- Email: [email protected]
656 Esophageal Intraluminal Baseline Impedance Differentiates Gastroesophageal Reflux Disease From Functional Heartburn in Patients With PPI Refractory Symptoms and Correlates With Morphological Changes
Towards Instant Detection of (Pre)Malignant Lesions by Ultrasensitive NearInfrared Fluorescence Endoscopy, Using Molecular Targeted Fluorescent Antibodies Jolien J. Tjalma, P. Beatriz Garcia-Allende, Anton Terwisscha van Scheltinga, Jürgen Glatz, Elmire Hartmans, Marleen van Oosten, Jan J. Koornstra, Jan H. Kleibeuker, Gooitzen M. van Dam, Vasilis Ntziachristos, Wouter B. Nagengast Quantitative targeted endoscopic images of Barrett's neoplasia in vivo. A) White light image shows region of Barrett's esophagus (arrow). B) Fluorescence image collected by molecular endoscope shows increased intensity from region (arrow) after peptide administration. C) Reflectance image of region (arrow) is collected consecutively. D) Quantitative results (blue) are obtained by performing a ratio of the fluorescence (red) and reflectance (green) images.
Introduction and aim: Colonoscopy is considered to be the most sensitive tool to detect colorectal neoplasms, including small and flat lesions. Nevertheless, detection of such lesions is still incomplete, giving rise to interval cancers, particularly in patients with Lynch syndrome and other hereditary disorders. To improve the detection of (pre)malignant lesions nearinfrared endoscopy, guided by molecular targeted fluorescent tracers, may hold great promise. Therefore, we developed a combined white-light and near-infrared (WLNI) endoscopic system, as well as two clinically approved fluorescent labeled antibodies, targeting vascular endothelial growth factor-A (VEGF-A) and epithelial growth factor receptor (EGFR). In the present study we investigated VEGF-A and EGFR expression in adenomas and CRC of Lynch patients, as potential targets for (pre)malignant detection. Subsequently we validated the newly developed WLNI endoscopic system, together with intravenously injected fluorescent anti-VEGF-A and anti-EGFR antibodies, in human CRC mouse models. Materials and Methods: VEGF-A and EGFR expression was analyzed by immunohistochemical staining of 34 CRC, 64 high-grade dysplastic (HGD) adenomas, and adjacent normal colon crypts of Lynch patients. The endoscopic system consisted of a custom made clinically approved WLNI camera, comprising a color camera and an ultra-sensitive camera for concurrent white-light and NIR fluorescence acquisition, attached to either a clinical fiberscope or a multi-modal fiber-optic bundle. The WLNI system was evaluated in vivo in CRC subcutaneous (sc) and intraperitoneal (ip) mouse models of bioluminescent CRC cell lines (HCT116-luc, HT29luc2), using bevacizumab-800CW (anti-VEGF-A) and cetuximab-800CW (anti-EGFR). Results: VEGF-A and EGFR were significantly overexpressed in 95% and 74% of HGD adenomas and in 100% and 85% of CRC, compared to adjacent normal crypts (resp. P,0.001/P,0.05 and P,0.05/P,0.001). Tumor-to-background ratio was high (3.2±0.9 for bevacizumab-800CW and 5.7±2.9 for cetuximab-800CW in the HCT116-luc sc model). WLNI endoscopy demonstrated excellent instant visualization of the sc and ip tumors (diameter ≥1 mm), with clear tumor boundaries and a low background fluorescence, demonstrating very high sensitivity and specificity of our WLNI endoscopic system. Conclusion: VEGF-A and EGFR are attractive targets for molecular targeted fluorescence endoscopy in Lynch patients based on their expression patterns. The newly developed WLNI endoscopic system using clinically approved molecular targeted fluorescent antibodies, enables instant visualization of very small tumor lesions in CRC mice models, with an excellent sensitivity and specificity. These results support clinical evaluation of WLNI endoscopy, in order to enhance early detection of colorectal (pre)malignancies and improve potentially outcome in patients.
655 Optical Sensing of Dysplasia and Field Effect of Carcinogenesis in Barrett's Esophagus Using Elastic-Scattering Spectroscopy Eladio Rodriguez-Diaz, Lisa I. Jepeal, Ashish Sharma, Hiroshi Mashimo, Qin Huang, Sandra R. Cerda, Huihong Xu, Irving J. Bigio, Satish K. Singh OBJECTIVE: Endoscopic surveillance of Barrett's esophagus (BE) is based on random biopsy protocols fraught by sampling error, inability to assess areas of visible concern during endoscopy, and poor consensus among pathologists. Elastic-scattering spectroscopy (ESS), mediated by fiberoptic probes integrated into standard forceps, can assess in vivo and in real time the scattering and absorption properties of tissue. Thus ESS has the potential to be a cost-effective method for guiding mucosal biopsies to dysplastic tissue. The aim of this study is to evaluate the performance of ESS in detecting BE dysplasia, and to investigate whether ESS can sense a field effect of carcinogenesis (FEC) that may be present in BE. METHODS: In an IRB-approved study, patients undergoing endoscopic surveillance for BE dysplasia at VA Boston were examined with ESS. The ESS fiberoptic probe was integrated into biopsy forceps, providing optimal co-registration between spectroscopic measurements and tissue sampling, while being minimally-disruptive to clinical flow. The integrated forceps were used whenever tissue sampling was indicated per current standards of care. Spectra were correlated to the consensus classification of the physical biopsies by H&E staining as determined by three GI pathologists. These labeled spectra were then analyzed using statistical classification computer algorithms. RESULTS: A total of 341 ESS spectra from 39 patients were analyzed: 196 spectra from Barrett's Metaplasia (BM), 21 from Barrett's Neoplasia (BN - LGD, HGD, and cancer), 42 from normal squamous (NS), and 82 from gastric columnar epithelium (GE). Retrospective leave-one-out cross-validation resulted in a sensitivity (Se) of 93% and specificity (Sp) of 94% for distinguishing GE from NS, 79% and 90% for BM vs. NS, and 87% and 84% in BM vs. GE. Distinguishing BN from BM resulted in Se of 95% and Sp of 75%. Subanalysis of 74 spectra from BM in patients with synchronous BN revealed a statistically significant (p , 0.05) increase in the rate of false positives, 0.42 (31/74), when compared to the false positive rate in patients without BN, 0.15 (18/122). A classification algorithm for the FEC in BM biopsies yielded Se of 90% and Sp of 85% for the presence of synchronous BN. CONCLUSION: The increased false positive rate in BM spectra from patients with BN suggests that ESS might be sensitive to changes occurring as a result of a FEC. A targeted approach to BE dysplasia screening using ESS could result in a decreased number of required biopsies. Based on our preliminary results, this reduction would be ~85% in patients with low malignant potential and ~60% in patients with malignant potential, while maintaining a high dysplasia detection rate. This may have profound implications in screening for BE dysplasia and may lead to a reliable, less-invasive, and cost-effective approach.
