[PDF] - Outline Total Knee Arthroplasty Project Motivation and Rationale PDF Document

SLIDE 1

1

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

SLIDE 2

2

Total Knee Arthroplasty (TKA)

 Treatment for severe knee osteoarthritis  Goals: Relieve pain, restore joint alignment  Over 500,000 TKA procedures per yr in North America1  Over 80% of patients satisfied with overall results and

improved quality of life (15 year follow-up study)2

3

Tibia Femur Patella Femoral Component Tibial Component

1. Kozak 2006. 2. Loughead 2008.

Project Motivation

4

Knee Pain

10-25% report postoperative

anterior knee pain (AKP)1

Patellar tracking is a likely cause
f AKP but not proven
Does patellar tracking differ for

patients with & w/o AKP?

Gender Differences

Tracking differences exist2
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]

1. Helmy N, 2008. 2. Anglin C, Ho KCT et al. 2008.

SLIDE 3

3

Project Rationale

5

Postoperative AKP Patellar Tracking Gender (ML/AP Ratio)

? Post-TKA Mechanics

6

X-ray & Fluoroscopy1

 2D  poor accuracy  single flexion angle (X-ray)  sagittal view (fluoro)

Computed Tomography (CT)

 3D  Better cortical bone

definition

 Higher flexion angles  Faster acquisition time  Less noise discomfort

Limitations

 Metal Artifact  Radiation

Magnetic Resonance Imaging (MRI)2,3

 3D  limited accuracy (low res.)  accuracy not validated for

patellar kinematics

 no contact area  metal artifact

1. Stiehl JB, 2001. 2. von Eisenhart-Rothe R, 2007. 3. Lee KY, 2005.

SLIDE 4

4

Objectives

7

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-

bserver repeatability of the CT protocol

Implementation Evaluate the effects of gender-specific implant design on knee mechanics

CT Protocol Development

8

Image Acquisition Segmentation Registration Joint Mechanics & Contact Areas

CT scans at multiple flexion angles Partial weight-bearing knee rig1 Metal artifact reduction Components matched to 3D TKA models Segment 3 bones + 3 components Each flexion angle registered to full extension position 6-DOF patellofemoral (PF) mechanics 6-DOF tibiofemoral (TF) mechanics PF & TF contact areas Patellar tracking relative to femoral groove

1. Connelly K, MSc Thesis, University of Calgary, 2006

SLIDE 5

5

Defining Patellofemoral Mechanics

Patellar Maltracking

 defined as shift > 5 mm or tilt > 5˚  could lead to clinical differences1

9

1. Shih H, 2004. 2. Katchburian M, 2003.

Clinical Standard: 30˚ or 45˚ radiographs Biomechanics Reference: Mechanical Axis vs. Prosthetic Femoral groove2

CT Protocol Validation

10

Accuracy Repeatability Expected Outcomes

At least 3 cadaveric knee specimens Compare CT method against Optotrak system using bone-mounted markers Intra-observer repeatability perform protocol at least 3 times Inter-observer repeatability 2 different observers perform protocol Accuracy of at least an order of magnitude smaller than clinically relevant differences in mechanics: 5 mm or 5º (translations and rotations)1

1. Shih H, 2004.

SLIDE 6

6

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

12

Goal Achieve adequate segmentation such that key distinctive features of each knee component are accurate for 3D TKA model registration

Patellar Button Pins Tibial Plateau Flange / Holes Femoral Component Pins

SLIDE 7

7

Segmentation Stages (1)

13

Original Image

Resampled data 0.8 mm slice resolution

Rough Threshold

Sawbone [-650 to -300] Metal [+2000 to +3071]

Segmentation Stages (2)

14

Mask Artifact Tibial Component

Threshold [+2000 to +3071]

SLIDE 8

8

Segmentation Stages (3)

15

Femoral Component

Morphological Filter: Closing Operation Island Removal

Dilation Erosion

Segmentation Stages (4)

16

Patellar Component

Mask patellar region Threshold to identify pins [-200 to 3071]

SLIDE 9

9

Segmentation Results

17

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

CAD Model Registration

18

SLIDE 10

10

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)

20

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

SLIDE 11

11

Patient Positioning

Artifact band appears in region of greatest material thickness

21

Full Extension 30˚ Flexion 60˚ Flexion Foot on CT Table Foot Raised

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

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

SLIDE 12

12

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