Semi-automatic bone scaffold design workflow

Abstract

Scaffold-guided bone regeneration (SGBR) is a novel method to treat large bone defects which has successfully been applied in a few clinical cases [1]. A SGBR scaffold is designed to match the defect of the patient in shape and size (patient-specific), then 3D printed using a bioresorbable polymer and is finally surgically implanted. Practical challenges in the design stage of the scaffold have been identified which hamper the routine clinical translation of this therapy. The scaffold design should be patient-specific, insertable from the planned surgical approach without being obstructed by existing bone, securely attachable to the host bone to prevent dislodgement, possess controlled porosity and pore architecture, and finally, the designed model should not have any surface mesh errors and thereby, be readily 3D printable. A semi-automatic modular workflow was developed to create patient-specific SGBR scaffold designs for a given bone defect model and a pre-determined surgical approach to overcome the above challenges and create scaffolds based on a multitude of pore architectures with varying porosities. The workflow was implemented within Rhinoceros 3D and Grasshopper (R&G) software (Robert McNeel & Associates, Washington, USA). Through its graphical programming interface an algorithmic workflow was developed to efficiently manipulate scaffold geometries. A dedicated plugin for R&G was written which enables Functional-representation (F-rep) modelling techniques that enable fast and robust Boolean operations without creating spontaneous surface mesh errors that inhibit 3D printing. The workflow was validated by applying it to a complex multi-fragmentary femoral bone defect. The workflow was able to design scaffolds for a given surgical approach complete with fixation points as requested by surgeons with minimal user input, with near real-time responsiveness which grants real-time surgeon feedback. The designs were inspected for patient-specific fit and unobstructive insertion digitally as well as physically via 3D printed prototypes which was successful. The output models were found to have no mesh errors when checked with commercial slicing software. In conclusion, the developed workflow is successful in designing patient-specific scaffolds with real-time responsiveness to overcome the above-mentioned challenges.