably more screening because of the multiple patient positions used and number of swallows administered. Although individual protocols vary, overall dosimetry may be 10 times higher, increasing relative organ doses. Need for Esophageal Imaging in Dysphagia. Traditional modified barium swallow studies have focused on the oral cavity, oropharynx, and pharyngoesophageal regions to evaluate complaints of dysphagia or suspected aspiration. The dynamic swallowing study is an excellent tool for detecting problems occurring above or at the level of the cricopharyngeus. However, one third of patients presenting with complaints of cervical dysphagia, cough, or globus may have an esophageal cause of their symptoms. 1 Smith et al 1 have reported the poor ability of patients to localize Videofluoroscopic Esophageal Screening Versus Esophagram HEAD & NECK DOI /hed February

5 site of disease by their symptoms. In their study, 58% of patients with a distal esophageal ring causing impaction of a marshmallow, had localized symptoms to the neck. An additional 17% localized the symptoms to the mid chest region. 1 A normal dynamic swallowing study in a patient with cervical dysphagia does not exclude esophageal disease. In our investigation, esophageal screening has shown a positive predictive value of 1, meaning that anything abnormal on the screen will demonstrate abnormality on comprehensive esophagram. The negative predictive value is significantly lower. A normal screening result therefore does not allow the clinician to rule out esophageal disease, and further investigation may be required if clinical suspicion is high. Limitations of This Investigation. This study was conducted at a tertiary center for voice and swallowing, serving individuals with severe dysphagia. The study cohort therefore has a high prevalence of oropharyngeal and esophageal disease, which may increase the sensitivity of both the esophagram and esophageal screening. Findings from our institution may therefore not be applicable to the general patient population at another institution. However, because we determined sensitivity and specificity with our own esophagram findings, the accuracy of esophageal screening, in this setting, is still valid. Estimation of screening times and radiation dosage relate only to the protocol at our institution and thus may also vary at other sites. The use of the fluoroscopy C-arm necessitated patient cooperation in performance of esophageal screening; that is, the patient needed to actively stand from a slight squat position so that the bolus could be followed distally. This was primarily the case in patients whose height exceeded the available C-arm coverage. In a setting where a complete view of the pharynx to stomach could be obtained in the same image (eg, a radiology department), obviating the need for patient cooperation, there may have been increased identification of esophagopharyngeal reflux and gastroesophageal reflux by the esophageal screening. Hiatal hernias and esophageal rings are best visualized with distention caused by a large barium bolus. The sample size in our investigation precludes determination of the sensitivity of esophageal screening in detecting hernias and rings. It is possible that, because of the limited bolus administered with esophageal screening, the screen will miss some hernias and rings. The lower rate of detection of cricopharyngeal bar (14 [esophagram] vs 6 [esophageal screening], p ¼.09) on esophageal screening compared with esophagram would be balanced in a clinical context, by identification of most cricopharyngeal bars on the dynamic swallow study portion of the procedure. Analysis of the esophageal screen and esophagram recordings was blinded and conducted by 2 separate otolaryngologists who were not aware of the patient identity or referral diagnosis. Raters analyzed only 1 study type (eg, esophagram or esophageal screening), to avoid contamination of reporting by recognition of abnormal findings. If both types of study had been reviewed, a tendency to interpret data seen in one study (for example, esophagram) into the other study type may have introduced reporting bias. Therefore interrater reliability was not able to be calculated. Because this was a proof of concept study, repeat examination of the studies to determine intrarater reliability was not formally recorded, although examiners were free to watch each study as long or as often as desired to reach a conclusive report. We are now performing a further blinded prospective study that will review