Although major strides have been made in understanding the kinematic and kinetic properties of the human musculoskeletal system, our knowledge is still quite limited due to a lack of reliable, accurate, non-invasive, in vivo, musculoskeletal measurement techniques. To address this issue, we first tested the accuracy and feasibility of using Cine Phase Contrast (Cine-PC) Magnetic Resonance Imaging (MRI) [1] to non-invasively measure 3-dimensional, in vivo, skeletal velocity.

Cine-PC MRI was originally developed to directly and non-invasively measure in vivo blood and heart velocity by providing one anatomic image and three velocity (v_x, v_y and v_z) images for each time frame. Although Cine-PC MRI has been shown capable of accurately measuring skeletal muscle fiber velocity in vivo during dynamic tasks [2], the accuracy of tracking trabecular bone has yet to be quantified and is potentially lower because the honeycomb nature of the trabeculation results in rapid spatial changes in magnetic properties [3], which could result in velocity errors.

The accuracy of tracking trabecular bone with Cine-PC MRI was determined to be high, with in-plane displacement errors of ~0.5 mm. Since no standard of reference exists for in vivo measurement of trabecular bone motion, a motion phantom (consisting of a series of paired gears that moved a sample box containing a human femoral bone sample) was built to assess the accuracy of tracking trabecular bone with Cine-PC MRI.

Cine-PC MRI was then applied to the study of normal (healthy) patellar-femoral-tibial (PFT) joint kinematics. Three-dimensional velocity profiles of the patella, femur and tibia were measured in 15 healthy subjects (28 knees) during leg extension resisted by a 34N weight (a low load in comparison to maximum isometric extensor torque). With the aid of a metronome subjects were able to maintain a consistent motion rate (35±0.5 cycles/min) and data quality was high. The orientation of the patella with respect to the femur ("patellar tracking") was described in terms of three orientation angles, θ1, θ2, and θ3 (based on an xyz body-fixed rotation system), which are commonly referred to as patellar flexion, tilt, and rotation, respectively. Based on the accuracy of tracking the phantom we predicted the error in the orientation angles to be ~3.5% on average. The difference in orientation angles measured in two different exams of 5 knees was 2.6 degrees. This demonstrates that Cine-PC MRI data can be produced repeatably and that 3D bone displacement and rotation can be accurately characterized.

Based on the results from 18 of the original 28 knees, patellar flexion (θ1) lags knee flexion, the patella tilts (02) laterally, and patellar rotation (θ3) changes little as the leg extends. Unique to this study, three distinct trends were found for patellar tilt (θ2):​ 7 knees tilted laterally, 7 knees tilted slightly medial, and 4 knees tilted both laterally and medially as the leg extended. The patellar kinematic data were then translated into common clinical parameters. Although a typical clinical study involves measurements at a single knee angle under static conditions, rather than over a range of knee angles under dynamic conditions, our results were similiar to previous clinical results. Thus, for the first time, 3D patellar kinematics were non-invasively measured in vivo during a dynamic task. The average maximum patellar tendon strain was 7.1% during this low load extension task.

It is concluded that Cine-PC MRI is a promising technique for the non-invasive measurement of in vivo skeletal dynamics and, based on our previous work, muscular dynamics as well. Also, this work demonstrates that patellar kinematics are much more complex than previously believed and patellar tendon strain to be quite high even during a low-load task.