Development and validation of statistical shape models of the primary functional bone segments of the foot

• DOI: 10.7717/peerj.8397
• Publication Date: 2020-02-04T04:34:48Z
• Authors: Tamara M. Grant; Laura E. Diamond; Claudio Pizzolato; Bryce A. Killen; Daniel Devaprakash; Luke Kelly; Jayishni N. Maharaj; David J. Saxby
• Keywords: Foot (prosody); Primary (astronomy)

Introduction Musculoskeletal models are important tools for studying movement patterns, tissue loading, and neuromechanics. Personalising bone anatomy within improves analysis accuracy. Few studies have focused on personalising foot anatomy, potentially incorrectly estimating the foot’s contribution to locomotion. Statistical shape been created a subset of foot-ankle bones, but not validated. This study aimed develop validate statistical functional segments in foot: first metatarsal, midfoot (second-to-fifth metatarsals, cuneiforms, cuboid, navicular), calcaneus, talus; then, assess reconstruction accuracy these using sparse anatomical data. Methods Magnetic resonance images 24 individuals feet (age = 28 ± 6 years, 52% female, height 1.73 0.8 m, mass 66.6 13.8 kg) were manually segmented generate three-dimensional point clouds. Point clouds registered analysed principal component analysis. For each segment, model components created, describing population variation. validated by assessing leave-one-out cross validation. excluding participant’s segment used reconstruct that same excluded full segmentations data (i.e. three discrete points segment), all combinations dataset. Tali reconstructed due lack externally accessible landmarks. Reconstruction was assessed Jaccard index, root mean square error (mm), Hausdorff distance (mm). Results Reconstructions generated had indices between 0.77 0.04 0.89 0.02, errors 0.88 0.19 1.17 0.18 mm, distances 2.99 0.98 mm 6.63 3.68 mm. 0.67 0.06 0.83 0.05, 1.21 0.54 1.66 0.41 3.21 0.94 7.19 3.54 index higher ( P < 0.01) lower reconstructions from compared calcaneus Conclusion time, primary developed Foot can be with minimal In future, larger training datasets could increase robustness, extending use paediatric or pathological populations.