Computer-assisted hip resurfacing planning using Lie group shape models.

INTRODUCTION: Hip resurfacing is a surgical option for osteoarthritis young and active patients. Early failures has been reported due to improper implant placement. Computer-assisted surgery is a promising avenue for more successful procedures. PURPOSE: This paper presents a novel automatic surgical planning for computer-assisted hip resurfacing procedures. The plan defined the femoral head axis that was used to place the implant. The automatic planning was based on a Lie group statistical shape model. METHODS: A statistical shape model was constructed using 50 femurs from osteoarthritis patients who underwent computer-assisted hip resurfacing. The model was constructed using product Lie groups representation of shapes and nonlinear analysis on the manifold of shapes. A surgical plan was drawn for the derived base shape. The base shape was transformed to 14 femurs with known manual plans. The transformed base plan was used as the computed plan for each femur. Both actual and computed plans were compared. RESULTS: The method showed a success by computing plans that differ from the actual plans within the surgical admissible ranges. The minimum crossing distance between the two plans had a mean of 0.75 mm with a standard deviation of 0.54 mm. The angular difference between the two plans had the mean of 5.94° with a standard deviation of 2.145.94°. CONCLUSION: Product Lie groups shape models were proved to be successful in automatic planning for hip resurfacing computer-assisted surgeries. The method can be extended to other orthopedic and general surgeries.

A matrix lie group approach to statistical shape analysis of bones.

Statistical shape models using a principal-component analysis are inadequate for studying shapes that are in non-linear manifolds. Principal tangent components use a matrix Lie group that maps a non-linear manifold to a corresponding linear tangent space. Computations that are performed on the tangent space of the manifold use linear statistics to analyze non-linear shape spaces. The method was tested on bone surface from proximal femurs. Using only three components, the new model recovered 94% of the medical dataset, whereas a conventional method that used linear principal components needed 24 components to achieve the same reconstruction accuracy.

The aspherical human hip: implication for early osteoarthritis.

One marker for early-onset hip arthritis is femoral acetabular impingement. The current standard way of quantifying impingement is manual calculation of anatomical measures on plain radiographs, including the α-angle. Such measurements are user-dependent and prone to error. We provided a robust computational alternative and proposed using numerical fitting of geometrical shapes. We applied least-squares fitting of an ellipse to the femoral head contour and used the difference between the ellipse axes as a quantification method. The results showed a good correlation between the new measure and previous definitions of the α-angle.

A prospective, randomized assessment of a spatial orientation device in natural orifice transluminal endoscopic surgery.

BACKGROUND: One of the challenges in natural orifice transluminal endoscopic surgery (NOTES) is spatial orientation. The Queen's NOTES group has devised a novel method of orientation by using a magnetic device that passes within an endoscope channel allowing for 3-dimensional imaging of the shape and orientation of the endoscope. OBJECTIVE: To assess the feasibility and utility of a novel orientation device. DESIGN: Randomized, controlled trial. SETTING: Animal research laboratory study on four 25-kg pigs. INTERVENTION: The device was tested by 6 endoscopists and 6 laparoscopic surgeons. Starting at the gastrotomy, the time to identify 4 targets was recorded. Participants were required to identify and touch the gallbladder, the fallopian tube, a clip on the abdominal wall, and the liver edge. Use of the orientation device was randomized for each session. MAIN OUTCOME MEASUREMENTS: Time to identify targets with and without the device. Secondary analysis assessed differences between medical specialties and level of training. RESULTS: The mean time to identify all 4 targets with the device was 75.08 ± 42.68 seconds versus 100.20 ± 60.70 seconds without the device (P <.001). The mean time to identify all 4 targets on the first attempt was 102.29 ± 61.36 seconds versus 72.99 ± 40.19 seconds on the second attempt (P <.001). No differences based on specialty or level of training were identified. LIMITATIONS: Small sample size and simplicity of tasks. CONCLUSION: Regardless of randomization order, both groups were faster with the device. These encouraging results warrant further study using more complex scenarios.