Radiographic flexor cortical lysis indicates advanced degenerative change and its earlier recognition may improve case outcome. Aims of this prospective, diagnostic accuracy study were to determine effects of radiographic beam angle and observer on accuracy of lesion detection.
The sample included 36 limbs from 31 horses. Palmaroproximal-palmarodistal oblique (skyline) radiographs were acquired at standard (n = 38) and more shallow (alternate) angles (n = 58). Images were independently reviewed by four experience levels of five observers each (n = 20) for the presence and severity of flexor cortical lysis. Observers also reported their confidence in these answers. Responses were compared based on seeing a standard skyline or multiple projections. The definitive presence (or absence) and severity of lysis was based upon radiologist consensus agreement. When assessed by observer, the identification of lysis and the assessment of its severity was most similar to that of radiologists when observers of all levels of experience were able to view multiple skyline projections (P = 0.399, P = 0.174).
Using multiple views to detect lysis resulted in improved sensitivity (85.3% vs. 97.2%, P < 0.001), decreased specificity (82.8% vs. 74.5%, P = 0.03), and improved interobserver agreement (86.0% vs. 90.2%, P = 0.21). On average, observers of all levels of experience became more confident viewing multiple projections (P < 0.001).
Skyline radiographs using a flatter angle of incidence improve radiographic detection and characterization of flexor cortical lysis severity, may allow the beam to be tangential to the more distal portion of the navicular bone, and are recommended when flexor cortical lysis is suspected.
