Ten embalmed human cadavers were selected. The lateral plateau was sagitally sectioned, and the joint was reduced under direct visualization. Lateral, anterior-posterior (AP), and joint line fluoroscopic views were obtained. The same fluoroscopic views were obtained with 2 mm displacement and 5 mm displacement. The images were randomised, and eight orthopaedic traumatologists were asked whether the plateau was reduced. Within each pair of conditions (view and displacement from 0 mm to 5 mm) sensitivity, specificity, and intraclass correlations (ICC) were evaluated.
The AP-lateral view with 5 mm displacement yielded the highest accuracy for detecting reduction at 90% (95% CI: 83–94%). For the other conditions, accuracy ranged from (37–83%). Sensitivity was highest for the reduced lateral view (79%, 95% CI: 57–91%). Specificity was highest in the AP-lateral view 98% (95% CI: 93–99%) for 5 mm step-off. ICC was perfect for the AP-lateral view with 5 mm displacement, but otherwise agreement ranged from poor to moderate at ICC = 0.09–0.46. Finally, there was no additional benefit to including the joint-line view with the AP and lateral views.
Using both AP and lateral views for 5 mm displacement had the highest accuracy, specificity, and ICC. Outside of this scenario, agreement was poor to moderate and accuracy was low. Applying this clinically, direct visualization of the articular surface may be necessary to ensure malreduction less than 5 mm.