Abstract
INTRODUCTION: Spontaneous resorption of a herniated lumbar disc and disappearance of clinical symptoms without repair is a well-known but not well-studied phenomenon. This prospective study uses magnetic resonance images to search for predictors as to the possibility of spontaneous herniation resorption without any intervention and patients' predisposition to benefit from conservative treatment. MATERIALS AND METHODS: Of the 125 patients examined by magnetic resonance imaging, 22 had clinical symptoms that spontaneously (without any intervention) disappear. The spinal condition of each was classified using Fardon's, Modic's, and Pfirman's classifications, and physical dimensions of the affected disc and herniated disc were measured. Inter- and intrareproducibility of this measurement were determined. Predictors for spontaneous disappearance of clinical symptoms were selected using multivariable logistic regression and receiver operating characteristic (ROC) analysis. RESULTS: The measurement uncertainty was less than four pixels for most parameters. Fardon's classification and middle height of the affected disc were the only clinically relevant parameters statistically proven to be predictors of clinical symptoms resolution (p < 0.01). By combining these appropriately, we are able to identify a group of patients who have up to 22.5 times greater chance of spontaneous regression compared to others (odds ratio = 22.5, p < 0.001). CONCLUSION: By a suitable combination of the two parameters, we can select patients who are suitable candidates for conservative treatment of lumbar disc herniation and unlikely to require surgery.