PURPOSE
To evaluate short echo time (TE) projection reconstruction magnetic resonance (MR) imaging in the detection of cartilage lesions.
MATERIALS AND METHODS
Twenty-seven cartilage regions of 10 human patellar specimens were examined with the following MR sequences: short TE projection reconstruction (repetition time msec/TE msec, 400/0.15), fat-suppressed three-dimensional spoiled gradient-recalled acquisition in the steady state (Spoiled GRASS) (50/10, 60 degrees flip angle), and magnetization transfer contrast (MTC) subtraction (400/6). MR findings were correlated with histopathologic grading of the cartilage.
RESULTS
For detection of cartilage lesions, sensitivity of projection reconstruction imaging (100%) was significantly greater (P = .03) than that of MTC (62%) but not significantly greater (P > .05) than that of Spoiled GRASS (81%) imaging. Accuracy of projection reconstruction was significantly greater than that of MTC (P = .004) and Spoiled GRASS (P = .03) imaging. Unmasking of collagen fibers was most predictive of abnormal signal intensity of the cartilage with all sequences.
CONCLUSION
In vitro, short TE projection reconstruction MR imaging provides superior delineation of cartilage lesions when compared with two other sequences. On Spoiled GRASS and MTC images, signal intensity of the superficial layer of cartilage is not a reliable sign for surface integrity.