OBJECTIVE : To estimate the effect of the knee bolster use during an MRI on lumbar spinal stenosis parameters and low back pain-related disability. METHODS : A repeated-measurement study of 27 males and 19 females with mean age 55.78 ± 14.36, referred for an MRI of the lumbar spine due to low back pain, performed with and without standard knee bolster. A radiologist evaluated the lumbar lordosis Cobb's angle, the cross-sectional area of the right and left intervertebral foramina and spinal canal at L1-L2, L2-L3, L3-L4 spinal levels. Spinal symptoms were evaluated by the Oswestry Disability Questionnaire. RESULTS : The Cobb angle of lumbar lordosis was found significantly greater on an MRI performed without knee bolster than with bolster (47.30 ± 9.90 vs. 42.57 ± 10.62, p < 0.001). The cross-sectional area of the intervertebral foramina and spinal canal at all evaluated levels was smaller when performed without knee bolster than with bolster. However, differences were significant only at the L1-L2 level and in the spinal canal at all evaluated levels. The Cobb angle, measured with and without knee bolster, showed significant positive correlations with back pain while standing and walking. The spinal canal area without knee bolster showed greater correlations with the Oswestry score parameters than with knee bolster. CONCLUSIONS : MR images at evaluated spinal levels taken without knee bolster showed greater correlations of the spinal canal cross-sectional area with the Oswestry score than ones with knee bolster. Thus, one may prefer MR images acquired without a bolster below the knee compared to an MRI with a knee bolster.