ACSAuto-semi-automatic assessment of human vastus lateralis and rectus femoris cross-sectional area in ultrasound images

Abstract

Open-access scripts to perform muscle anatomical cross-sectional area (ACSA) evaluation in ultrasound images are currently unavailable. This study presents a novel semi-automatic ImageJ script (named “ACSAuto”) for quantifying the ACSA of lower limb muscles. We compared manual ACSA measurements from 180 ultrasound scans of vastus lateralis (VL) and rectus femoris (RF) muscles to measurements assessed by the ACSAuto script. We investigated inter- and intra-investigator reliability of the script. Consecutive-pairwise intra-class correlations (ICC) and standard error of measurement (SEM) with 95% compatibility interval were calculated. Bland–Altman analyses were employed to test the agreement between measurements. Comparing manual and ACSAuto measurements, ICCs and SEMs ranged from 0.96 to 0.999 and 0.12 to 0.96 cm2 (1.2–5.9%) and mean bias was smaller than 0.5 cm2 (4.3%). Inter-investigator comparison revealed ICCs, SEMs and mean bias ranging from 0.85 to 0.999, 0.07 to 1.16 cm2 (0.9–7.6%) and − 0.16 to 0.66 cm2 (− 0.6 to 3.2%). Intra-investigator comparison revealed ICCs, SEMs and mean bias between 0.883–0.998, 0.07–0.93 cm2 (1.1–7.6%) and − 0.80 to 0.15 cm2 (− 3.4 to 1.8%). Image quality needs to be high for efficient and accurate ACSAuto analyses. Taken together, the ACSAuto script represents a reliable tool to measure RF and VL ACSA, is comparable to manual analysis and can reduce time needed to evaluate ultrasound images.

DeepACSA: Automatic Segmentation of Cross-Sectional Area in Ultrasound Images of Lower Limb Muscles Using Deep Learning

“…Furthermore, ACSAuto is somewhat subjective because the user must validate the proposed muscle outline. In addition, not correcting the suggested outline led to large measurement errors compared with manual analysis and thus unusable results (20). By using trainable CNN, DeepACSA is more robust to variation in ultrasound image pixel characteristics because the whole image texture is considered, and more complex features are computed.…”

“…Therefore, image quality is still an important factor for correct muscle ACSA segmentation using deep learning. Images for which muscle ACSA was predicted incorrectly should be analyzed using ASCAuto (20) or manually.…”

“…To our knowledge, the semiautomated ACSAuto program (20) is the only open-access tool available for analyzing panoramic lower limb ultrasound images. Thus, it should be determined whether CNN approaches could be used to fully automate the process of ACSA analysis from panoramic ultrasound images.…”

“…However, manual analysis is subjective, laborious, and requires great experience (18,19). Semiautomated and automated algorithms have been developed to accelerate the ACSA segmentation process and reduce the subjectivity of manual segmentation in lower limb muscle ultrasound images (20,21). These approaches used sophisticated image processing steps to localize features and objects within an image.…”

Innovations in the Assessment of Skeletal Muscle Health: A Glimpse into the Future

“…However, emerging technologies are likely to streamline the measurement process, reduce subjectivity, and enhance the accuracy of echo intensity analysis. Automated or semi-automated region of interest selection algorithms have recently been introduced to target specific muscle regions [87,88]. In the future, computerized analysis of ultrasound images will enable the precise quantification of echo intensity values, ensuring consistent and reproducible evaluations.…”