Abstract Body:

Introduction: Additive manufacturing, or 3D printing, can create on-demand patient-specific orthopedic implants. A promising application is devices to replace the meniscus, a tissue that stabilizes the knee and is often injured in sports [1]. Clinically, musculoskeletal implant failures are difficult to diagnose. Utilizing magnetic resonance imaging (MRI) for detection is generally inconclusive due to high-intensity T1 and T2 signal changes from scar tissue, necessitating invasive surgical techniques [2, 3]. Computed tomography (CT) is an alternative, but orthopedic applications require differentiation of radiopaque implants from bone. This can be accomplished with the maturing technology of spectral photon-counting CT (SPCCT) which can be used to discriminate radiopaque contrast agents and bone based on unique attenuation properties of elements [4].

Methods: To fabricate on-demand, patient-specific implants, ex vivo swine legs were scanned with clinical CT and medial menisci were digitally extracted. Menisci were converted to STL models (PrusaSlicer 2.6.1) and 3D printed from polycaprolactone (PCL) incorporating 4wt% bismuth oxide (Bi₂O₃) nanoparticles (American Elements) using a Lulzbot Taz 6, and rectilinear infill of 40-50%. Implants were inserted into the original stifle joint of pig legs after surgical extraction of the native meniscus and secured in place with surgical glue. Devices were implanted both in and out of alignment, and imaged post-implantation using clinical CT and a clinical SPCCT prototype (Philips), using scan parameters similar to a clinical protocol: 120 kVp, 182 mA, 190 mAs. The SPCCT system enables 500 mm in-plane field-of-view, and z-coverage of 17.6 mm in the isocenter, realized with 64 rows of detectors (270×270 µm²). The application-specific integrated circuit supports the readout of 5 different energy bins, set at 30, 51, 63, 78, and 92 keV for scans of bismuth nanoparticles. Reconstruction parameters were set to provide both conventional HU images and color K-edge images of bismuth, with 256 mm field-of-view, matrix size of 1024 X 1024 pixels, a 250 μm isotropic voxel size, with detailed bone (sharp) reconstruction kernel (e.g., Detailed 2). More technical details of the reconstruction chain are provided in [5].

Results & Discussion: Introduction of Bi₂O₃ nanoparticles into PCL effectively produced implants which were easily distinguished from muscle (contrast to noise ratio: 16). Utilizing additive manufacturing, a meniscal device could be created at a human scale. The physiological similarities between the ex vivo swine model and humans ensure that the imaging results in this study are readily translatable to clinical applications.

Patient specific, 3D printed menisci fit into the meniscal space with excellent alignment. When an intentional misalignment was introduced, it was easily distinguishable by clinical CT. After validation of CT monitoring to distinguish implant alignment, the ex vivo model was imaged using clinical SPCCT. Unlike traditional CT, SPCCT measures the energy of transmitted X-rays and “bins” them to perform a re-convolution of the elemental distribution based on elemental K-edge values. Bismuth has a distinct K-edge at 90.5 keV, allowing differentiation from the calcium phosphates present in bone [6]. Thus, the composite meniscal implant is readily identifiable in the ultra-high-resolution scan. The voxels above the Bi₂O₃ threshold content were isolated to generate the reconstruction of the anatomically derived, 3D printed radiopaque meniscus. The fusion image demonstrates excellent radiographic distinguishability, even in applications where implants possess similar X-ray attenuation to bone in a conventional CT.

Conclusion: Here we demonstrate how non-invasive CT of radiopaque implants can be used for enhanced diagnosis of orthopedic implant failure. Additionally, with the growth of clinical SPCCT, the incorporation of metal oxide-based nanoparticles is poised to enhance diagnostic power and longitudinal traceability to implanted biomedical devices.

Author