Panoramic ultrasound vs. MRI for the assessment of hamstrings cross-sectional area and volume in a large athletic cohort

We investigated the validity of panoramic ultrasound (US) compared to magnetic resonance imaging (MRI) for the assessment of hamstrings cross-sectional area (CSA) and volume. Hamstrings CSA were acquired with US (by an expert operator) at four different sites of femur length (FL) in 85 youth competitive alpine skiers (14.8 +/- 0.5 years), and successively compared to corresponding scans obtained by MRI, analyzed by a trained vs. a novice rater. The agreement between techniques was assessed by Bland-Altman analyses. Statistical analysis was carried out using Pearson's product moment correlation coefficient (r). US-derived CSA showed a very good agreement compared to MRI-based ones. The best sites were 40% FL (0=mid patellar point) for biceps femoris long head (r=0.9), 50% for semitendinosus (r=0.9), and 30% for semimembranosus (r=0.86) and biceps femoris short head (BFsh, r=0.8). US-based vs. MRI-based hamstrings volume showed an r of 0.96. Poorer r values were observed for the novice compared to the trained rater, with the biggest difference observed for BFsh at 50% (r=0.001 vs. r=0.50, respectively) and semimembranosus at 60% (r=0.23 vs. r=0.42, respectively). Panoramic US provides valid CSA values and volume estimations compared to MRI. To ensure optimal US-vs.-MRI agreement, raters should preferably possess previous experience in imaging-based analyses.