Is Handheld Lower-Resolution Ultrasound Evaluation Sufficient to Measure the Cross-Sectional Area of the Median Nerve?

Abstract

This study aimed to assess the accuracy of median nerve cross-sectional area (CSA) measurements using a handheld ultrasound (HHUS) and a cart-based high-resolution ultrasound. Fifteen healthy subjects were prospectively enrolled. Each subject underwent standardized ultrasound evaluations of their median nerves with HHUS and a high-resolution ultrasound by three different providers with different levels of training in musculoskeletal ultrasound. CSA calculations were performed using a standardized method. The primary outcome was the mean difference in CSA (ΔCSA) between the gold standard high-resolution ultrasound and the lower resolution handheld ultrasound, with a noninferiority margin set at ΔCSA < 2 mm². Inter-rater reliability was calculated using Lin concordance correlation coefficient (LCCC), and the intraclass coefficient (ICC) was used to calculate intra-rater reliability. The average CSA of the median nerve across all patients and reviewers using the high-resolution US was > 0.9 mm² for the HHUS. In paired comparisons within subjects, the ΔCSA of the median nerve using the HHUS was 1.8 mm² (95% CI, 0.8-2.7 mm²), and the LCCC was weak 0.258 (0.018-0.470). Noninferiority was not achieved. The intrarater reliability between devices was poor at 0.201. However, the inter-rater reliability among different examiners was good when using the high-resolution device, but it was poor using the HHUS. The findings suggest caution for the use of handheld ultrasound in measurements of median nerve CSA because the difference between HHUS and gold standard high-resolution ultrasound was not within the noninferiority margin of 2 mm². Diagnostic II.