Assessment of clavicle muscle attachment site variability through a non-rigid registration framework

Last decade, a shift towards operative treatment of midshaft clavicle fractures has been observed [T. Huttunen et al., Injury, 2013]. Current fracture fixation plates are however suboptimal, leading to reoperation rates up to 53% [J. G. Wijdicks et al., Arch. Orthop. Trauma Surg, 2012]. Plate irritation, potentially caused by a bad geometric fit and plate prominence, has been found to be the most important factor for reoperation [B. D. Ashman et a.l, Injury, 2014]. Therefore, thin plate implants that do not interfere with muscle attachment sites (MAS) would be beneficial in reducing plate irritation. However, little is known about the clavicle MAS variation. The goal of this study was therefore to assess their variability by morphing the MAS to an average clavicle. 14 Cadaveric clavicles were dissected by a medical doctor (MH), laser scanned (Nikon, LC60dx) and a photogrammetry was created with Agisoft photoscan (Agisoft, Russia). Subsequently a CT-scan of these bones was acquired and segmented in Mimics (Materialise, Belgium). The segmented bone was aligned with the laser scan and MAS were indicated in 3-matic (Materialise, Belgium). Next, a statistical shape model (SSM) of the 14 segmented clavicles was created. The average clavicle from the SSM was then registered to all original clavicle meshes. This registration assures correspondences between source and target mesh. Hence, MAS of individual muscles of all 14 bones were indicated on the average clavicle. Mean area is 602 mm2 ± 137 mm2 for the deltoid muscle, 1022 mm2 ±207 mm2 for the trapezius muscle, and 683 mm2 ± 132 mm2 for the pectoralis major muscle. The sternocleidomastoid muscle has a mean area of 513 mm2 ± 190 mm2 and the subclavius muscle had the smallest mean area of 451 mm2 ± 162 mm2. Visualization of all MAS on the average clavicle resulted in 72% coverage of the surface, visualizing only each muscle's largest MAS led to 52% coverage. The large differences in MAS surface areas, as shown by the standard deviation, already indicate their variability. Difference between coverage by all MAS and only the largest, shows that MAS location varies strongly as well. Therefore, design of generic plates that do not interfere with individual MAS is challenging. Hence, patient-specific clavicle fracture fixation plates should be considered to minimally interfere with MAS.status: publishe