Novel and simplified optimisation pathway using response surface and design of experiments methodologies for dental...

Dental implants are widely used as a long-term treatment solution for missing teeth. A titanium implant is inserted into the jawbone, acting as a replacement for the lost tooth root and can then support a denture, crown or bridge. This allows discreet and high-quality aesthetic and functional improvement, boosting patient confidence. The use of implants also restores normal functions such as speech and mastication. Once an implant is placed, the surrounding bone will fuse to the titanium in a process known as osseointegration. The success of osseointegration is dependent on stress distribution within the surrounding bone and thus implant geometry plays an important role in it. Optimisation analyses are used to identify the geometry which results in the most favourable stress distribution, but the traditional methodology is inefficient, requiring analysis of numerous models and parameter combinations to identify the optimal solution. A proposed improvement to the traditional methodology includes the use of Design of Experiments (DOE) together with Response Surface Methodology (RSM). This would allow for a well-reasoned combination of parameters to be proposed. This study aims to use DOE, RSM and finite element models to develop a simplified optimisation analysis method for dental implant design. Drawing on data and results from previous studies, two-dimensional finite element models of a single Branemark implant, a multi-unit abutment, two prosthetic screws, a prosthetic crown and a region of mandibular bone were built. A small number of combinations of implant diameter and length were set based on the DOE method to analyse the influence of geometry on stress distribution at the bone-implant interface. The results agreed with previous studies and indicated that implant length is the critical parameter in reducing stress on cortical bone. The proposed method represents a more efficient analysis of multiple geometrical combinations with reduced time and computational cost, using fewer than a third of the models required by the traditional methods. Further work should include the application of this methodology to optimisation analyses using three-dimensional finite element models.