A prognostic model for the presence of neurogenic lesions in atypical idiopathic scoliosis.

Abstract

Consecutive series of patients with idiopathic scoliosis with atypical features. The purpose of this study is to define a specific yet sensitive set of signs and symptoms to indicate the use of MRI in patients with atypical idiopathic scoliosis. Specifically, this study empirically defines a new diagnostic test for the presence of neurogenic lesions based on clinical and radiologic data and then reports the properties of this test in relation to MRI as the gold standard. The reported prevalence of brain stem and spinal cord abnormalities in patients with idiopathic scoliosis associated with atypical features varies from 0% to 60%. This wide range most likely results from the fact that the samples studied are either not well defined or are heterogeneous across studies. Because of these issues, the likelihood of neurogenic lesions in atypical idiopathic scoliosis is not known; consequently, the decision to order an MRI is controversial.METHODS A total of 1,206 patients coded as having idiopathic scoliosis were identified from our institutional database. Of these, 72 patients had one or more atypical features: early-onset scoliosis, atypical curve, severe curve despite immaturity (>45 degrees ), rapidly progressive curve (>1 degrees per month), back pain, headache, or neurologic abnormalities on clinical examination. All 72 patients underwent brain and spinal cord MRI. Logistic regression was used to determine significant predictors of positive MRI and to define the prognostic model. Eleven patients (15%) had abnormal findings on MRI. Eight had an Arnold-Chiari type I malformation associated with a syrinx; 1 had an Arnold-Chiari type I malformation; 1 a syrinx; and 1 a cervical syrinx with a conus lipoma. MRI was positive in 5 of 9 patients (55%) with severe curves despite immaturity. Twenty patients had one or more abnormal neurologic signs. Of these, 8 (40%) had a positive MRI, while only 3 of the 52 patients (6%) with a normal neurologic examination (but other atypical features) had a positive MRI. The most predictive model included the variables neurologic abnormalities (yes or no) and severe curve despite immaturity (yes or no). Using this model, patients with atypical characteristics other than severe curvatures or abnormal neurologic abnormalities(s) had a 3% probability (95% confidence interval [CI], 1-12%) of having a positive MRI; patients with abnormal neurologic change(s), but a nonsevere curve, had a 29% probability of a positive MRI (95% CI, 12-53%) and patients with severe curves and no neurologic change(s) had a 32% probability of positive MRI (95% CI, 8-71%). Patients with both a severe curve and abnormal neurologic change(s) had an 86% probability of positive MRI (95% CI, 46-98%). Agreement between this test and the MRI was 75%, with a sensitivity of 82% (95% CI, 48-97%) and a specificity of 74% (95% CI, 61-83%). The model derived in this study indicates that the probability of neurogenic lesions is extremely low in most patients with idiopathic scoliosis with atypical features. However, patients with severe curves despite skeletal immaturity and an abnormal neurologic examination have a significant probability of neurogenic lesions. Therefore, clinical efficiency will be enhanced by narrowing the indications for MRI to those patients with these risk factors.