654 Quantitative Targeted Wide Area In Vivo Imaging of Neoplasia in Barrett's Esophagus Cyrus R. Piraka, Yi-Chen Chen, Xiyu Duan, B. Joseph Elmunzer, Richard S. Kwon, Henry D. Appelman, Scott R. Owens, D. K. Turgeon, Thomas D. Wang Background: The incidence of esophageal adenocarcinoma (EAC) has increased dramatically. Currently, endoscopic surveillance with random 4-quadrant biopsies is recommended. However, regions of high-grade dysplasia (HGD) and EAC can be difficult to identify on white light endoscopy because of lack of architectural features and sampling error. Peptides are promising for use as targeting agents because their high clonal diversity, small size, and broad compatibility with fluorescence dyes, and rapid binding kinetics are compatible for use in real time in vivo "immunohistochemistry." We have previously developed the fluorescentlabeled heptapeptide (ASY*-FITC) and demonstrated specific binding on confocal endomicroscopy in vivo. Aims: We aim to quantify the fluorescence intensity from binding of ASY*FITC to esophageal neoplasia in vivo on wide-field molecular endoscopy. Methods: Patients with Barrett's esophagus undergoing routine endoscopy were recruited for this study, and Prague criteria was used to identify landmarks for histology registration. ASY*-FITC was topically applied to the esophagus via a spray catheter. After a 5 minute incubation period, the unbound peptide was removed with extensive saline rinsing. Fluorescence and reflectance images were collected consecutively in vivo with a novel molecular endoscope (Olympus Medical Systems Corp). The fluorescence and reflectance images were ratioed, and the resulting image was evaluated at thresholds in increments of 10 (range 0 to 100). All pixels above the threshold were defined as target and those below as background. The signal-tonoise ratio (SNR) was calculated by dividing the mean intensity and standard deviation from the target, and the target-to-background ratio was determined by ratioing the mean intensities of the target and background regions. Results: The endoscopic appearance of the lesions found in n=18 patients were identified per Paris classification as 0-Is (n=1), 0-IIa (n=9), and 0-IIb (n=12). The image intensities from autofluorescence, prior to peptide administration, were neglible. Fluorescence was collected from the surface epithelium of squamous (n=6), BE (n=14), LGD (n=2), HGD (n=8), and EAC (n=6). We measured a SNR of 5.9±4.1 and a TBR of 8.5±5.8 from the target region at a threshold of 40. No serious adverse events were observed. Conclusions: We demonstrate a molecular endoscope that is sensitive to fluorescence from topically administered ASY*-FITC peptide in vivo, and show that the intensities can be quantified by correcting for differences in distance to the mucosa. This integrated imaging strategy has potential as a "red flag" technology to guide tissue biopsy.
656 Esophageal Intraluminal Baseline Impedance Differentiates Gastroesophageal Reflux Disease From Functional Heartburn in Patients With PPI Refractory Symptoms and Correlates With Morphological Changes Arne Kandulski, Carlos Caro, Doerthe Kuester, Jochen Weigt, Thomas Wex, Peter Malfertheiner BACKGROUND: Intraluminal baseline impedance (BI) allows to assess mucosal integrity in GERD. In patients with GERD BI levels are impaired. In patients with PPI refractory reflux symptoms, functional heartburn (FH) is the major differential diagnosis of GERD. Whether baseline impedance levels are affected in patients with FH remains uncertain. AIM: To compare baseline impedance levels in patients with refractory symptoms and differentiate patients with GERD from FH. METHODS: 52 patients (male 16, female 36; age 55 years (23-78 years)) with PPI refractory reflux symptoms underwent EGD and MII-pH monitoring at the same day after deescalating and pausing PPI medication at least for 3 weeks. Symptoms were quantified according to the reflux disease questionnaire (RDQ). In recumbent position baseline impedance was assessed at 3, 5, 7, 9, 15 and 17 cm proximal to the lower esophageal sphincter (LES) for at least 30 minutes. Swallows and reflux induced changes were excluded. At upper gastrointestinal endoscopy, biopsies were taken 3 cm above the gastroesophageal junction for histomorphological evaluation of dilated intercellular spaces (DIS), papillary elongation (PE) and basal cell hyperplasia (BCH). RESULTS: Based on endoscopy and MIIpH, GERD was diagnosed in 35 patients (NERD: 19; ERD: 16) and FH in 17 patients. According to symptoms, there were no differences in the RDQ scores between the groups. In the distal esophagus, BI differed significantly at 3 cm (p=0.0006) and 5 cm (p=0.0002). Posthoc analysis reveals significant differences for FH when compared with NERD and ERD (Table 1). In the proximal esophagus, no differences were assessed for BI levels. BI levels correlate negatively with acid exposure time (r:-0.45, p=0.0008), reflux episodes in pH analyses (r:-0.4, p=0.003) as well as number of acidic reflux episodes (r:-0.45, p=0.001) and proximal extend (r:-0.4, p=0.004) in MII analyses. Histomorphologically, DIS in NERD and ERD differed significantly from FH. Correlation analysis revealed a negative correlation of DIS and distal BI levels that marginally missed statistical significance (r:-0.28, p=0.06). CONCLUSIONS: Patients with ERD and NERD have lower distal BI levels as compared with
S-117
AGA Abstracts
AGA Abstracts
653
FH. Impaired BI levels are associated with abnormal acid exposure to the distal esophagus and low BI correlated with DIS which might reflect impaired mucosal integrity.
