OBJECTIVE: Our goal was the development of an accurate and objective technique to depict and quantitate articular cartilage using magnetic resonance imaging (MRI) and 3D data processing. METHODS: A method of 3D image analysis has been developed that provides a noninvasive technique of measuring cartilage volume from standard diagnostic MR images. RESULTS: Using a computer workstation, cartilage from Fast Field Echo (FFE) MR images was reconstructed into 3 dimensions (3D). The accuracy and reproducibility of both thickness and volume measurements obtained from the existing computer software was tested using calibration test objects. We describe the sources of error we encountered when attempting to quantitate cartilage using the existing computer software, the methods developed to reduce these errors, and a preliminary cartilage volume study using healthy human volunteers. CONCLUSION: We have identified errors involved in attempting cartilage volume estimations using the existing 3D computer software and have developed a data processing technique that minimizes these errors. With these objective data processing techniques we have improved the reproducibility of the technique to +/- 10-15% error. This modified technique provides a promising new method for viewing and quantifying cartilage volume from standard diagnostic MR images.
OBJECTIVE: Our goal was the development of an accurate and objective technique to depict and quantitate articular cartilage using magnetic resonance imaging (MRI) and 3D data processing. METHODS: A method of 3D image analysis has been developed that provides a noninvasive technique of measuring cartilage volume from standard diagnostic MR images. RESULTS: Using a computer workstation, cartilage from Fast Field Echo (FFE) MR images was reconstructed into 3 dimensions (3D). The accuracy and reproducibility of both thickness and volume measurements obtained from the existing computer software was tested using calibration test objects. We describe the sources of error we encountered when attempting to quantitate cartilage using the existing computer software, the methods developed to reduce these errors, and a preliminary cartilage volume study using healthy human volunteers. CONCLUSION: We have identified errors involved in attempting cartilage volume estimations using the existing 3D computer software and have developed a data processing technique that minimizes these errors. With these objective data processing techniques we have improved the reproducibility of the technique to +/- 10-15% error. This modified technique provides a promising new method for viewing and quantifying cartilage volume from standard diagnostic MR images.