How to summarise or combine lumbar MRI-findings for studying associations with low back pain: A modified Delphi study

There is a lack of understanding on how to optimally combine MRI findings to assess their association with low back pain (LBP).

To explore expert opinions on optimal approaches to sum or combine lumbar spine MRI findings, and to identify specific combinations of MRI findings considered most strongly associated with LBP.

A two-round modified Delphi study.

The participants indicated whether a sum score or the most severe finding was most optimal to sum different MRI findings. They also suggested combinations of MRI findings believed to be most strongly associated with 5 different LBP phenotypes (acute/chronic LBP with or without non-radicular leg pain; acute/chronic LBP with radicular leg pain; chronic LBP with bilateral leg pain, and neurogenic claudication). Data were analysed descriptively and visualised using graphs.

Fifty-five multidisciplinary experts participated in round 1 and 39 in round 2. Sum scores were preferred for findings such as disc degeneration, Modic changes, and facet joint degeneration, especially for chronic LBP phenotypes. Using the most severe finding was favoured for nerve root-related or stenotic findings, especially for acute LBP and radicular phenotypes. Suggested combinations varied by phenotype, with disc degeneration, Modic changes, and disc herniation frequent in non-radicular phenotypes, and foraminal stenosis, spondylolisthesis, and nerve root compression more frequent for radicular phenotypes.

The optimal approach to sum or combine lumbar MRI findings depends on the LBP phenotypes. This study provides hypotheses for future research to test the clinical value of combinations of lumbar MRI findings in LBP.