659
AGA Abstracts
Pepsin in Saliva and Gastroesophageal Reflux Monitoring in 100 Healthy Asymptomatic Subjects Jamal O. Hayat, Etsuro Yazaki, Jin-Yong Kang, Andrew Woodcock, Peter W. Dettmar, Jerry Mabary, Daniel Sifrim Background: Pepsin is a protease originating from pepsinogen secreted into gastric juice from chief cells, only found in the stomach. Its presence in the esophagus or more proximally (pharynx or airways) suggests gastro-esophageal reflux (GER). Several studies have measured pepsin in saliva to determine its value as marker of pathological GER. However, appropriate normal values and correlation with acid/non acid reflux are still limited. The aim of this study was to measure pepsin in expectorated saliva together with objective assessment of GER by pH-impedance in a large cohort of asymptomatic subjects. Methods: 100 healthy subjects, age 30.7 y (range 19-55), with no typical or atypical reflux symptoms underwent MII-pH monitoring "off" PPI. Esophageal pH was measured 5cm above the LES and impedance sensors were positioned at 3,7,9,12 and 15cm above LES. Subjects collected expectorated saliva on waking, one hour after lunch and one hour after dinner. Saliva was collected into tubes containing 0.5ml of 0.01M citric acid and analysed for the presence of pepsin using a lateral flow test comprising two unique human monoclonal antibodies to pepsin (Peptest™, RDBiomed Ltd). The cut off value to determine pepsin positivity was 25ng/ml. Results: Of 300 saliva samples tested, 19% were +ve for pepsin. 64% of subjects had all three saliva samples negative; 20% had 1 sample positive, 12% had 2 samples positive and 4% had 3 samples positive. A similar percentage of samples were positive after lunch (24%) and dinner (22%), but lower on waking (10%). Median acid exposure time was 0.3% (IQR - 0.1-0.8%, 95thcentile 3.5%). Median no. of reflux events was 32 (15-42, 77) being acid 11 (5-22,47) and non-acid 15 (8-25, 46). Saliva samples positive for pepsin were preceded by significantly more reflux events during the 60 min interval before sampling compared to negative samples both after lunch and dinner (+ve pepsin 6 reflux (4-9) vs. -ve pepsin 3 reflux (1-5) p ,0.0001). Supine acid exposure and no. of reflux episodes was not significantly different with +ve or -ve morning samples. Subjects with 3 saliva samples +ve for pepsin had a higher ratio of proximal reflux episodes than subjects with no +ve samples (37%(range 29-40%) vs. 19%(1233%), p,0.02). Only 6/300 samples contained more than 250 ng/ml pepsin. Conclusion: Pepsin was found in the expectorated saliva of a proportion of healthy individuals who did not experience reflux symptoms, particularly post-prandially. However, only 4% of healthy subjects had 3 positive samples. An increased number of reflux episodes were found prior to giving saliva samples with detectable levels of pepsin. Our results suggest that the presence of pepsin in saliva can be a potential screening tool for GERD when at least 3 samples throughout a day are positive or samples contain more than 250 ng/ml pepsin.
657 Use of a Non-Invasive Pepsin Diagnostic Test to Detect GERD: Correlation With MII-pH Evaluation in a Series of Suspected NERD Patients. A Pilot Study Nicola de Bortoli, Edoardo Savarino, Manuele Furnari, Irene Martinucci, Patrizia Zentilin, Lorenzo Bertani, Riccardo Franchi, Massimo Bellini, Vincenzo Savarino, Santino Marchi BACKGROUND: Presence of pepsin in bronchoalveolar lavage fluid, laryngeal biopsy and sputum may be a consequence of gastroesophageal reflux disease (GERD). A novel noninvasive test to detect it in saliva/sputum (PEP-Test) has been proposed to diagnose GERD. A correlation between PEP-Test and multichannel impedance pH monitoring (MII-pH) has never been performed. AIM: The aim was to evaluate the PEP-Test accuracy for the diagnosis of GERD in patients with reflux symptoms by means of MII-pH. PATIENTS AND METHODS: 35 patients with GER symptoms were studied. All patients with negative endoscopy underwent pathophysiological examinations, after wash-out from proton pump inhibitors. Samples of saliva/sputum were obtained by requesting the patient to cough up and spit into a tube containing 0.01 M citric acid within 15 minutes from experiencing reflux symptoms. Patients were grouped on the basis of MII-pH results as follows: True-NERD (increased acid exposure time, AET/reflux number); Hypersensitive Esophagus, HE (normal AET/reflux number, positive symptom association probability index, SAP); no-GERD patients (normal AET/reflux number, negative SAP). Roc curve was performed to obtain diagnostic accuracy of test. RESULTS: Male/Female was 18/17, mean age was 49.8 yrs, mean BMI was 24.9. The mean BMI was similar in three sub-groups. Nine patients were abitudinary smokers and five had a regular alimentary alcohol use. Eleven out of thirty-five patients presented hiatal hernia. No patients showed abnormal esophageal motility. MII-pH results showed: 16 True-NERD patients (median AET 9.5); 12 HE (median AET 3); 7 no-GERD (median AET 1.1). PEPTest was positive in 93.7% of True-NERD, in 58.3% of HE, and negative in 100% of noGERD patients. Accuracy of PEP-Test is reported in Table 1. CONCLUSIONS: PEP-Test is a simple, economic, reproducible, highly specific test to detect the presence of GERD. Accuracy PEP-Test
