•  
  •  
 

RELIABILITY OF MANUAL VS AUTOMATIC ULTRASOUND ANALYSES

Abstract

Kealey J. Wohlegemuth1, Malia M. Blue2, Jacob A. Mota1. 1University of Alabama, Tuscaloosa, AL. 2University of North Carolina at Chapel Hill, Chapel Hill, NC.

BACKGROUND: Brightness-mode (B-mode) ultrasonography has become a popular imaging modality. Ultrasonography may be employed to examine skeletal muscle architecture (i.e., pennation angle [PA] and fascicle length [FL]), which may provide unique insights to muscle function. Unfortunately, manually analyzing ultrasound scans is a time-consuming task. An automated analysis software program has been recently published, which may be a useful tool to increase reliability and speed of ultrasound muscle architecture analyses. The purpose of this project was to compare the reliability of manual versus automatic ultrasound analyses of muscle architecture. METHODS:Twenty-three participants (23 ± 4 yrs) completed one laboratory visit which consisted of two separate data collection trials separated by 10 minutes. During each trial, participants had measures of the vastus lateralis (VL) PA and FL assessed with a B-Mode ultrasound imaging device. Ultrasound image settings were held constant (i.e., depth = 6 cm, frequency = 12 mHz, gain = 52 dB) for each participant. Muscle architecture images were taken by scanning the length of the VL longitudinally, in conjunction with the extended field of view algorithm. All images were manually analyzed in an open-source imaging program, ImageJ, after pixel scaling. The automatic analyses were completed using the Simple Muscle Architecture (SMA) 17 macro function in FIJI. Both manual and automatic analyses were completed on the same image for trial 1 and trial 2, respectively. Test-retest reliability statistics (i.e., intraclass correlation coefficient [ICC] model 2, 1, standard error of measure expressed as a percentage of the mean [SEM%], and the minimal differences [MD] values needed to be considered real) were calculated for each variable. RESULTS: Manual analyses for PA (ICC2,1 = 0.75, SEM (%) = 9.61%, MD = 5.18) were more reliable than automated PA analyses (ICC2,1 = 0.27, SEM (%) = 28.27%, MD = 13.39). Similarly, manual analyses for FL (ICC2,1 = 0.86, SEM (%) = 8.05%, MD = 1.62) were more reliable than automated FL calculations (ICC2,1 = 0.16, SEM (%) = 26.90%, MD = 5.89). CONCLUSION: The outcomes of this study suggest that this specific automated analysis tool may be less reliable when compared to our manual ultrasound analysis technique when measuring muscle architecture of the VL.

This document is currently not available here.

Share

COinS