Use of roentgenography and magnetic resonance imaging to predict meniscal geometry determined with a three-dimensional coordinate digitizing system.

To evaluate and improve on the procedures used by a tissue bank in selecting donor menisci for transplantation, this study was designed to fulfill four objectives: (a) define and quantify a set of independent parameters for describing the geometry of the medial and lateral menisci, (b) determine how well the sizing protocol of the tissue bank (i.e., two transverse roentgenographic measurements obtained from the injured knee or six transverse magnetic resonance imaging measurements obtained from the contralateral knee) predicts the four standard transverse parameters of the menisci, (c) determine if including one additional transverse roentgenographic measurement for each compartment improves the ability of roentgenograms to predict transverse meniscal parameters, and (d) determine if five magnetic resonance imaging measurements at three different meniscal cross sections of the contralateral knee predict the 15 standard cross-sectional parameters of the meniscus in the injured knee. A laser-based, noncontacting three-dimensional coordinate digitizing system was used to determine surface coordinates from which menisci were reconstructed in a computer. For each reconstructed meniscus, four parameters in the transverse plane and five cross-sectional parameters in each of three regions (i.e., anterior, middle, and posterior) were defined, yielding a set of 19 standard parameters to describe the geometry. Through a correlation analysis, these standard parameters were shown to be largely unrelated to one another, thus confirming that the parameters form an independent set describing the three-dimensional geometry of the menisci. The two roentgenographic measurements were poor predictors of transverse standard meniscal parameters, predicting only one of four standard parameters for the medial meniscus and none of four standard parameters for the lateral meniscus with coefficients of determination greater than or equal to 0.5. Including one additional roentgenographic measurement to the tissue bank protocol increased the number of standard transverse parameters predicted to three of four for the medial meniscus and two of four for the lateral meniscus. Magnetic resonance imaging was better than roentgenography for predicting the three-dimensional meniscal geometry. The transverse measurements from magnetic resonance imaging predicted three of four standard transverse parameters for the medial meniscus and all four for the lateral meniscus. With the addition of the cross-sectional measurements by magnetic resonance imaging, seven of 15 standard cross-sectional parameters were predicted for both the medial and lateral menisci. Assuming that a successful clinical outcome depends on how well an allograft matches the size and shape of the original meniscus, magnetic resonance imaging rather than roentgenography should be used for allograft size-matching by tissue banks.