660 Esophageal Mucosal Impedance Measurement: Time to Move Beyond pH Monitoring for GERD Diagnosis Elif Saritas Yuksel, James C. Slaughter, Tina Higginbotham, Jerry E. Mabary, Michael F. Vaezi INTRODUCTION: Current GERD diagnostics are suboptimal due to limited sensitivity and specificity, and they are constrained by measurement of esophageal reflux during a single time point at a specified location. They measure presence of reflux but do not measure the long-term mucosal consequences of GERD, a significant limitation of existing platforms. The feasibility of a novel, minimally invasive technology to assess esophageal mucosal conductivity changes due to GERD was recently established. The aims of this study were to test the accuracy and reliability of this device: 1) pre- and post-acid suppressive therapy and 2) relative to the standard pH monitoring in detecting GERD. METHODS: Six combinations of single channel Mucosal Impedance (MI) catheters with 360 degree rings were tested by varying ring length to 0.2 or 0.3 mm and ring separation to 0.2, 0.3 or 0.4 mm. Each catheter could be easily traversed through the working channel of an upper endoscope. 98 patients with GERD and controls were enrolled: Erosive Esophagitis (E+) (n=28); pH+/Nonerosive GERD (pH+) (n=30) and pH-/non-erosive controls (n=40). All patients underwent endoscopy and wireless 48-hour pH monitoring 10-days off PPI therapy. During index endoscopy, MI were measured from the esophagitis site as well as 2-, 5- and 10-cm above squamocolumnar junction (SCJ). The optimal MI sensor ring size and separation were determined based on ROC analysis. MI measurements were repeated in those with esophagitis after 2-4months of aggressive PPI therapy to determine test reliability pre- and post-therapy. Predictive value positive of MI measurements were then compared to pH monitoring in diagnosing GERD. RESULTS: Optimum sensitivity (0.9) and specificity (1.0) was achieved by 2mm sensor ring size with separation of 3mm. Median (IQ) mucosal measurements were significantly (p=0.001) lower at the site of erosive esophageal mucosa [1147 (530-1307)] than non-erosive regions at the same level [3654 (2315-6146)]. Among patients with erosive esophagitis, baseline MI [900 (520-1200)] significantly (p,0.001) increased post PPI therapy [4503 (4673-5100)]. The predicted probability of GERD for MI at 540 ohms was 90% and overall MI was superior to pH monitoring for predicting GERD (Figure). CONCLUSIONS: 1) Esophageal mucosal impedance is a reliable indicator of GERD. 2) MI is decreased with mucosal injury and the values increase with healing of mucosa post-acid suppressive therapy. 3) The positive predictive value for GERD is superior with MI than the traditionally employed pH monitoring. 4) The ability of MI device to diagnose GERD simply, efficiently and reliability by through the endoscope application during index endoscopy is an innovative step forward in reflux monitoring.
658 Surgical and Medical Treatment of GERD Lead to a Comparable Increase in Baseline Impedance Levels Nicolaas Fedde Rinsma, Nicole D. Bouvy, Ad Masclee, José M. Conchillo Recent data indicate that baseline esophageal intraluminal impedance levels reflect the electrical resistance of the esophageal wall and may serve as an instrument for the in vivo evaluation of impaired mucosal integrity in GERD patients. Reduction of acid reflux events by therapy may lead to recovery of mucosal integrity as reflected by increased baseline impedance levels. The aim of this study was to evaluate the effect of therapy by acid suppressive medication and through surgical treatment on baseline impedance levels in GERD patients. Methods: In 48 GERD patients (31 males; mean age 45 yrs, range 20-67 yrs) with abnormal acid exposure time (pH ,4 .4.0%) and hiatal hernia ≤2 cm, baseline impedance levels were studied after cessation of acid-suppressive medication for ≥7 days. Endoscopic fundoplication was performed in 33 patients and 24-h pH-impedance monitoring was repeated 6 months after the procedure. Fifteen patients were treated with a daily high dose of PPI and underwent 24-h pH-impedance monitoring on PPI after 6 months. Baseline impedance levels were assessed 3 cm above the lower esophageal sphincter every 2 hours during the 24-h measurement. Endoscopy was performed for assessment of esophagitis. Symptoms were evaluated by a disease-specific quality of life questionnaire (GERD-HRQL). Results: Distal baseline impedance levels at baseline of all patients correlated with acid exposure time and the number of acid and proximal reflux episodes (resp. r:-0.60; p ,0.001, r:-0.42, p,0.01 and r:-0.50; p,0.001), but not with symptom scores. Patients with esophagitis at baseline endoscopy (n=17) showed lower baseline impedance levels compared to patients without esophagitis (n=31) (1416±193 V vs. 1999±167 V, p=0.03). Acid exposure time was reduced by endoscopic fundoplication (from 9.7±0.8 % to 6.7±1.1 %, p ,0.05). Baseline impedance levels were higher after the procedure (from 1908±154 V to 2272±180 V, p,0.05) and correlated with acid exposure time (r:-0.63, p ,0.001). Baseline impedance levels of patients with normalized acid exposure time (pH ,4 ,4.0%) after fundoplication (n=18) were higher than in patients with abnormal acid exposure time (n=15) (2924±211 V vs. 1490±131 V, p,0.001). Acid exposure time was also reduced by PPI treatment (from 12.5±1.8 % to 5.9±1.9 %, p ,0.01) and baseline impedance levels were higher after 6 months (from 1415±221 V to 2487±311 V, p,0.01). Baseline impedance levels after surgical and medical treatment were comparable (2272±180 V vs. 2487±311 V, p=0.53). No correlation was found between baseline impedance levels and symptom scores after endoscopic fundoplication or PPI treatment. Conclusion: Reduction of acid exposure by either endoscopic fundoplication or PPI treatment leads to a comparable increase in baseline impedance levels in GERD patients which may indicate repair of mucosal integrity and therapy success.
