Michel Kahaleh1, Isaac Raijman2, Monica Gaidhane3, Amy Tyberg3, Amrita Sethi4, Adam Slivka5, Douglas G Adler6, Divyesh Sejpal7, Haroon Shahid3, Avik Sarkar3, Fernanda Martins8, Christine Boumitri9, Samuel Burton9, Helga Bertani10, Paul Tarnasky11, Frank Gress12, Ian Gan13, Jose C Ardengh14, Prashant Kedia11, Urban Arnelo15, Priya Jamidar16, Raj J Shah17, Carlos Robles-Medranda18. 1. Department of Medicine, Rutgers Robert Wood Johnson Medical School, Robert Wood Johnson University Hospital, Rutgers University, 1 RWJ Place, MEB 464, New Brunswick, NJ, 08901, USA. mkahaleh@gmail.com. 2. Gastroenterology, Greater Houston Gastroenterology, Houston, USA. 3. Department of Medicine, Rutgers Robert Wood Johnson Medical School, Robert Wood Johnson University Hospital, Rutgers University, 1 RWJ Place, MEB 464, New Brunswick, NJ, 08901, USA. 4. Columbia University Medical Center, New York, USA. 5. Gastroenterology, University of Pittsburgh Medical Center, Pittsburgh, USA. 6. Gastroenterology, University of Utah Health, Salt Lake City, USA. 7. Baylor College of Medicine Houston, Long Island, USA. 8. Hospital Israelita Albert Einstein, São Paulo, Brazil. 9. Saint Louis University School of Medicine, St. Louis, MO, USA. 10. NuovoOspedale Civile S. Agostino Endoscopy Unit, Dr., Baggiovara, Italy. 11. Digestive Health Associates of Texas, Dallas, USA. 12. Division of Digestive and Liver Diseases, Columbia University, New York, USA. 13. Vancouver General Hospital, GI, Vancouver, Canada. 14. HCFMRP-USP Department of Surgery and Anatomy, Surgery and Anatomy of HCFMRP-USP, São Paulo, Brazil. 15. Department of Upper Gastrointestinal Diseases, Division of Surgery, CLINTEC, Karolinska Institutet, Stockholm, Sweden. 16. Gastroenterology/Medicine, Yale University, Guilford, USA. 17. Gastroenterology, University of Colorado, Aurora, USA. 18. Instituto Ecuatoriano De Enfermedades Digestivas (IECED)-University Hospital Omni, Espiritu Santo University, Guayaquil, Ecuador.
Abstract
BACKGROUND: Digital single-operator cholangioscopy (DSOC) (SpyGlass DS™, Boston Scientific, MA, USA) allows for high-definition imaging of the biliary tree. The superior visualization has led to the development of two different sets of criteria to evaluate and classify indeterminate biliary strictures: the Monaco criteria and the criteria in Carlos Robles-Medranda's publication (CRM). Our objective was to assess the interrater agreement (IA) of DSOC interpretation for indeterminate biliary strictures using the two newly published criteria. METHODS: Forty de-identified DSOC video recordings were sent to 15 interventional endoscopists with experience in cholangioscopy. They were asked to score the videos based on the presence of Monaco Classification criteria: stricture, lesion, mucosal changes, papillary projections, ulceration, white linear bands or rings, and vessels. Next, they scored the videos using CRM criteria: villous pattern, polypoid pattern, inflammatory pattern, flat pattern, ulcerate pattern and honeycomb pattern. The endoscopists then diagnosed the recordings as neoplastic or non-neoplastic based on the criteria. Intraclass correlation (ICC) analysis was done to evaluate interrater agreement for both criteria set and final diagnosis. RESULTS: Recordings of 26 malignant lesions and 14 benign lesions were scored. The IA using both the Monaco criteria and CRM criteria ranged from poor to excellent (range 0.1-0.76) and (range 0.1-0.62), respectively. Within the Monaco criteria, IA was excellent for lesion (0.75) and fingerlike papillary projections (0.74); good for tortuous vessels (0.7), mucosal features (0.62), uniform papillary projections (0.53), and ulceration (0.58); and fair for white linear bands (0.4). Within the CRM criteria, the IA was good for villous pattern (0.62), flat pattern (0.62), and honeycomb pattern; fair for ulcerated pattern (0.56), polypoid pattern (0.52) and inflammatory pattern (0.54). The diagnostic IA using Monaco criteria was good (0.65), while the diagnostic IA using CRM was fair (0.58). The overall diagnostic accuracy using the Monaco classification was 61% and CRM criteria were 57%. CONCLUSION: The IOA and accuracy rate of DSOC using visual criteria from both Monaco Criteria and CRM are similar. However, some criteria from both sets suffer from poor IA, thus affecting the overall diagnostic accuracy. More formal training and refinements in visual criteria with additional validation are needed to improve diagnostic accuracy. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02166099.
