No abstract

Aim The objective of this study was to assess whether inversion of grey scale radiograph images has any impact on the diagnostic sensitivity and inter-observer reliability in diagnosing neck of femur fractures. Method We performed a retrospective, single centre study, using 50 randomly selected AP pelvis radiographs. The images included a combination of normal, intracapsular and extracapsular neck of femur fractures, which had been confirmed on CT, MRI and/or subsequent surgery. Four independent observers (two orthopaedic specialists, one registrar and one SHO in orthopaedics) reviewed the images and graded each radiograph image using the Likert scale in response to the statement “there is a fracture”. Following this, the same radiographs were inverted to grey scale images and sent for a second review using the same method. RAND correlation was used for statistical analysis. Results Overall, observers appeared to have similar accuracy and equivocal rates with normal and inversion radiographic imaging. One of the T&O Consultant's had much higher equivocal rate when analysing inverted radiographs compared to conventional digital radiographs but had slightly higher accuracy with inversion (95.35% compared to 93.88%) as well as the Registrar (83.33% compared to 81.63%). The second Trauma and Orthopaedics Consultant had slightly more accuracy with conventional digital radiographs (89.58% compared with 87.50%). Conclusions Inversion of digital radiographs did not affect the diagnostic sensitivity of neck of femur fractures in our study, and similarly did to not have a statistically significant impact in clinician confidence in diagnosing neck of femur fractures.