Reliability of panoramic ultrasound imaging and agreement with magnetic resonance imaging for the assessment of lumbar multifidus anatomical cross-sectional area

Abstract

The aim of this study was to investigate the reliability of panoramic ultrasound (US) imaging and agreement with magnetic resonance imaging (MRI) for assessing the average lumbar multifidus anatomical cross-sectional area between the lumbar vertebral bodies L3–L5 (i.e., LMF ACSA_L3–L5). US and MRI scans of 20 male youth competitive alpine skiers were collected. To test the intra- and interrater reliability of US, transversal panoramic scans were analyzed on two different days by the same rater and the analysis of the first day was compared with the analysis of a second rater. To examine the agreement between US and MRI, Bland–Altman analysis was performed. Intrarater reliability was excellent, and interrater reliability was weak to good for both sides. The bias between MRI and US was − 0.19 ± 0.90 cm² (2.68 ± 12.30%) for the left side and − 0.04 ± 0.98 cm² (− 1.11 ± 12.93%) for the right side (i.e., for both sides US slightly overestimated LMF ACSA_L3–L5 on average). The limits of agreement were − 1.95 to 1.57 cm² (− 26.70 to 21.30%) for the left side and − 1.95 to 1.88 cm² (− 26.46 to 24.24%) for the right side. Panoramic US imaging may be considered a method with excellent intrarater and weak to good interrater reliability for assessing LMF ACSA_L3–L5. Comparison with MRI showed large individual differences in some cases, but an acceptable bias between the two imaging modalities.