© Georg Thieme Verlag KG Stuttgart · New York
Reflux and Barrett’s disease
13 January 2011 (online)
Multispectral scanning during endoscopy guides biopsy of dysplasia in Barrett’s esophagus (Qiu et al., Nat Med 2010 )
In this study, a team of Harvard researchers conducted human testing of a novel technique known as endoscopic polarized scanning spectroscopy (EPSS) in Barrett’s esophagus. This technology relies on the fact that dense and enlarged epithelial nuclei are a key pathological feature of dysplasia. Because dysplastic nuclei scatter light differently from normal nuclei, the back-scattering of spectra can be analyzed to detect foci of dysplasia. Using polarized light, the researchers were able to isolate the spectroscopic signal of the epithelial layer, as light back-scattered from deeper tissues becomes depolarized and can therefore be cancelled out. The EPSS instrument uses a 2.5-mm stainless-steel probe, which can be inserted into the working channel of a standard upper endoscope ([Fig. 1]).
Fig. 1 Endoscopic polarized scanning spectroscopy (EPSS) of Barrett’s esophagus. a Illustration of probe inserted into working channel of upper endoscope with yellow arrows indicating linear rise of probe tip before each scan and rotary motion during scanning. b Endoscopic image showing actual EPSS probe during scanning of Barrett’s esophagus segment showing illumination spot on the esophageal wall at the upper right of the image. (Reproduced with permission from Macmillan Publishers Ltd. ©2010. Qui et al., Nat Med 2010; 16: 603 – 606.)
Scanning of the Barrett’s segment is accomplished by both linear and rotary movement of the probe, which is driven by an external control box. The instrument is able to collect spectra regardless of the orientation and distance between the probe tip and the esophageal wall, so the test results are not altered by peristaltic motion or the level of insufflation of the esophagus.
In this proof of concept study, EPSS data from five patients were compared with pathology results from standard of care 4-quadrant biopsies performed during endoscopic surveillance of Barrett’s esophagus. A total of 95 biopsies were obtained from all patients, revealing 13 dysplastic sites (nine high grade dysplasia [HGD] and four low grade dysplasia [LGD]). The exact site of each biopsy was recorded with reference to the EPSS scanning grid, and the biopsy locations were then overlaid on the spectroscopic results using “pseudocolor maps” ([Fig. 2]).
Fig. 2 Pseudocolor maps produced from endoscopic polarized scanning spectroscopy (EPSS) data overlaid with circles indicating biopsy sites and confirmed pathology. Vertical axis indicates the angle of rotation from the start of each rotary scan, and the horizontal axis indicates distance from upper incisors. Blue and green areas are sites unlikely for dysplasia, and pink/red areas are sites suspicious for dysplasia as determined by EPSS. Color of circles indicates pathology results (green, nondysplastic; pink, low grade dysplasia; red, high grade dysplasia). The two images depicted represent one patient whose initial exam was negative for dysplasia based on standard-of-care 4-quadrant biopsies (i. e. all green circles) but whose EPSS map identified several areas suspicious for dysplasia (pink and red areas). The patient was recalled and three EPSS-guided biopsies confirmed high grade dysplasia. (Reproduced with permission from Macmillan Publishers Ltd. © 2010. Qui et al., Nat Med 2010; 16: 603 – 606.)
Blinded comparison was then performed to test the accuracy of the EPSS analysis at each of the biopsy sites. The EPSS technique had a sensitivity, specificity, and accuracy of 92 %, 96 %, and 96 %, respectively. These test characteristics only reflect the performance of EPSS at each of the biopsy sites, however, not the accuracy of the entire scanned segment; it is also worth noting that localization of biopsies and matching of biopsy sites with EPSS maps was likely imperfect, which may have accounted for some false-negative and false-positive results. Additionally, the authors report that in one patient, the endoscopic examination (with high resolution endoscopy and narrow-band imaging) and 4-quadrant biopsies were negative for suspicious lesions or dysplasia, but the EPSS scan did reveal several probable dysplastic sites and the patient was recalled. EPSS-targeted biopsies were subsequently performed and confirmed HGD at three EPSS-positive sites ([Fig. 2]).
This technique has several benefits over the current standard of care of random 4-quadrant biopsies every 1 – 2 cm. First, because this technology allows for scanning of the entire segment of Barrett’s epithelium for dysplasia, it offers the potential for targeted biopsies at suspicious sites, and more efficient surveillance. Secondly, the possibility of sampling error is reduced, which is a downfall of the current standard of care biopsy approach. Thirdly, if this technology is shown to allow rapid scanning of the entire Barrett’s segment with real-time reporting and ease of interpretation, it may be superior to other advanced microimaging technologies such as narrow-band imaging, optical coherence tomography, and confocal laser microscopy, which have been complicated by imperfect sensitivity and the need for methodical exams and specialized operator training. This technique may also be useful for dysplasia screening and surveillance challenges elsewhere in the gastrointestinal tract, such as high risk gastric cancer screening, gastric ulcer evaluation, polyposis syndromes, and inflammatory bowel disease. As with all new technologies, rigorous prospective studies are required until efficacy and safety can be assured, and practical aspects (e. g. ease of use, reliability of results, operator training needs, cost, and inter-operator variability) can be optimized. However, if the device can really achieve real-time targeting of biopsies for areas highly enriched in dysplasia, this technology is a potential game-changing tool for optimizing dysplasia surveillance in patients with Barrett’s esophagus.