Jessica R Mann1, Ged G Wieschhoff2, Ryan Tai3, William C Wrobel4, Nehal Shah2, Jacob C Mandell5. 1. Department of Radiology, Brigham and Women's Hospital, 75 Francis St., Boston, MA, 02115, USA. 2. Division of Musculoskeletal Imaging and Intervention, Brigham and Women's Hospital, 75 Francis St., Boston, MA, 02115, USA. 3. Department of Radiology, University of Massachusetts, 55 Lake Ave North, Worcester, MA, 01655, USA. 4. X-Ray Professional Association, Department of Radiology, Concord Hospital, 250 Pleasant St., Concord, NH, 03301, USA. 5. Division of Musculoskeletal Imaging and Intervention, Brigham and Women's Hospital, 75 Francis St., Boston, MA, 02115, USA. jmandell@bwh.harvard.edu.
Abstract
OBJECTIVE: To compare the diagnostic performance and inter-reader agreement of an abbreviated (5 min) MR protocol compared to a complete (25 min) protocol, for evaluation of suspected tibial bone stress injury. MATERIALS AND METHODS: This IRB-approved retrospective study consisted of 95 consecutive MR examinations in 88 patients with suspected tibial bone stress injury. Three musculoskeletal radiologists independently classified all examinations utilizing both an abbreviated protocol consisting only of axial T2-weighted images with fat suppression, and after a washout period again classified the complete examinations. Accuracy was calculated as proportion of cases classified exactly, within 1 grade, within 2 grades, and also utilizing a simplified "clinically relevant" classification combining grades 2, 3, and 4A into a single group. Significance testing was performed with the chi-test, and a post-hoc power analysis was performed. Inter-reader agreement was calculated with Kendall's coefficient of concordance, with significance testing performed utilizing the z-test after bootstrapping to obtain the standard error. RESULTS AND CONCLUSIONS: There was no significant difference in accuracy of grading tibial bone stress injuries between complete and abbreviated examinations. For complete exams, pooled exact accuracy was 47.8%; accuracy within 1 grade was 82.8%; and accuracy within 2 grades was 96.1%. For the abbreviated protocol, corresponding accuracies were 50.2, 82.0, and 93.9%. With the "clinically relevant" simplified classification, accuracy was 58.6% for complete exams and 64.2% for abbreviated exams. There was no significant difference in inter-reader agreement, with substantial agreement demonstrated for both complete (Kendall coefficient of concordance 0.805) and abbreviated examinations (coefficient of 0.767).
OBJECTIVE: To compare the diagnostic performance and inter-reader agreement of an abbreviated (5 min) MR protocol compared to a complete (25 min) protocol, for evaluation of suspected tibial bone stress injury. MATERIALS AND METHODS: This IRB-approved retrospective study consisted of 95 consecutive MR examinations in 88 patients with suspected tibial bone stress injury. Three musculoskeletal radiologists independently classified all examinations utilizing both an abbreviated protocol consisting only of axial T2-weighted images with fat suppression, and after a washout period again classified the complete examinations. Accuracy was calculated as proportion of cases classified exactly, within 1 grade, within 2 grades, and also utilizing a simplified "clinically relevant" classification combining grades 2, 3, and 4A into a single group. Significance testing was performed with the chi-test, and a post-hoc power analysis was performed. Inter-reader agreement was calculated with Kendall's coefficient of concordance, with significance testing performed utilizing the z-test after bootstrapping to obtain the standard error. RESULTS AND CONCLUSIONS: There was no significant difference in accuracy of grading tibial bone stress injuries between complete and abbreviated examinations. For complete exams, pooled exact accuracy was 47.8%; accuracy within 1 grade was 82.8%; and accuracy within 2 grades was 96.1%. For the abbreviated protocol, corresponding accuracies were 50.2, 82.0, and 93.9%. With the "clinically relevant" simplified classification, accuracy was 58.6% for complete exams and 64.2% for abbreviated exams. There was no significant difference in inter-reader agreement, with substantial agreement demonstrated for both complete (Kendall coefficient of concordance 0.805) and abbreviated examinations (coefficient of 0.767).
Authors: Naveen Subhas; Alex Benedick; Nancy A Obuchowski; Joshua M Polster; Luis S Beltran; Jean Schils; Gina A Ciavarra; Soterios Gyftopoulos Journal: AJR Am J Roentgenol Date: 2017-01-31 Impact factor: 3.959
Authors: Amir Hadid; Daniel S Moran; Rachel K Evans; Yael Fuks; Mark E Schweitzer; Nogah Shabshin Journal: Radiology Date: 2014-07-14 Impact factor: 11.105
Authors: Anousheh Sayah; Ann K Jay; Jacob S Toaff; Erini V Makariou; Frank Berkowitz Journal: AJR Am J Roentgenol Date: 2016-06-08 Impact factor: 3.959
Authors: Alexis A Wright; Eric J Hegedus; Leon Lenchik; Karin J Kuhn; Laura Santiago; James M Smoliga Journal: Am J Sports Med Date: 2015-03-24 Impact factor: 6.202
Authors: Maria G Longo; Joana Fagundes; Susie Huang; William Mehan; Thomas Witzel; Himanshu Bhat; Keith Heberlein; Bruce R Rosen; Daniel Rosenthal; Ramon G Gonzalez; Pamela W Schaefer; Otto Rapalino Journal: J Neuroimaging Date: 2017-06-02 Impact factor: 2.486