CLINICAL RELEVANCE: Low back pain from lumbar spinal stenosis is a significant source of morbidity, especially among the elderly population. Accurate diagnosis is imperative for effective treatment to be initiated. This paper presents a quantitative method for the evaluation of spinal stenosis that, when used in conjunction with CT and MRI, may greatly aid the clinician in the diagnosis of this debilitating condition. OBJECTIVE: Precise clinical tools for the diagnosis of spinal stenosis are severely lacking. Low back pain and dysfunction derived from lumbar spinal stenosis is a significant source of morbidity, especially among the elderly. Despite its importance, there has been little progress made towards establishing valid, quantitative criteria for the diagnosis of spinal stenosis. We present a new quantitative tool for the diagnosis of lumbar stenosis, the Stenosis Ratio (SR). METHODS: CT scans and MRI scans of 43 patients presenting with clinico-radiographic evidence of lumbar stenosis were used. The patient group consisted of 13 males and 30 females between the ages 49 and 82 with average age of 67. CT and MRI/scans of 43 patients were digitized and computer analyzed. Measurements of SR, defined as the ratio of the cross-sectional dural area of the motion segment to that of the stable segment, were established for L3-L4, L4-L5 and L5-S1 stenotic levels and compared to SR values for a non-stenotic (internal control) level, L2-L3. RESULTS: The L4-L5 level had the lowest SR value of 0.71, followed by 0.74 at L3-L4, and 0.87 at L5-S1. Ninety-five percent confidence intervals of (0.66, 0.81), (0.62, 0.81), and (0.73, 1.00) were found for SR values at levels L3-L4, L4-L5 and L5-S1 respectively. The SR at L2-L3 had a mean value of 1.37 with a 95% confidence interval of (0.970, 1.78). At all levels, SRs were significantly lower for the spinal stenotic L3-S1 levels than for the L2-L3 control as confirmed by a student's t-test (p < 0.05). CONCLUSION: In a select population of patients with spinal stenosis confirmed by neuroradiological assessment, values of SRs were consistently and significantly lower than controls. We believe that measurements of SRs may provide reproducible quantitative measures for the diagnosis of spinal stenosis. SR values below the 95% confidence limit may be indicative of lumbar stenosis. Through the use of ratios, inherent differences in patient size are controlled for, thus allowing comparison of values between patients and treatment groups and effective clinical diagnosis of spinal stenosis.