To report accuracy, repeatability, and agreement of Cobb angle measurements on radiographs and/or stereo-radiographs (EOS) compared against one another or against other imaging modalities.
This review follows Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines. A literature search was conducted on 21 July 2021 using Medline, Embase, and Cochrane. Two researchers independently performed title/abstract/full-text screening and data extraction. Studies were eligible if they reported Cobb angles, and/or their repeatability and agreement, measured on radiographs and/or EOS compared against one another or against other imaging modalities.
Of the 2993 records identified, 845 were duplicates and 2212 were excluded during title/abstract/full-text screening. Two more relevant studies were identified from references of eligible studies, leaving 14 studies for inclusion. Two studies compared Cobb angles from EOS vs CT, while 12 compared radiographs vs other imaging modalities: EOS, CT, MRI, digital fluoroscopy, or dual-energy x-ray absorptiometry. Angles from standing radiographs tended to be higher than those from supine MRI and CT, and angles from standing EOS tended to be higher than those from supine or prone CT. Correlations across modalities were strong (R = 0.78–0.97). Inter-observer agreement was excellent for all studies (ICC = 0.77–1.00), except one (ICC = 0.13 radiographs and ICC = 0.68 for MRI).
Differences of up to 11º were found when comparing Cobb angles across combinations of imaging modalities and patient positions. It is not possible, however, to determine whether the differences observed are due to the change of modality, position, or both. Therefore, clinicians should be careful when utilizing the thresholds for standing radiographs across other modalities and positions for diagnosis and assessment of scoliosis.