BACKGROUND: We describe a novel computational method for assessing the fit of an osteochondral graft. We applied our software to five normal wrist computed tomography (CT) scans to determine the fit of the scaphoid to the lunate fossa of the distal radius. METHODS: CT scans of five wrists were digitally rendered. The capitate facet of the scaphoid was fit to the lunate fossa of the distal radius using custom software based on the iterative closest point (ICP) algorithm. This approach iteratively determines the optimal position of a model surface to minimize the sum of squares of distances from all points on a target surface. The fit of the two surfaces was reported by calculating the mean residual distance (MRD) between each point on one surface and its nearest neighbor on the other. RESULTS: The MRD for the five subjects was found to be 0.25 mm, with 82.8-98.3 % of the articular surfaces within 0.5 mm of each other. CONCLUSIONS: We have developed a software algorithm for comparing two articular surfaces to test fit for a proposed joint reconstruction. The software is versatile and may be applied to any bony surface to identify new graft donor sites. The fit assessment renders a richer, three-dimensional understanding of the fit of the graft as compared to traditional two-dimensional assessments. LEVEL OF EVIDENCE: Decision analysis, Level V.
BACKGROUND: We describe a novel computational method for assessing the fit of an osteochondral graft. We applied our software to five normal wrist computed tomography (CT) scans to determine the fit of the scaphoid to the lunate fossa of the distal radius. METHODS: CT scans of five wrists were digitally rendered. The capitate facet of the scaphoid was fit to the lunate fossa of the distal radius using custom software based on the iterative closest point (ICP) algorithm. This approach iteratively determines the optimal position of a model surface to minimize the sum of squares of distances from all points on a target surface. The fit of the two surfaces was reported by calculating the mean residual distance (MRD) between each point on one surface and its nearest neighbor on the other. RESULTS: The MRD for the five subjects was found to be 0.25 mm, with 82.8-98.3 % of the articular surfaces within 0.5 mm of each other. CONCLUSIONS: We have developed a software algorithm for comparing two articular surfaces to test fit for a proposed joint reconstruction. The software is versatile and may be applied to any bony surface to identify new graft donor sites. The fit assessment renders a richer, three-dimensional understanding of the fit of the graft as compared to traditional two-dimensional assessments. LEVEL OF EVIDENCE: Decision analysis, Level V.
Authors: John T Capo; Hill Hastings; Edward Choung; Tosca Kinchelow; William Rossy; Bruce Steinberg Journal: J Hand Surg Am Date: 2008 May-Jun Impact factor: 2.230