Volumetric Laser Endomicroscopy With Intelligent Real-time Image Segmentation (IRIS)

Barrett's Esophagus With Dysplasia, Barrett's Esophagus With Dysplasia, Unspecified, Barrett's Esophagus With High Grade Dysplasia, Barrett's Esophagus With Low Grade Dysplasia, Barrett's Esophagus Without Dysplasia

VLE, Volumetric laser endomicroscopy, IRIS, Intelligent real-time image segmentation, Barrett's Esophagus

This is a prospective randomized clinical trial examining how IRIS (Intelligent Real-time Image Segmentation) affects biopsy patterns in VLE (Volumetric laser endomicroscopy).

Patients will undergo a VLE exam with and without IRIS per the standard of care. All patients regardless of the study participation in this study would receive VLE with and without IRIS clinically. They will be randomized into VLE without IRIS first vs VLE with IRIS first. The BUDDY Randomization System, provided for use by the Northwell Health Biostatistics Unit, will be used for this. Randomization will be stratified based on prior diagnosis of dysplasia. The order in which the patient is randomized to will be recorded on the case report form. There will be concealed allocation as the study coordinator performing the randomization will not know the order of the next allocation. Both VLE and IRIS imaging are being performed as standard of care. However, randomization of the order allows for comparison of the two. Regions of interest (ROI) will be recorded using a full scan and recorded on the case report form. The time to identification of ROIs will also be recorded. Each group will then cross over such that the VLE without IRIS group will then have IRIS turned on. ROI will then be recorded for each group based on full scans. A consensus ROI will be recorded based on the two exams. Only one ROI per centimeter will be allowed to avoid overlapping of laser marks. In addition, VLE without IRIS and IRIS marks within 75 frames of each other or 2 hours (on a clock face orientation) will be considered the same target. Laser marking will then be performed. A double laser mark will be applied to all IRIS ROIs. A single laser mark will be applied to VLE ROIs. Targeted biopsies will be taken of all laser marks and placed in separate biopsy jars. Biopsies will be taken in between the laser marks for double laser marked areas. For single laser marks, biopsies will be taken on either side of the laser mark. Resection of visible lesions will then occur per standard of care (if present) followed by random biopsies of the segment. Random biopsies are pinch biopsies every 1 cm the length of the Barrett's in a 4 quadrant fashion per gastrointestinal society guidelines. There are no additional research biopsies being performed outside this study. The targeted biopsies are based on the VLE features that are suspicious for dysplasia and thus standard of care. The other biopsies being performed here are random biopsies which are also being performed for standard of care. Biopsied samples will be stored per standard of care procedures per the pathology department. They will not be stored for future research purposes. Following each procedure, the physician will be asked a series of Likert Scale questions to assess the utility of IRIS in that procedure, including how it impacts their confidence in image interpretation, their ability to assimilate data more easily and quickly, and their overall perception of the technology. Although, the VLE with IRIS and VLE without IRIS are being performed as standard of care, all adverse events will be recorded and reported to the IRB. All data will be entered into a central encrypted REDCap database. Expert pathologists who specialize in gastrointestinal pathology will read the histology. A second gastrointestinal pathologist will confirm any histologic diagnosis of dysplasia. This is the standard of care at Northwell Health and academic medical centers per gastrointestinal society guidelines. All procedures (upper endoscopy, volumetric laser endomicroscopy, and IRIS enhanced VLE) will be performed as standard of care. The prospective data collection and initial randomization will be the research component. All data collected will be de-identified before being transferred into a database. This will be kept in a safe place that is only accessible to the research team.

No related papers found yet.

United States

Participants by arm