intrarater reliability by repetition of study analyses. Improving the Esophageal Screening. Esophageal screening begins at the oropharynx and follows the bolus through the upper esophageal sphincter and esophagus to the lower esophageal sphincter, gastroesophageal junction, and stomach. Because it is not possible to move the fluoroscopic C-arm fast enough to keep up with the bolus, we ask the patient to squat slightly and then to stand up straight as they swallow. As the patient extends their knees, the view of the bolus is followed distally, but the view of the upper esophageal sphincter region can be lost, and thus immediate assessment of esophagopharyngeal reflux is not possible. Likewise gastroesophageal reflux is often minimized when a patient is standing upright (compared with lying down as in the water siphon test of formal esophagram) but might be revealed with a longer screening time than the 10 to 15 seconds used in our screen. A 13-mm barium tablet was not routinely given during esophageal screening, whereas it is always given during formal esophagram. Addition of a 13-mm barium tablet during the esophageal screening would have increased the number of patients in this study subsequently referred for esophagoscopy and increased the sensitivity of the esophageal screening. We hypothesize that routinely incorporating a 13-mm barium tablet in the esophageal screening, and screening for a full 15 to 20 seconds with each swallow, would increase the sensitivity of the esophageal screening. CONCLUSIONS Esophageal screening identified 44/70 (63%) of patients with esophageal disease that might otherwise have been missed. An abnormal screening examination result was 100% predictive of an abnormal esophagram result. In almost three quarters of subjects the esophageal screening result was in close agreement with the esophagram result or predicted the need for further investigation. The esophageal screening may be improved by the inclusion of a 13-mm barium tablet swallow in all patients. Esophageal screening is a 268 Videofluoroscopic Esophageal Screening Versus Esophagram HEAD & NECK DOI /hed February 2012

6 simple tool that may guide further esophageal investigation or management, while reducing overall fluoroscopy time. For example, evidence of obstruction might lead to esophagoscopy, whereas a motility disturbance might be referred for manometry. Sensitivity of the screening, however, is limited, and where clinical suspicion is high, a formal esophagram should be considered. REFERENCES 1. Smith DF, Ott DJ, Gelfand DW, Chen MYM. Lower esophageal mucosal ring: correlation of referred symptoms with radiologic findings using a marshmallow bolus. AJR Am J Roentgend 1998;171: Leonard R, Mackenzie S. Dynamic swallow studies: measurement techniques. In: Leonard R, Kendall K. Dysphagia assessment and treatment planning: a team approach. San Diego: Plural Publishing; 2008: Sharitzer M, Pokieser P, Schober E, et al. Morphological findings in dynamic swallowing studies of symptomatic patients. Eur Radiol 2002;12: Grishaw EK, Ott DJ, Frederick MG, Gelfand DW, Chen MYM. Functional abnormalities of the esophagus: a prospective analysis of radiographic findings relative to age and symptoms. AJR Am J Roentgenol 1996;167: Ron E. Cancer risks from medical radiation. Health Phys 2003;85: Herfarth H, Palmer L. Risk of radiation and choice of imaging. Dig Dis 2009;27: Raelson CA, Kanal KM, Vavilala MD, et al. Radiation dose and excess risk of cancer in children undergoing Neuroangiography. AJR Am J Roentgenol 2009;193: Kim S, Yoshizumi TT, Frush DP, Toncheva G, Yin FF. Radiation dose from cone beam CT in a pediatric phantom: risk estimation of cancer incidence. AJR Am J Roentgenol 2010;194: Fazel R, Krumholz HM, Wang Y et al. Exposure to low-dose ionizing radiation from medical imaging procedures. NEJM 2009;361: Smith-Bindman R, Lipson J, Marcus R, et al. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med 2009;169: Berrington de González A, Mahesh M, Kim KP et al. Projected cancer risks from computed tomographic scans performed in the United States in Arch Intern Med 2009;169: Redberg RF. Cancer risks and radiation exposure from computed tomographic scans. Arch Intern Med 2009;169: Torrico S, Corazziari E, Habib FI. Barium studies for detecting esophagopharyngeal reflux events. Am J Med 2003;115(3A):124S 129S. 14. Barloon TJ, Bergus GR, Lu CC. Diagnostic imaging in the evaluation of dysphagia. Am Fam Physician 1996;53: Videofluoroscopic Esophageal Screening Versus Esophagram HEAD & NECK DOI /hed February