Do Oblique Views Add Value in the Diagnosis of Spondylolysis in Adolescents?
Nicholas A. Beck, BS1; Robert Miller, BS1; Keith Baldwin, MD, MSTP, MPH1; X. Zhu, MS1; David Spiegel, MD1; Denis Drummond, MD1; Wudbhav N. Sankar, MD1; John M. Flynn, MD1
1 Division of Orthopaedic Surgery (N.A.B., R.M., K.B., D.S., D.D., W.N.S., J.M.F.) and Department of Radiology (X.Z.), The Children’s Hospital of Philadelphia, 34th and Civic Center Boulevard, Philadelphia, PA 19104. E-mail address for J.M. Flynn: firstname.lastname@example.org
View Disclosures and Other Information
J Bone Joint Surg Am, 2013 May 15;95(10):e65 1-7. doi: 10.2106/JBJS.L.00824
Anteroposterior, lateral, and right and left oblique lumbar spine radiographs are often a standard part of the evaluation of children who are clinically suspected of having spondylolysis. Recent concerns regarding radiation exposure and costs have brought the value of oblique radiographs into question. The purpose of the present study was to determine the diagnostic value of oblique views in the diagnosis of spondylolysis.
Radiographs of fifty adolescents with L5 spondylolysis without spondylolisthesis and fifty controls were retrospectively reviewed. All controls were confirmed not to have spondylolysis on the basis of computed tomographic scanning, magnetic resonance imaging, or bone scanning. Anteroposterior, lateral, and right and left oblique radiographs of the lumbar spine were arranged into two sets of slides: one showing four views (anteroposterior, lateral, right oblique, and left oblique) and one showing two views (anteroposterior and lateral only). The slides were randomly presented to four pediatric spine surgeons for diagnosis, with four-view slides being presented first, followed by two-view slides. The slides for twenty random patients were later reanalyzed in order to calculate of intra-rater agreement. A power analysis demonstrated that this study was adequately powered. Inter-rater and intra-rater agreement were assessed on the basis of the percentage of overall agreement and intraclass correlation coefficients (ICCs). PCXMC software was used to generate effective radiation doses. Study charges were determined from radiology billing data.
There was no significant difference in sensitivity and specificity between four-view and two-view radiographs in the diagnosis of spondylolysis. The sensitivity was 0.59 for two-view studies and 0.53 for four-view studies (p = 0.33). The specificity was 0.96 for two-view studies and 0.94 for four-view studies (p = 0.60). Inter-rater agreement, intra-rater agreement, and agreement with gold-standard ICC values were in the moderate range and also demonstrated no significant differences. Percent overall agreement was 78% for four-view studies and 82% for two-view studies. The radiation effective dose was 1.26 mSv for four-view studies and 0.72 mSv for two-view studies (difference, 0.54 mSv). The charge for four-view studies was $145 more than that for two-view studies.
There is no difference in sensitivity and specificity between four-view and two-view studies. Although oblique views have long been considered standard practice by some, our data could not identify a diagnostic benefit that might outweigh the additional cost and radiation exposure.