Introduction: The purpose of this study was to assess the added clinical value of oblique knee radiographs four-view (4V) compared to orthogonal anteroposterior (AP) and lateral radiographs in a two-view (2V) series.
Methods: We obtained 200 adult, 4V knee radiographs in 200 patients in the ED and randomly divided them into two groups with 100 series in each group. Ten reviewers - three musculoskeletal radiologists and seven orthopedic surgeons - performed radiograph analyses. These reviewers were randomly divided evenly into group one and group two. Reviewers were blinded to patient data and first reviewed 2V radiographs (AP/lateral) only, and then reviewed 4V radiographs, including AP/lateral, and two additional oblique views for the same patients at least four weeks later. Acute pathology identification and the need for further imaging was assessed for all reviewers, and clinical decision-making (operative vs nonoperative treatment, need for admission, need for additional imaging) was assessed only by the seven orthopaedic surgeon reviewers.
Results: Mean sensitivity for pathology identification was 79% with 2V and 81% with 4V (P =0.25). Intra-observer kappa value was 0.81 (range 0.54-1.00). Additional oblique radiographs led orthopaedic reviewers to change their treatment recommendations in 62/329 patients (18.84%) (P <0.001). Eight of 329 radiographic series were identified as "critical misses." (2.43%) (P =0.004), when pathology was reported as normal or reviewers recommended nonoperative treatment on 2V radiographs but changed their recommendation to operative management after the addition of oblique radiographs. The number needed to treat (NNT) for any treatment change and for "critical misses" was 83 and 643, respectively.
Conclusion: Although the addition of oblique radiographs may improve a clinician's ability to identify subtle pathologic findings not identified on 2V, it rarely leads to significant changes in treatment recommendations. Given the high NNT, limiting the usage of these oblique radiographs in the general patient population may reduce costs without significantly affecting patient care.