Enhanced pre-clinical assessment of total knee replacement using computational modelling with experimental corroboration & probabilistic applications

Abstract

Demand for Total Knee Replacement (TKR) surgery is high and rising; not just in numbers of procedures, but in the diversity of patient demographics and increase of expectations. Accordingly, greater efforts are being invested into the pre-clinical analysis of TKR designs, to improve their performance in-vivo. A wide range of experimental and computational methods are used to analyse TKR performance pre-clinically. However, direct validation of these methods and models is invariably limited by the restrictions and challenges of clinical assessment, and confounded by the high variability of results seen in-vivo.
Consequently, the need exists to achieve greater synergy between different pre-clinical analysis methods. By demonstrating robust corroboration between in-silico and in-vitro testing, and both identifying & quantifying the key sources of uncertainty, greater confidence can be placed in these assessment tools. This thesis charts the development of a new generation of fast computational models for TKR test platforms, with closer collaboration with in-vitro test experts (and consequently more rigorous corroboration with experimental methods) than previously.
Beginning with basic tibiofemoral simulations, the complexity of the models was progressively increased, to include in-silico wear prediction, patellofemoral & full lower limb models, rig controller-emulation, and accurate system dynamics. At each stage, the models were compared extensively with data from the literature and experimental tests results generated specifically for corroboration purposes.
It is demonstrated that when used in conjunction with, and complementary to, the corresponding experimental work, these higher-integrity in-silico platforms can greatly enrich the range and quality of pre-clinical data available for decision-making in the design process, as well as understanding of the experimental platform dynamics. Further, these models are employed within a probabilistic framework to provide a statistically-quantified assessment of the input factors most influential to variability in the mechanical outcomes of TKR testing. This gives designers a much richer holistic visibility of the true system behaviour than extant 'deterministic' simulation approaches (both computational and experimental).
By demonstrating the value of better corroboration and the benefit of stochastic approaches, the methods used here lay the groundwork for future advances in pre-clinical assessment of TKR. These fast, inexpensive models can complement existing approaches, and augment the information available for making better design decisions prior to clinical trials, accelerating the design process, and ultimately leading to improved TKR delivery in-vivo to meet future demands.

More information

Identifiers

ORCID for M. Browne: orcid.org/0000-0001-5184-050X

Catalogue record

Export record