To analyze whether there is sufficient data from published literature to demonstrate that ultrasound, including elastography, present good metric properties (truth, discrimination and feasibility) in autoimmune myositis (AIM).
A population, intervention, comparator and outcome-structured (PICO) search was performed in Medline, Cochrane Library and Embase database from 01/01/1973 to 08/05/2019. The inclusion criteria required original research involving adult humans, reported in English, assessing ultrasound and elastography in patients with an AIM. Conference abstracts and computer-assisted diagnostics that focused on technique and not ultrasound domains were excluded.
Approximately 2670 articles were identified. Forty-one full-text articles were included in the final analysis. There were 551 AIM patients studied. Eighteen studies (43.9%) had a control group, of which 15 (63.3%) were healthy controls. The age of participants (including controls) varied from 18 to 86 years, and most were females (59%). Diagnosis of AIM was largely biopsy-proven, although some were derived through clinical presentation, positive clinical imaging (ultrasound or otherwise) and/or electromyography and steroid responsiveness. The features examined with ultrasound in the 41 included articles consisted of: muscle echogenicity, bulk, atrophy, architecture, power Doppler, perfusion characteristics, shear wave modulus, shear wave velocity, elasticity index and fasciculations. Twelve studies (29.2%) used quantitative methods to assess these characteristics, whilst others used semi-quantitative, dichotomous/binary and descriptive scoring systems. Criterion validity was met in 14 studies (12/14, 85.7%) and construct validity in 22 studies (22/25, 88.0%). Most published articles reported Level 3b to Level 5 evidence with varying degrees of bias. There was only one longitudinal study examining discrimination. Reliability and feasibility were under-reported.
This is the first systematic review studying the utility of ultrasound, including elastography, in AIM. There is some evidence for criterion and construct validity, suggesting that ultrasound may be a promising outcome measurement instrument in AIM. Agreement on the standardization of acquisition, and the definitions of target domains, is required. Additionally, further validation studies are required to determine discrimination, reliability and feasibility.