Premium
Pilot assessment of a radiologic classification system for segmentation defects of the vertebrae
Author(s) -
Offiah Amaka,
Alman Benjamin,
Cornier Alberto S.,
Giampietro Philip F.,
Tassy Olivier,
Wade Angie,
Turnpenny Peter D.
Publication year - 2010
Publication title -
american journal of medical genetics part a
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.064
H-Index - 112
eISSN - 1552-4833
pISSN - 1552-4825
DOI - 10.1002/ajmg.a.33361
Subject(s) - kappa , scoliosis , medical diagnosis , medicine , multidisciplinary approach , cohen's kappa , confidence interval , group (periodic table) , segmentation , reliability (semiconductor) , surgery , radiology , artificial intelligence , statistics , mathematics , computer science , chemistry , social science , geometry , organic chemistry , sociology , power (physics) , physics , quantum mechanics
Abstract Existing nomenclature systems for describing and reporting congenital segmentation defects of the vertebrae (SDV) are confusing, inconsistently applied, and lack molecular genetic advances. Our aim was to develop and assess a new classification system for SDV. A multidisciplinary group of the International Consortium for Vertebral Anomalies and Scoliosis (ICVAS) developed a new classification system for SDV, and 5 members group (Group 1) independently classified 10 previously unseen cases using this system. Inter‐observer reliability was assessed using kappa, which compares observed agreement with that expected by chance. Seven independent general radiologists unaffiliated with the ICVAS (Group 2) classified the same 10 cases (total, 70 scores) before and after the ICVAS system was explained. We demonstrated the following: Inter‐observer reliability for Group 1 yielded a kappa value of 0.21 (95% confidence intervals (CI) 0.052, 0.366, P = 0.0046); A consensus diagnosis was established for the 10 cases. For Group 2, before the ICVAS system was explained, 1 of 70 scores (1.4%) agreed with the Group 1 consensus diagnoses; Group 2 offered 12 different diagnoses, but 38 of 70 (54.3%) responses were “Don't Know.” After the ICVAS system was explained, 47 of 70 responses (67.1%; 95% CI 55.5, 77.0) agreed with the Group 1 consensus, an improvement of 65.7% (95% CI 52.5, 75.6, P < 0.00005), with no “Don't Know” responses. Group 2 average reporting times, before and after explanation of the ICVAS system, were 148 and 48 min, respectively. We conclude that the ICVAS radiological classification system was found to be reliable and applicable for 10 SDV phenotypes. © 2010 Wiley‐Liss, Inc.