Method

A subject implanted with a new Percutaneous Osseointegrated Prosthesis (POP) (DJO Surgical, Austin, TX) had CTs taken of their residual femur at 6-weeks and 12-months post-op in a FDA Early Feasibility Study with Institutional Review Board approval. Three-dimensional models of the femur were created from dicom files of the CT slices using Mimics (v21.0, Materialise, Leuven, Belgium). Each scan was segmented into four objects: cortical bone, medullary cavity, total volume (cortical bone plus the medullary cavity) and endoprosthetic stem (Fig. 1).

Following segmentation, models were uploaded to 3-Matic Research (v13.0, Materialise, Leuven, Blegium) in STL format for alignment to a common world coordinate system (Fig. 2). A common origin was set by taking the average distance between planes of the femoral head and the greater trochanter. Once aligned to the coordinate system, biomechanical length (BML) was calculated from the proximal origin to the distal end of the amputated femur.

BML and STLs of the aligned medullary cavity and femur volume were entered into custom Matlab code designed to measure cortical and medullary morphology in transverse cross sections of the femur. Morphology data from 6-weeks and 12-month time points were compared in order to determine if bone remodeling around the POP implant could be detected using these methods.

Methods

Twenty-five goats underwent anterior cervical discectomy and fusion using a Trabecular Metal or PEEK device for 6, 12 or 26 weeks. The GH of each device was filled with autograft bone morsels harvested from the animal at implantation. Fluorochrome labeling oxytetracycline was administered to the animals and used to determine bone viability in the device regions. Following necropsy, the vertebral segments were embedded in poly(methyl methacrylate) sectioned and analyzed using fluorescence and backscatter electron (BSE) imaging. The percent of bone tissue present within the GH was measured as a volume percent using BSE images (Fig. 1).