Author(s): BIZET, B., TRINCHI, M., MAGRIS, R., MONTE, A., ZAMPARO, P., Institution: UNIVERSITÀ DI VERONA , Country: FRANCE, Abstract-ID: 820

INTRODUCTION:
At rest, ultrasound imaging most reliable way to analyze muscle architecture is the extended field of view (Franchi et al., 2020). This method could not be applied in dynamic conditions, where muscle architecture characterization usually involves ‘best fitting’ of fascicles and aponeurosis in the field of view (FOV) and a linear extrapolation of both fascicles and aponeurosis outside the FOV. This method can easily be automated, thus reducing the time required for data analysis. Although most fascicles run as a “straight lines” from the proximal to the distal aponeuroses (Narici et al., 1996), fascicles could present a curvilinear path (even at rest) in different regions within a pennate muscle) (Blazevich et al., 2006); in these conditions, automated linear extrapolation could be biased. The aim of this study was to propose and test a new semi-automatic software for tracking “curvilinear fascicles” during dynamic muscle shape changes.
METHODS:
12 healthy adults performed a maximal knee extension at 75°.s-1 on an isokinetic dynamometer. Vastus lateralis ultrasound data were collected in correspondence of the 1st (proximal) and 2nd third (distal) of the thigh by means of a linear array probe (6 cm). B-mode videos were analyzed during the isokinetic phase: i) by using the automated software “ultratrack” (linear extrapolation within and outside the FOV) (Farris & Lichtwark, 2016); ii) by using a new semi-automatic software programmed in Matlab (MLE: manual linear extrapolation). With this software, the visible part of the fascicle was divided into 4 (MLE4) or 2 (MLE2) segments (to take into account the visible curvature) whereas fascicles and aponeurosis outside the FOV were linearly extrapolated (as in the case of the “ultratrack” software). The analysed parameters were: fascicle length (FL), pennation angle (PA), muscle thickness (Th), and belly length (BL). Moreover, one video was analyzed 3 times by the same investigator using MLE4 to perform ICC reliability tests.
RESULTS:
Ultratrack showed higher FL, BL, and Th values and lower PA values compared to MLE4 and MLE2 (about 20% in the proximal and 40% in the distal regions). On the other hand, no significant differences were observed between MLE4 and MLE2 analysis in each of the investigated parameters. Excellent to good absolute concordance, using the 2 factors random model with “single rate” for FL, PA, BL, and Th was found (ICC=0.93; 0.90; 0.93; 0.88 respectively; P<0.001).
CONCLUSION:
Failing to consider the curvilinear behavior in fascicles with significant curvature, as in the distal region of the vastus lateralis, can result in either overestimation or underestimation of muscle geometry. It appears that MLE analysis with a visible fascicle divided into 2 or 4 segments does not affect muscle geometry evaluation.