Statistical shape modeling of carpal tunnel cross section

Carpal tunnel shape variation has implications for hand and wrist musculoskeletal health. Analyzing this variation can improve understanding of carpal tunnel morphological characteristics. Ultrasound images of the distal carpal tunnel were taken of healthy participants (n = 95). A point-distribution statistical shape model and principal component analysis were used to analyze tunnel shape variation. Feature point location, ellipse axes length, and area were correlated with each principal component (α = 0.05). Principal components 1, 2, and 3 were most strongly correlated with the radial-ulnar location of the thenar muscle ulnar attachment point (r = 0.997, p < 0.001), the radial-ulnar location of the ridge of trapezium (r = -0.988, p = 0.002), and the volar-dorsal location of the hook of hamate (r = -0.935, p = 0.0196). Principal components 2 and 3 were strongly correlated with minor axis length (r = -0.986, p = 0.0019) and area (r = -0.994, p < 0.001). Statistical shape modeling identifies prominent features underlying healthy carpal tunnel shape variation.