AGA Abstracts
S-118
Introduction and aim: Colonoscopy is considered to be the most sensitive tool to detect colorectal neoplasms, including small and flat lesions. Nevertheless, detection of such lesions is still incomplete, giving rise to interval cancers, particularly in patients with Lynch syndrome and other hereditary disorders. To improve the detection of (pre)malignant lesions nearinfrared endoscopy, guided by molecular targeted fluorescent tracers, may hold great promise. Therefore, we developed a combined white-light and near-infrared (WLNI) endoscopic system, as well as two clinically approved fluorescent labeled antibodies, targeting vascular endothelial growth factor-A (VEGF-A) and epithelial growth factor receptor (EGFR). In the present study we investigated VEGF-A and EGFR expression in adenomas and CRC of Lynch patients, as potential targets for (pre)malignant detection. Subsequently we validated the newly developed WLNI endoscopic system, together with intravenously injected fluorescent anti-VEGF-A and anti-EGFR antibodies, in human CRC mouse models. Materials and Methods: VEGF-A and EGFR expression was analyzed by immunohistochemical staining of 34 CRC, 64 high-grade dysplastic (HGD) adenomas, and adjacent normal colon crypts of Lynch patients. The endoscopic system consisted of a custom made clinically approved WLNI camera, comprising a color camera and an ultra-sensitive camera for concurrent white-light and NIR fluorescence acquisition, attached to either a clinical fiberscope or a multi-modal fiber-optic bundle. The WLNI system was evaluated in vivo in CRC subcutaneous (sc) and intraperitoneal (ip) mouse models of bioluminescent CRC cell lines (HCT116-luc, HT29luc2), using bevacizumab-800CW (anti-VEGF-A) and cetuximab-800CW (anti-EGFR). Results: VEGF-A and EGFR were significantly overexpressed in 95% and 74% of HGD adenomas and in 100% and 85% of CRC, compared to adjacent normal crypts (resp. P,0.001/P,0.05 and P,0.05/P,0.001). Tumor-to-background ratio was high (3.2±0.9 for bevacizumab-800CW and 5.7±2.9 for cetuximab-800CW in the HCT116-luc sc model). WLNI endoscopy demonstrated excellent instant visualization of the sc and ip tumors (diameter ≥1 mm), with clear tumor boundaries and a low background fluorescence, demonstrating very high sensitivity and specificity of our WLNI endoscopic system. Conclusion: VEGF-A and EGFR are attractive targets for molecular targeted fluorescence endoscopy in Lynch patients based on their expression patterns. The newly developed WLNI endoscopic system using clinically approved molecular targeted fluorescent antibodies, enables instant visualization of very small tumor lesions in CRC mice models, with an excellent sensitivity and specificity. These results support clinical evaluation of WLNI endoscopy, in order to enhance early detection of colorectal (pre)malignancies and improve potentially outcome in patients.
655 Optical Sensing of Dysplasia and Field Effect of Carcinogenesis in Barrett's Esophagus Using Elastic-Scattering Spectroscopy Eladio Rodriguez-Diaz, Lisa I. Jepeal, Ashish Sharma, Hiroshi Mashimo, Qin Huang, Sandra R. Cerda, Huihong Xu, Irving J. Bigio, Satish K. Singh OBJECTIVE: Endoscopic surveillance of Barrett's esophagus (BE) is based on random biopsy protocols fraught by sampling error, inability to assess areas of visible concern during endoscopy, and poor consensus among pathologists. Elastic-scattering spectroscopy (ESS), mediated by fiberoptic probes integrated into standard forceps, can assess in vivo and in real time the scattering and absorption properties of tissue. Thus ESS has the potential to be a cost-effective method for guiding mucosal biopsies to dysplastic tissue. The aim of this study is to evaluate the performance of ESS in detecting BE dysplasia, and to investigate whether ESS can sense a field effect of carcinogenesis (FEC) that may be present in BE. METHODS: In an IRB-approved study, patients undergoing endoscopic surveillance for BE dysplasia at VA Boston were examined with ESS. The ESS fiberoptic probe was integrated into biopsy forceps, providing optimal co-registration between spectroscopic measurements and tissue sampling, while being minimally-disruptive to clinical flow. The integrated forceps were used whenever tissue sampling was indicated per current standards of care. Spectra were correlated to the consensus classification of the physical biopsies by H&E staining as determined by three GI pathologists. These labeled spectra were then analyzed using statistical classification computer algorithms. RESULTS: A total of 341 ESS spectra from 39 patients were analyzed: 196 spectra from Barrett's Metaplasia (BM), 21 from Barrett's Neoplasia (BN - LGD, HGD, and cancer), 42 from normal squamous (NS), and 82 from gastric columnar epithelium (GE). Retrospective leave-one-out cross-validation resulted in a sensitivity (Se) of 93% and specificity (Sp) of 94% for distinguishing GE from NS, 79% and 90% for BM vs. NS, and 87% and 84% in BM vs. GE. Distinguishing BN from BM resulted in Se of 95% and Sp of 75%. Subanalysis of 74 spectra from BM in patients with synchronous BN revealed a statistically significant (p , 0.05) increase in the rate of false positives, 0.42 (31/74), when compared to the false positive rate in patients without BN, 0.15 (18/122). A classification algorithm for the FEC in BM biopsies yielded Se of 90% and Sp of 85% for the presence of synchronous BN. CONCLUSION: The increased false positive rate in BM spectra from patients with BN suggests that ESS might be sensitive to changes occurring as a result of a FEC. A targeted approach to BE dysplasia screening using ESS could result in a decreased number of required biopsies. Based on our preliminary results, this reduction would be ~85% in patients with low malignant potential and ~60% in patients with malignant potential, while maintaining a high dysplasia detection rate. This may have profound implications in screening for BE dysplasia and may lead to a reliable, less-invasive, and cost-effective approach.
