MRI-Based Characterization of Bone Anatomy in the Human Knee for Size Matching of a Medial Meniscal Implant

Abstract

Allograft or synthetic menisci have been suggested as a means to restore contact pressures following meniscectomy. However, when the natural meniscus is severely damaged/absent, the necessary size cannot be determined according to the recipient size and there is a need to estimate it from magnetic resonance imaging (MRI) of the contralateral knee or the injured knee bones. The use of the contralateral-knee for size matching is problematic due to economic and practical reasons. Hence, there are significant advantages for a sizing algorithm based only on the candidate knee geometry. The aim of this study is to characterize midrange values and variability of knee dimensions and to develop a set of mathematical relations representing knee dimensions using a minimum of imaging-based bone measurements. Tibia, femur, and meniscus measurements were taken in 118 MRI scans and used to develop a representative parametric knee model in which all dimensions are expressed using tibia plateau width. The model was verified by comparing the predicted values to direct MRI measurements for 20 additional subjects by means of the Pearson correlation and Bland and Altman (1986, "Statistical Methods for Assessing Agreement Between Two Methods of Clinical Measurement," Lancet, 1, pp. 307-310) plot. Anatomical parameters in the male knee were significantly larger (∼17%) compared with corresponding female measurements. However, most relations between tibia, femur, and meniscus measurements (43/56) were not significantly different between male and female populations (p ≥ 0.05), indicating that differences between male and female joints are generally related to scaling and not shape. Dimensions predicted by the knee model were in a good agreement with dimensions measured directly from the MRI (R(2)>0.96) and the Bland and Altman plot indicated that ∼95% of data points were well within the ± 2 standard deviation lines of agreement. The model proposed in this study is advantageous in being able to describe typical knee proportions for a given tibial width and can be used to predict the dimensions of a candidate knee based on a single measurement. The anatomical/anthropometric data presented in the study can be utilized in a sizing algorithm for artificial meniscal implants or in the design of artificial meniscus prostheses.