Background This prospective study assessed the accuracy of MRI and ultrasound (US) measurements as a preprocedural assessment tool for predicting clinical loss of resistance depth (CLORD) during fluoroscopy-guided lumbar epidural steroid injections (ESIs).

Materials and methods Sixty patients enrolled received lumbar ESIs at an academic chronic pain clinic. The MRI measurement calculated the distance between the skin and the posterior epidural space, while US measurements included transverse and parasagittal oblique views of the interlaminar space. The epidural space measurements were compared with the CLORD during the performance of the prone epidural injections. The differences in measurements were analyzed using two one-sided tests for equivalency with a 0.5 equivalency margin. The intraclass correlation coefficients between CLORD and the imaging modalities were estimated using mixed effects models.

Results MRI was equivalent to CLORD with a mean difference of –0.2 cm (95% CI –0.39 to –0.11). US transverse and US parasagittal oblique measurements were not equivalent to and underestimated CLORD with mean differences of –0.98 cm (90% CI –1.8 to –0.77) and –0.79 cm (90% CI −1.0 to −5.9), respectively. The intraclass correlation coefficients between MRI and CLORD were the highest at 0.85, compared with 0.65 and 0.73 for transverse and parasagittal oblique US views, respectively.

Conclusions MRI measurements are preferable over US for preprocedural assessment of patients receiving lumbar ESIs for predicting CLORD.