654 Quantitative Targeted Wide Area In Vivo Imaging of Neoplasia in Barrett's Esophagus Cyrus R. Piraka, Yi-Chen Chen, Xiyu Duan, B. Joseph Elmunzer, Richard S. Kwon, Henry D. Appelman, Scott R. Owens, D. K. Turgeon, Thomas D. Wang Background: The incidence of esophageal adenocarcinoma (EAC) has increased dramatically. Currently, endoscopic surveillance with random 4-quadrant biopsies is recommended. However, regions of high-grade dysplasia (HGD) and EAC can be difficult to identify on white light endoscopy because of lack of architectural features and sampling error. Peptides are promising for use as targeting agents because their high clonal diversity, small size, and broad compatibility with fluorescence dyes, and rapid binding kinetics are compatible for use in real time in vivo "immunohistochemistry." We have previously developed the fluorescentlabeled heptapeptide (ASY*-FITC) and demonstrated specific binding on confocal endomicroscopy in vivo. Aims: We aim to quantify the fluorescence intensity from binding of ASY*FITC to esophageal neoplasia in vivo on wide-field molecular endoscopy. Methods: Patients with Barrett's esophagus undergoing routine endoscopy were recruited for this study, and Prague criteria was used to identify landmarks for histology registration. ASY*-FITC was topically applied to the esophagus via a spray catheter. After a 5 minute incubation period, the unbound peptide was removed with extensive saline rinsing. Fluorescence and reflectance images were collected consecutively in vivo with a novel molecular endoscope (Olympus Medical Systems Corp). The fluorescence and reflectance images were ratioed, and the resulting image was evaluated at thresholds in increments of 10 (range 0 to 100). All pixels above the threshold were defined as target and those below as background. The signal-tonoise ratio (SNR) was calculated by dividing the mean intensity and standard deviation from the target, and the target-to-background ratio was determined by ratioing the mean intensities of the target and background regions. Results: The endoscopic appearance of the lesions found in n=18 patients were identified per Paris classification as 0-Is (n=1), 0-IIa (n=9), and 0-IIb (n=12). The image intensities from autofluorescence, prior to peptide administration, were neglible. Fluorescence was collected from the surface epithelium of squamous (n=6), BE (n=14), LGD (n=2), HGD (n=8), and EAC (n=6). We measured a SNR of 5.9±4.1 and a TBR of 8.5±5.8 from the target region at a threshold of 40. No serious adverse events were observed. Conclusions: We demonstrate a molecular endoscope that is sensitive to fluorescence from topically administered ASY*-FITC peptide in vivo, and show that the intensities can be quantified by correcting for differences in distance to the mucosa. This integrated imaging strategy has potential as a "red flag" technology to guide tissue biopsy.
656 Esophageal Intraluminal Baseline Impedance Differentiates Gastroesophageal Reflux Disease From Functional Heartburn in Patients With PPI Refractory Symptoms and Correlates With Morphological Changes Arne Kandulski, Carlos Caro, Doerthe Kuester, Jochen Weigt, Thomas Wex, Peter Malfertheiner BACKGROUND: Intraluminal baseline impedance (BI) allows to assess mucosal integrity in GERD. In patients with GERD BI levels are impaired. In patients with PPI refractory reflux symptoms, functional heartburn (FH) is the major differential diagnosis of GERD. Whether baseline impedance levels are affected in patients with FH remains uncertain. AIM: To compare baseline impedance levels in patients with refractory symptoms and differentiate patients with GERD from FH. METHODS: 52 patients (male 16, female 36; age 55 years (23-78 years)) with PPI refractory reflux symptoms underwent EGD and MII-pH monitoring at the same day after deescalating and pausing PPI medication at least for 3 weeks. Symptoms were quantified according to the reflux disease questionnaire (RDQ). In recumbent position baseline impedance was assessed at 3, 5, 7, 9, 15 and 17 cm proximal to the lower esophageal sphincter (LES) for at least 30 minutes. Swallows and reflux induced changes were excluded. At upper gastrointestinal endoscopy, biopsies were taken 3 cm above the gastroesophageal junction for histomorphological evaluation of dilated intercellular spaces (DIS), papillary elongation (PE) and basal cell hyperplasia (BCH). RESULTS: Based on endoscopy and MIIpH, GERD was diagnosed in 35 patients (NERD: 19; ERD: 16) and FH in 17 patients. According to symptoms, there were no differences in the RDQ scores between the groups. In the distal esophagus, BI differed significantly at 3 cm (p=0.0006) and 5 cm (p=0.0002). Posthoc analysis reveals significant differences for FH when compared with NERD and ERD (Table 1). In the proximal esophagus, no differences were assessed for BI levels. BI levels correlate negatively with acid exposure time (r:-0.45, p=0.0008), reflux episodes in pH analyses (r:-0.4, p=0.003) as well as number of acidic reflux episodes (r:-0.45, p=0.001) and proximal extend (r:-0.4, p=0.004) in MII analyses. Histomorphologically, DIS in NERD and ERD differed significantly from FH. Correlation analysis revealed a negative correlation of DIS and distal BI levels that marginally missed statistical significance (r:-0.28, p=0.06). CONCLUSIONS: Patients with ERD and NERD have lower distal BI levels as compared with
S-117
AGA Abstracts
AGA Abstracts
653
FH. Impaired BI levels are associated with abnormal acid exposure to the distal esophagus and low BI correlated with DIS which might reflect impaired mucosal integrity.
