Outline Total Knee Arthroplasty Project Motivation and Rationale - PDF document

Computed Tomography Methods for Assessing Total Knee Replacement Mechanics Karen C.T. Ho University of Calgary, Canada Centre for Bioengineering Research & Education Imaging and Measurements in Biomedical Engineering Paris, France October 2-3, 2008 Outline Total Knee Arthroplasty Project Motivation and Rationale Postoperative Knee Mechanics Protocol Development and Validation Preliminary Results Future Work & Significance 2 1

Total Knee Arthroplasty (TKA)  Treatment for severe knee osteoarthritis  Goals: Relieve pain, restore joint alignment  Over 500,000 TKA procedures per yr in North America 1  Over 80% of patients satisfied with overall results and improved quality of life (15 year follow-up study) 2 Femur Patella Femoral Component Tibial Tibia Component 3 1. Kozak 2006. 2. Loughead 2008. Project Motivation Knee Pain • 10-25% report postoperative anterior knee pain (AKP) 1 • Patellar tracking is a likely cause of AKP but not proven • Does patellar tracking differ for patients with & w/o AKP? Gender Differences • Tracking differences exist 2 • Gender-specific (GS) implants are designed to suit female femurs (ML/AP ratio) • Does GS implant design affect patellar tracking? Pain? Zimmer Gender Solutions Knee [www.genderknee.com] 4 1. Helmy N, 2008. 2. Anglin C, Ho KCT et al. 2008. 2

Project Rationale Postoperative Gender AKP (ML/AP Ratio) ? Patellar Tracking 5 Post-TKA Mechanics X-ray & Fluoroscopy 1 Computed Tomography (CT)  2D  3D  poor accuracy  Better cortical bone definition  single flexion angle (X-ray)  Higher flexion angles  sagittal view (fluoro)  Faster acquisition time Magnetic Resonance  Less noise discomfort Imaging (MRI) 2,3  3D Limitations  limited accuracy (low res.)  Metal Artifact  accuracy not validated for  Radiation patellar kinematics  no contact area  metal artifact 6 1. Stiehl JB, 2001. 2. von Eisenhart-Rothe R, 2007. 3. Lee KY, 2005. 3

Objectives Development Develop an accurate protocol to measure in vivo 3D knee arthroplasty mechanics and joint contact areas using computed tomography Validation Validate the accuracy and the intra- and inter- observer repeatability of the CT protocol Implementation Evaluate the effects of gender-specific implant design on knee mechanics 7 CT Protocol Development CT scans at multiple flexion angles Image Acquisition Partial weight-bearing knee rig 1 Metal artifact reduction Segmentation Components matched to 3D TKA models Segment 3 bones + 3 components Each flexion angle registered to full Registration extension position 6-DOF patellofemoral (PF) mechanics Joint Mechanics 6-DOF tibiofemoral (TF) mechanics & Contact Areas PF & TF contact areas Patellar tracking relative to femoral groove 8 1. Connelly K, MSc Thesis, University of Calgary, 2006 4

Defining Patellofemoral Mechanics Patellar Maltracking  defined as shift > 5 mm or tilt > 5˚  could lead to clinical differences 1 Clinical Standard: Biomechanics Reference: 30˚ or 45˚ radiographs Mechanical Axis vs. Prosthetic Femoral groove 2 9 1. Shih H, 2004. 2. Katchburian M, 2003. CT Protocol Validation At least 3 cadaveric knee specimens Accuracy Compare CT method against Optotrak system using bone-mounted markers Intra-observer repeatability  perform protocol at least 3 times Repeatability Inter-observer repeatability  2 different observers perform protocol Accuracy of at least an order of Expected magnitude smaller than clinically Outcomes relevant differences in mechanics: 5 mm or 5º (translations and rotations) 1 10 1. Shih H, 2004. 5

Preliminary CT Scan TKA components on Sawbone CT scanner: Siemens Sensation 64 Image analysis software: Amira 4.1.2 Scan Parameters:  120 kVp, 108 mAs  FOV: 150mm  Matrix: 512 × 512  Slice Thickness: 2 mm  Number of slices: 38 11 Segmentation Goal Achieve adequate segmentation such that key distinctive features of each knee component are accurate for 3D TKA model registration Patellar Tibial Plateau Femoral Button Pins Flange / Holes Component Pins 12 6

Segmentation Stages (1) Original Image Rough Threshold Sawbone [-650 to -300] Resampled data Metal [+2000 to +3071] 0.8 mm slice resolution 13 Segmentation Stages (2) Mask Artifact Tibial Component Threshold [+2000 to +3071] 14 7

Segmentation Stages (3) Femoral Component Morphological Filter: Closing Operation Island Removal Dilation Erosion 15 Segmentation Stages (4) Patellar Component Mask patellar region Threshold to identify pins [-200 to 3071] 16 8

Segmentation Results Surface generation using unconstrained 3D smoothing • After segmenting implants and masking the artifact, the remainder of the image is bone  segment bones • If preop CT available, register preop to postop • Assign coordinate systems to bones and components 17 CAD Model Registration 18 9

Metal Artifact  Photon starvation (not enough photons hitting the detectors) producing incomplete data  Beam hardening (increase in effective energy as lower energy photons are attenuated)  Streak artifact is a combined effect 19 Metal Artifact Reduction (MAR) Before scanning Material selection (no control for our study) During scanning Patient positioning Standard MAR algorithm available on scanner Scanning parameters (e.g. adjust kVp, mAs) After scanning Use iterative forward & backward projections to correct and reduce metal artifact given known implant geometry (future research) Register to preop CT, when available 20 10

Patient Positioning Artifact band appears in region of greatest material thickness Raising the patient’s leg may move the patella out of the artifact and decrease the width of the artifact band Full Extension 30˚ Flexion 60˚ Flexion Foot on CT Table Foot Raised Full Extension 30˚ Flexion 60˚ Flexion 21 Radiation  Radiation dose limits number of flexion angles & number of patients  Ethically reasonable because in extremity and in older (non child-bearing) individuals  Will optimize acquisition parameters to reduce radiation dose  Aiming to combine with fluoroscopic imaging so only need one CT image (future work) 22 11

Conclusion Anticipated Outcome  Accurate, validated in vivo 3D CT method able to detect clinically relevant differences in patellofemoral & tibiofemoral mechanics  Patellar mechanics reported relative to femoral groove, to improve clinical relevance Potential Applications  Gold standard for testing other techniques  Testing clinical hypotheses (gender-specific vs. standard components, pain vs. no pain)  Use results to improve implant design, surgical technique, preop screening, postop diagnosis, prehab & rehab 23 Acknowledgements Research Team Dr. Carolyn Anglin – Supervisor Dr. Carol Hutchison – Orthopaedic Surgeon Jack Fu – MSc Student Jeff Wai – Research Assistant Funding CAOS-International Travel Fellowship NSERC Alberta Ingenuity Fund 24 12