Identifying Acute Lumbar Spondylolysis in Young Athletes with Low Back Pain

Abstract

Case-control study.Objective. The aim of this study was to establish an algorithm to distinguish acute lumbar spondylolysis (LS) from nonspecific low back pain (NSLBP) among patients in junior high school by classification and regression tree (CART) analysis. Summary of Background Data. Rapid identification of acute LS is important because delayed diagnosis may result in pseudarthrosis in the pars interarticularis. To diagnose acute LS, magnetic resonance imaging (MRI) or computed tomography is necessary. However, not all adolescent patients with low back pain (LBP) can access these technologies. Therefore, a clinical algorithm that can detect acute LS is needed. Methods. The medical records of 223 junior high school-aged patients with diagnosed acute NSLBP or LS verified by MRI were reviewed. A total of 200 patients were examined for establishing the algorithm and 23 were employed for testing the performance of the algorithm. CART analysis was applied to establish the algorithm using the following data; age, sex, school grades, days after symptom onset, history of LBP, days of past LBP, height, passive straight leg raising test results, hours per week spent in sports activities, existence of spina bifida, lumbar lordosis angle, and lumbosacral joint angle. Sensitivity and specificity of the algorithm and the area under the ROC curve were calculated to assess algorithm performance. Results. The algorithm revealed that sex, days after symptom onset, days of past LBP, hours per week spent in sports activities, and existence of spina bifida were key predictors for identifying acute LS versus NSLBP. Algorithm sensitivity was 0.64, specificity was 0.92, and the area under the ROC curve was 0.79. Conclusion. The algorithm can be used in clinical practice to distinguish acute LS from NSLBP in junior high school athletes, although referral to MRI may be necessary for definitive diagnosis considering the algorithm's sensitivity.

CORR Synthesis: Can Decision Tree Learning Advance Orthopaedic Surgery Research?

“…Aside from predicting postoperative complications, trees have been used diagnostically and operationally. CART distinguished acute lumbar spondylolysis from nonspecific low back pain in 223 patients, with an area under the curve of 0.79, 92% sensitivity, and 92% specificity [4]. A single classification tree identified the most influential intraoperative and postoperative blood transfusion variables in 1029 patients with adult spinal deformities, with an area under the curve of 0.79 [11].…”

Classification of Characteristic Factors of Brain Functional Network Using Decision Tree

“…Mudali et al used C4.5 decision tree to classify the Parkinson's disease and healthy individuals union principal component analysis method with fluorodeoxyglucose positron emission tomography data [5]. Aoyagi et al used the CART algorithm to classify the acute lumbar spondylolysis from non-specific low back pain patients [6]. In addition to using functional data, Zhang et al use the gray matter volume and lateralization index to discuss the classification performances on the ADNI database with structural MRI (sMRI) data [7].…”

Identifying Acute Lumbar Spondylolysis in Young Athletes with Low Back Pain: Retrospective Classification and Regression Tree Analysis

“…Subjects were retrospectively classified as suffering from LBP caused by spondylolysis (if “MRI showed a high-intensity signal at the pars interarticularis… regardless of duration of pain”), or suffering from nonspecific (NS) LBP (if “no remarkable changes were found in the x-ray and MRI”). CART analysis was used to develop an algorithm to distinguish one group from the other based on personal and clinical features 1 …”

“…CART analysis was used to develop an algorithm to distinguish one group from the other based on personal and clinical features. 1 We would like to point out that:1. This design assumes that (''acute'') spondylolysis always causes pain, and that images of spondylolysis rule out the diagnosis of NSLPB.…”

“…CART analysis was used to develop an algorithm to distinguish one group from the other based on personal and clinical features. 1 We would like to point out that:…”