659
AGA Abstracts
Pepsin in Saliva and Gastroesophageal Reflux Monitoring in 100 Healthy Asymptomatic Subjects Jamal O. Hayat, Etsuro Yazaki, Jin-Yong Kang, Andrew Woodcock, Peter W. Dettmar, Jerry Mabary, Daniel Sifrim Background: Pepsin is a protease originating from pepsinogen secreted into gastric juice from chief cells, only found in the stomach. Its presence in the esophagus or more proximally (pharynx or airways) suggests gastro-esophageal reflux (GER). Several studies have measured pepsin in saliva to determine its value as marker of pathological GER. However, appropriate normal values and correlation with acid/non acid reflux are still limited. The aim of this study was to measure pepsin in expectorated saliva together with objective assessment of GER by pH-impedance in a large cohort of asymptomatic subjects. Methods: 100 healthy subjects, age 30.7 y (range 19-55), with no typical or atypical reflux symptoms underwent MII-pH monitoring "off" PPI. Esophageal pH was measured 5cm above the LES and impedance sensors were positioned at 3,7,9,12 and 15cm above LES. Subjects collected expectorated saliva on waking, one hour after lunch and one hour after dinner. Saliva was collected into tubes containing 0.5ml of 0.01M citric acid and analysed for the presence of pepsin using a lateral flow test comprising two unique human monoclonal antibodies to pepsin (Peptest™, RDBiomed Ltd). The cut off value to determine pepsin positivity was 25ng/ml. Results: Of 300 saliva samples tested, 19% were +ve for pepsin. 64% of subjects had all three saliva samples negative; 20% had 1 sample positive, 12% had 2 samples positive and 4% had 3 samples positive. A similar percentage of samples were positive after lunch (24%) and dinner (22%), but lower on waking (10%). Median acid exposure time was 0.3% (IQR - 0.1-0.8%, 95thcentile 3.5%). Median no. of reflux events was 32 (15-42, 77) being acid 11 (5-22,47) and non-acid 15 (8-25, 46). Saliva samples positive for pepsin were preceded by significantly more reflux events during the 60 min interval before sampling compared to negative samples both after lunch and dinner (+ve pepsin 6 reflux (4-9) vs. -ve pepsin 3 reflux (1-5) p ,0.0001). Supine acid exposure and no. of reflux episodes was not significantly different with +ve or -ve morning samples. Subjects with 3 saliva samples +ve for pepsin had a higher ratio of proximal reflux episodes than subjects with no +ve samples (37%(range 29-40%) vs. 19%(1233%), p,0.02). Only 6/300 samples contained more than 250 ng/ml pepsin. Conclusion: Pepsin was found in the expectorated saliva of a proportion of healthy individuals who did not experience reflux symptoms, particularly post-prandially. However, only 4% of healthy subjects had 3 positive samples. An increased number of reflux episodes were found prior to giving saliva samples with detectable levels of pepsin. Our results suggest that the presence of pepsin in saliva can be a potential screening tool for GERD when at least 3 samples throughout a day are positive or samples contain more than 250 ng/ml pepsin.
657 Use of a Non-Invasive Pepsin Diagnostic Test to Detect GERD: Correlation With MII-pH Evaluation in a Series of Suspected NERD Patients. A Pilot Study Nicola de Bortoli, Edoardo Savarino, Manuele Furnari, Irene Martinucci, Patrizia Zentilin, Lorenzo Bertani, Riccardo Franchi, Massimo Bellini, Vincenzo Savarino, Santino Marchi BACKGROUND: Presence of pepsin in bronchoalveolar lavage fluid, laryngeal biopsy and sputum may be a consequence of gastroesophageal reflux disease (GERD). A novel noninvasive test to detect it in saliva/sputum (PEP-Test) has been proposed to diagnose GERD. A correlation between PEP-Test and multichannel impedance pH monitoring (MII-pH) has never been performed. AIM: The aim was to evaluate the PEP-Test accuracy for the diagnosis of GERD in patients with reflux symptoms by means of MII-pH. PATIENTS AND METHODS: 35 patients with GER symptoms were studied. All patients with negative endoscopy underwent pathophysiological examinations, after wash-out from proton pump inhibitors. Samples of saliva/sputum were obtained by requesting the patient to cough up and spit into a tube containing 0.01 M citric acid within 15 minutes from experiencing reflux symptoms. Patients were grouped on the basis of MII-pH results as follows: True-NERD (increased acid exposure time, AET/reflux number); Hypersensitive Esophagus, HE (normal AET/reflux number, positive symptom association probability index, SAP); no-GERD patients (normal AET/reflux number, negative SAP). Roc curve was performed to obtain diagnostic accuracy of test. RESULTS: Male/Female was 18/17, mean age was 49.8 yrs, mean BMI was 24.9. The mean BMI was similar in three sub-groups. Nine patients were abitudinary smokers and five had a regular alimentary alcohol use. Eleven out of thirty-five patients presented hiatal hernia. No patients showed abnormal esophageal motility. MII-pH results showed: 16 True-NERD patients (median AET 9.5); 12 HE (median AET 3); 7 no-GERD (median AET 1.1). PEPTest was positive in 93.7% of True-NERD, in 58.3% of HE, and negative in 100% of noGERD patients. Accuracy of PEP-Test is reported in Table 1. CONCLUSIONS: PEP-Test is a simple, economic, reproducible, highly specific test to detect the presence of GERD. Accuracy PEP-Test