BACKGROUND: Digital single-operator cholangioscopy (DSOC) (SpyGlass DS™, Boston Scientific, MA, USA) allows for high-definition imaging of the biliary tree. The superior visualization has led to the development of two different sets of criteria to evaluate and classify indeterminate biliary strictures: the Monaco criteria and the criteria in Carlos Robles-Medranda's publication (CRM). Our objective was to assess the interrater agreement (IA) of DSOC interpretation for indeterminate biliary strictures using the two newly published criteria. METHODS: Forty de-identified DSOC video recordings were sent to 15 interventional endoscopists with experience in cholangioscopy. They were asked to score the videos based on the presence of Monaco Classification criteria: stricture, lesion, mucosal changes, papillary projections, ulceration, white linear bands or rings, and vessels. Next, they scored the videos using CRM criteria: villous pattern, polypoid pattern, inflammatory pattern, flat pattern, ulcerate pattern and honeycomb pattern. The endoscopists then diagnosed the recordings as neoplastic or non-neoplastic based on the criteria. Intraclass correlation (ICC) analysis was done to evaluate interrater agreement for both criteria set and final diagnosis. RESULTS: Recordings of 26 malignant lesions and 14 benign lesions were scored. The IA using both the Monaco criteria and CRM criteria ranged from poor to excellent (range 0.1-0.76) and (range 0.1-0.62), respectively. Within the Monaco criteria, IA was excellent for lesion (0.75) and fingerlike papillary projections (0.74); good for tortuous vessels (0.7), mucosal features (0.62), uniform papillary projections (0.53), and ulceration (0.58); and fair for white linear bands (0.4). Within the CRM criteria, the IA was good for villous pattern (0.62), flat pattern (0.62), and honeycomb pattern; fair for ulcerated pattern (0.56), polypoid pattern (0.52) and inflammatory pattern (0.54). The diagnostic IA using Monaco criteria was good (0.65), while the diagnostic IA using CRM was fair (0.58). The overall diagnostic accuracy using the Monaco classification was 61% and CRM criteria were 57%. CONCLUSION: The IOA and accuracy rate of DSOC using visual criteria from both Monaco Criteria and CRM are similar. However, some criteria from both sets suffer from poor IA, thus affecting the overall diagnostic accuracy. More formal training and refinements in visual criteria with additional validation are needed to improve diagnostic accuracy. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02166099.
Authors: Aatur D Singhi; Marina N Nikiforova; Jennifer Chennat; Georgios I Papachristou; Asif Khalid; Mordechai Rabinovitz; Rohit Das; Savreet Sarkaria; M Samir Ayasso; Abigail I Wald; Sara E Monaco; Michael Nalesnik; N Paul Ohori; David Geller; Allan Tsung; Amer H Zureikat; Herbert Zeh; J Wallis Marsh; Melissa Hogg; Kenneth Lee; David L Bartlett; James F Pingpank; Abhinav Humar; Nathan Bahary; Anil K Dasyam; Randall Brand; Kenneth E Fasanella; Kevin McGrath; Adam Slivka Journal: Gut Date: 2019-04-10 Impact factor: 23.059