Background
The diagnosis of slipped capital femoral epiphysis (SCFE) often is delayed. Although lack of clinical suspicion is the main cause of delayed diagnosis, typical radiographic changes may not be present during the initial phases of SCFE. The peritubercle lucency sign for follow-up of the contralateral hip in patients with unilateral SCFE may be beneficial in assisting the early diagnosis. However, the accuracy and reliability of this sign in patients with SCFE is unknown.
Questions/purposes
(1) What is the accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the peritubercle lucency sign on radiographs for the early diagnosis of SCFE compared with MRI as the gold standard? (2) What are the interobserver and intraobserver reliabilities of the peritubercle lucency sign on radiographs?
Methods
Between 2000 and 2017, 71 patients underwent MRI for an evaluation of pre-slip or a minimally displaced SCFE. Sixty percent of hips (43 of 71) had confirmed SCFE or pre-slip based on the presence of hip pain and MRI changes, and these patients underwent in situ pinning. Three independent experienced observers reviewed MR images of the 71 hips and agreed on the presence of a juxtaphyseal bright-fluid signal suggesting bone marrow edema in these 43 hips with SCFE, and absence MRI changes in the remaining 28 hips. The same three experienced observers and two inexperienced observers, including a general radiologist and an orthopaedic surgery resident, blindly assessed the radiographs for the presence or absence of the peritubercle lucency sign, without information about the diagnosis. Diagnostic accuracy measures including sensitivity, specificity, PPV, and NPV were evaluated. Intraobserver and interobserver agreements were calculated using kappa statistics.
Results
The overall accuracy of the peritubercle lucency sign on radiographs was 94% (95% CI 91 to 96), sensitivity was 97% (95% CI 95 to 99), specificity was 89% (95% CI 90 to 96), PPV was 93% (95% CI 90 to 96), and NPV was 95% (95% CI 92 to 99). All accuracy parameters were greater than 85% for the five observers, regardless of experience level. Intraobserver agreement was perfect (kappa 1.0), and interobserver agreement was excellent for the peritubercle lucency sign on radiographs across the five observers (kappa 0.81 [95% CI 0.73 to 0.88]). The reliability was excellent for experienced observers (kappa 0.88 [95% CI 0.74 to 1.00]) and substantial for inexperienced observers (kappa 0.70 [95% CI 0.46 to 0.93]), although no difference was found with the numbers available (p = 0.18).
Conclusions
The peritubercle lucency sign on radiographs is accurate and reliable for the early diagnosis of SCFE compared with MRI as the gold standard. Improving the early diagnosis of SCFE may be possible with increased awareness, high clinical suspicion, and a scrutinized evaluation of radiographs including an assessment of the peritubercle lucency sign.
Level of Evidence
Level III, diagnostic study.
