Articular cartilage is a highly ordered tissue with layered structure. It can be functionally and structurally divided into the superficial, transitional and radial bones, as well as the calcified cartilage (CC) ¹. Variations with intrinsic magnetic resonance (MR) properties of cartilage can be exploited to provide non-invasive assessment of normal and abnormal tissue components. Recently, ultrashort echo time (UTE) imaging sequences have been used to investigate the T2* relaxation times of bound water and free water components in different layers of articular cartilage ². Relatively constant short T2* values and reduced long T2* values and short T2* fractions are observed with increasing cartilage depth. However, the depth dependences of longitudinal relaxation times, as well as the T1s of bound and free water components are still unknown. In this study we aimed to determine the T1s of bound and free water in different layers of patella cartilage using UTE imaging sequences on a clinical 3T scanner.Cadaveric patellar (n=7) MR imaging was performed on a GE 3T Signa TwinSpeed MR scanner (GE Healthcare Technologies, Milwaukee, MI) with a maximum peak gradient amplitude of 40 mT/m, and a slew rate of 150 mT/m/s using a single-channel 3-inch surface coil. A UTE saturation recovery technique was used to measure the T1 of the patella samples. A two-component fitting model was used to account for the short T1 and long T1 components. A total number of 20 single slice UTE images were acquired with a series of saturation recovery time (TSR) times, including 10, 50, 100, 150, 200, 250, 300, 400, 500, 600, 750, 1000, 1500, 2000, 2500, 3000, 4000, 5000, 6000 and 8000 ms. Other imaging parameters included a TE of 8 ms, a flip angle of 60o, a bandwidth of ±62.5 kHz, a FOV of 8 cm, a slice thickness of 1.7 mm, 2 number of excitation (NEX), and a readout of 512 or 511 projections. The acquired voxel size was equal to 0.156×0.156×1.7 mm³.

Figure 1 shows selected saturation recovery UTE images of a patella sample arranged in order of increasing TSRs, as well as bi-component fitting of the saturation recovery curves derived from small ROIs in the CC and superficial layers of articular cartilage, respectively. The CC was visualized as a high signal line (thin arrow) with excellent image contrast compared to the superficial layers of cartilage and fatty marrow with TSRs of 100 to 800 ms. The CC can be seen prominently with TSR varies below 200 ms, suggesting a shorter T1 compared to that of fatty bone marrow and the superficial layers of cartilage. Excellent fitting was generated for both the CC and superficial layers, suggesting a short T1 of 312 ms for the CC and a long T1 of 1469 ms for the superficial layers of cartilage. In the CC the short T1 component accounts for about 70% of the signal, with another 30% from the longer T1 component. A mono-exponential fitting was performed for the bone marrow fat, yielding a T1 relaxation time of 362 ± 35 ms, which proves consistent with the literature ³.

Figure 2 shows depth profile of both short and long T1 values and their fractions across the whole cartilage depth from the CC to the superficial layers. The CC has a short T1 of ~270 ms and a long T1 of ~1100 ms, which gradually increased to ~360 ms and ~1700 ms, respectively, for the superficial layers. The short T1 fraction gradually reduced ~50% for the CC to ~32% for the superficial layers, consistently with reduced proteoglycan content near the articular surface.

On average, the calcified cartilage has T1s ranging from 259 ms to 343 ms, with a mean of 294 ± 29 ms. The superficial layers have a mean T1 of 1396 ± 127 ms, which is slightly longer than the reported values of 1240 ± 107 ms in the literature ^3,4.

Bi-component exponential fitting provides excellent modeling of both the UTE saturation recovery data and the T2* decay data (results not shown). In summary, our methods to measure T1 values in the calcified cartilage and superficial layers of patellar cartilage appear reproducible and provide values that are consistent with those predicted, though not previously proven, in the literature. These values provides a means to optimize pulse sequences in a tissue specific manner. They also provide a means to perform T1 mapping of tissueswith T2s in the short T2* range to serve as an objective and quantitative means for the determining tissue structural integrity and in disease.