DOES CHANGING THE ORIENTATION OF A THORACIC RADIOGRAPH AID DIAGNOSIS OF RIB FRACTURES?

Abstract

To test the hypothesis that changing the orientation of a thoracic radiograph aids diagnosis of rib fractures, a case-control study was carried out using 120 radiographs of small animal patients with recent thoracic trauma. Six independent observers representing three levels of experience viewed the radiographs in randomized order in conventional and unconventional orientations at 15-s intervals in a PowerPoint presentation. Sensitivity for rib fractures was 53-69% and specificity was 74-97%, depending on the observer. Radiologists had higher specificity than less experienced observers (P = 0.0001), but comparable sensitivity for rib fractures. Radiologists had significantly higher accuracy than residents (P = 0.03), and residents had higher accuracy than interns (P = 0.02). Accuracy of diagnosis was reduced significantly when based on lateral radiographs compared with dorsoventral (DV)/ventrodorsal (VD) views (P = 0.02) mainly because of low sensitivity of lateral radiographs of dogs with rib fractures. Rotating radiographs 90 degrees clockwise was associated with increased specificity of one intern (from 74 to 90%, P = 0.03), but had no effect on accuracy of diagnosis by other observers (P > 0.16). Turning DV/VD views upside down had no effect on accuracy of diagnosis of any observers (P > 0.40). Changing the orientation of a thoracic radiograph may make it easier for some novices to examine the ribs; hence it could be considered a training aid rather than a technique that will benefit an experienced radiologist.