660 Esophageal Mucosal Impedance Measurement: Time to Move Beyond pH Monitoring for GERD Diagnosis Elif Saritas Yuksel, James C. Slaughter, Tina Higginbotham, Jerry E. Mabary, Michael F. Vaezi INTRODUCTION: Current GERD diagnostics are suboptimal due to limited sensitivity and specificity, and they are constrained by measurement of esophageal reflux during a single time point at a specified location. They measure presence of reflux but do not measure the long-term mucosal consequences of GERD, a significant limitation of existing platforms. The feasibility of a novel, minimally invasive technology to assess esophageal mucosal conductivity changes due to GERD was recently established. The aims of this study were to test the accuracy and reliability of this device: 1) pre- and post-acid suppressive therapy and 2) relative to the standard pH monitoring in detecting GERD. METHODS: Six combinations of single channel Mucosal Impedance (MI) catheters with 360 degree rings were tested by varying ring length to 0.2 or 0.3 mm and ring separation to 0.2, 0.3 or 0.4 mm. Each catheter could be easily traversed through the working channel of an upper endoscope. 98 patients with GERD and controls were enrolled: Erosive Esophagitis (E+) (n=28); pH+/Nonerosive GERD (pH+) (n=30) and pH-/non-erosive controls (n=40). All patients underwent endoscopy and wireless 48-hour pH monitoring 10-days off PPI therapy. During index endoscopy, MI were measured from the esophagitis site as well as 2-, 5- and 10-cm above squamocolumnar junction (SCJ). The optimal MI sensor ring size and separation were determined based on ROC analysis. MI measurements were repeated in those with esophagitis after 2-4months of aggressive PPI therapy to determine test reliability pre- and post-therapy. Predictive value positive of MI measurements were then compared to pH monitoring in diagnosing GERD. RESULTS: Optimum sensitivity (0.9) and specificity (1.0) was achieved by 2mm sensor ring size with separation of 3mm. Median (IQ) mucosal measurements were significantly (p=0.001) lower at the site of erosive esophageal mucosa [1147 (530-1307)] than non-erosive regions at the same level [3654 (2315-6146)]. Among patients with erosive esophagitis, baseline MI [900 (520-1200)] significantly (p,0.001) increased post PPI therapy [4503 (4673-5100)]. The predicted probability of GERD for MI at 540 ohms was 90% and overall MI was superior to pH monitoring for predicting GERD (Figure). CONCLUSIONS: 1) Esophageal mucosal impedance is a reliable indicator of GERD. 2) MI is decreased with mucosal injury and the values increase with healing of mucosa post-acid suppressive therapy. 3) The positive predictive value for GERD is superior with MI than the traditionally employed pH monitoring. 4) The ability of MI device to diagnose GERD simply, efficiently and reliability by through the endoscope application during index endoscopy is an innovative step forward in reflux monitoring.
658 Surgical and Medical Treatment of GERD Lead to a Comparable Increase in Baseline Impedance Levels Nicolaas Fedde Rinsma, Nicole D. Bouvy, Ad Masclee, José M. Conchillo Recent data indicate that baseline esophageal intraluminal impedance levels reflect the electrical resistance of the esophageal wall and may serve as an instrument for the in vivo evaluation of impaired mucosal integrity in GERD patients. Reduction of acid reflux events by therapy may lead to recovery of mucosal integrity as reflected by increased baseline impedance levels. The aim of this study was to evaluate the effect of therapy by acid suppressive medication and through surgical treatment on baseline impedance levels in GERD patients. Methods: In 48 GERD patients (31 males; mean age 45 yrs, range 20-67 yrs) with abnormal acid exposure time (pH ,4 .4.0%) and hiatal hernia ≤2 cm, baseline impedance levels were studied after cessation of acid-suppressive medication for ≥7 days. Endoscopic fundoplication was performed in 33 patients and 24-h pH-impedance monitoring was repeated 6 months after the procedure. Fifteen patients were treated with a daily high dose of PPI and underwent 24-h pH-impedance monitoring on PPI after 6 months. Baseline impedance levels were assessed 3 cm above the lower esophageal sphincter every 2 hours during the 24-h measurement. Endoscopy was performed for assessment of esophagitis. Symptoms were evaluated by a disease-specific quality of life questionnaire (GERD-HRQL). Results: Distal baseline impedance levels at baseline of all patients correlated with acid exposure time and the number of acid and proximal reflux episodes (resp. r:-0.60; p ,0.001, r:-0.42, p,0.01 and r:-0.50; p,0.001), but not with symptom scores. Patients with esophagitis at baseline endoscopy (n=17) showed lower baseline impedance levels compared to patients without esophagitis (n=31) (1416±193 V vs. 1999±167 V, p=0.03). Acid exposure time was reduced by endoscopic fundoplication (from 9.7±0.8 % to 6.7±1.1 %, p ,0.05). Baseline impedance levels were higher after the procedure (from 1908±154 V to 2272±180 V, p,0.05) and correlated with acid exposure time (r:-0.63, p ,0.001). Baseline impedance levels of patients with normalized acid exposure time (pH ,4 ,4.0%) after fundoplication (n=18) were higher than in patients with abnormal acid exposure time (n=15) (2924±211 V vs. 1490±131 V, p,0.001). Acid exposure time was also reduced by PPI treatment (from 12.5±1.8 % to 5.9±1.9 %, p ,0.01) and baseline impedance levels were higher after 6 months (from 1415±221 V to 2487±311 V, p,0.01). Baseline impedance levels after surgical and medical treatment were comparable (2272±180 V vs. 2487±311 V, p=0.53). No correlation was found between baseline impedance levels and symptom scores after endoscopic fundoplication or PPI treatment. Conclusion: Reduction of acid exposure by either endoscopic fundoplication or PPI treatment leads to a comparable increase in baseline impedance levels in GERD patients which may indicate repair of mucosal integrity and therapy success.
AGA Abstracts
S-118