Algorithm for the calculation of proximity area and area centroid within the carpal joint

Abstract

The purpose of this study was to use Computed Tomography (CT) imaging technology and computer-generated reconstruction techniques in developing an automatic method for the determination of carpal joint proximity area and its area centroid within specific distance thresholds. Five fresh frozen cadaver specimen arms were used in this study. Each arm was first loaded with a total of 801b from the carpal flexor and extensor tendons. Then the entire wrist was CT imaged at an interval of 1.5mm. The weights were removed and the wrist was imaged again at the same interval. It was expected that with increased load, the proximity "area" in the radio-carpal joint would also increase. AU CT slices were contoured, then reconstructed as a three-dimensional model, and each bone shape examined for gross flaws. Proximity distances were calculated for each discrete point of the bone surface. An algorithm was written in C which used the discrete spatial locations of the bone surfaces, each location's corresponding proximity distance and projected bone information to calculate proximity "area" and its centroid. The programs accuracy was tested creating input files from a know geometry and testing the output for different thresholds. Each wrist was analyzed at both Olb and 801b loads at thresholds ranging from 2mm to 4mm. The difference between the loaded and unloaded condition was used to normalize for different wrist sizes. It was expected that the proximity area would increase as load increased, however, in all cases, the joint proximity "area" increased or decreased without presenting a definite pattern. A geometric simulation using spheres determined possible error in the distance measurements of reconstructed objects. This simulation showed that the current protocol had too much error to accurately predict proximity distances and their associated areas.