OpenLimbTT, a Transtibial Residual Limb Shape Model for Prosthetics Simulation and Design: creating a statistical anatomic model using sparse data

Poor socket fit is the leading cause of prosthetic limb discomfort. However, currently clinicians have limited objective data to support and improve socket design. Prosthesis fit could be predicted by finite element analysis to help improve the fit, but this requires internal and external anatomy models. While external 3D surface scans are often collected in routine clinical computer aided design practice, detailed imaging of internal anatomy (e.g. MRI or CT) is not. This paper presents a prototype Statistical Shape Model (SSM) describing the transtibial amputated residual limb, generated using a sparse dataset of 10 MRI scans. To describe the maximal shape variance, training scans are size-normalised to their estimated intact tibia length. A mean limb is calculated, and Principal Component Analysis used to extract the principal modes of shape variation. In an illustrative use case, the model is interrogated to predict internal bone shapes given a skin surface shape. The model attributes [~]82% of shape variance to amputation height and [~]7.5% to soft tissue profile. Leave-One-Out cross-validation allows mean shape reconstruction with 0.5-3.1mm root-mean-squared-error (RMSE) surface deviation (median 1.0mm), and left-out-shape reconstruction with 4.8-8.9mm RMSE (median 6.1mm). Linear regression between mode scores from skin- only- and full-model SSMs allowed prediction of bone shapes from the skin surface with 4.9-12.6mm RMSE (median 6.5mm). The model showed the feasibility of predicting bone shapes from skin surface scans, which will enable more representative prosthetic biomechanics research, and address a major barrier to implementing simulation within clinical practice. Impact StatementThe presented Statistical Shape Model answers calls from the prosthetics community for residual limb shape descriptions to support prosthesis structural testing that is representative of a broader population. The SSM allows definition of worst-case residual limb sizes and shapes, towards testing standards. Further, the lack of internal anatomic imaging is one of the main barriers to implementing predictive simulations for prosthetic socket interface fitting at the point-of-care. Reinforced with additional data, this model may enable generation of estimated finite element analysis models for predictive prosthesis fitting, using 3D surface scan data already collected in routine clinical care. This would enable prosthetists to assess their design choices and predict a sockets fit before fabrication, important improvements to a time-consuming process which comes at high cost to healthcare providers. Finally, few researchers have access to residual limb anatomy imaging data, and there is a cost, inconvenience, and risk associated with putting the small community of eligible participants through CT or MRI scanning. The presented method allows sharing of representative synthetic residual limb shape data whilst protecting the data contributors privacy, adhering to GDPR. This resource has been made available at https://github.com/abel-research/openlimb, open access, providing researchers with limb shape data for biomechanical analysis.