UTE T2∗ mapping detects sub-clinical meniscus degeneration

Abstract

The meniscus is increasingly being recognized as an important tissue in knee osteoarthritis (OA). We know that individuals who have undergone meniscectomy are at least 2.6 times more likely to develop knee OA within 22 years1 and that, by the time end-stage lateral knee OA is reached, complete maceration or resection of the meniscus has occurred in approximately 50% of cases2. However, relatively little is known about early stage meniscal degeneration, before tears, extrusion, myxoid degeneration, maceration or resection occurs because, until recently, the tools necessary to directly image the meniscus have not been available. With magnetic resonance imaging (MRI), it is possible to non-invasively visualize the meniscus; however, it appears dark (as signal void) on traditional MR images. This is because the meniscus has a short T2 relaxation time due to its highly organized collagen ultra-structure, with most fibers oriented in the circumferential direction. T2 relaxation time refers to the time required for the transverse component of magnetization to decay after a radiofrequency (RF) excitation pulse is applied. Most clinical scanners use echo times (TE) of 8–20 ms; however the mean T2 times in the meniscus have been reported to be approximately 11.4 ± 3.9 ms in healthy individuals, therefore, some individuals will have T2 relaxation times at the bottom end of this range3. The paper by Williams et al. shows that, with ultrashort echo time (UTE) sequences, high quality images of the meniscus with shorter TEs, can be acquired. The goal of UTE sequences is to shorten the TE time such that there is signal in the short T2 tissues4,5. To do so, shorter RF excitation pulses and faster readout methods are used. Pulses used to date include half sinc pulses5, nonselective pulses6, selective pulses7, and discrete pulses8. Readout techniques include radial trajectories for 2D sequences9 and spherical, spiral or twisted spiral trajectories for 3D sequences10–12. Due to the shorter RF excitation pulse and the faster readout methods, the TEs are 20–1000 times shorter than conventional MR sequences and the shortest TEs range from 0.008 to 0.07 ms for the RF excitation pulses mentioned above. It is important to note that in UTE imaging, the term TE can be used to define the ‘dead time’ between the end of the RF pulse and the beginning of the readout gradient ramp (data acquisition period) or the time elapsed from the center of the RF pulse to the time when data at the center of k space is sampled. Therefore, care must be taken when comparing TE times between studies and techniques. Regardless, with UTE it is possible to directly image short T2 tissues, such as the meniscus, tendons, ligaments and synovium. Since UTE is a relatively new method, researchers continue to develop improved 2D and 3D sequences for clinical scanners which are faster, have better resolution, and improved contrast. The 2D acquisitions are single or multi-slice image sets with in-plane resolutions of 0.3–0.8 mm9,13, slice thicknesses of 3–5 mm, and imaging times of 3–17 min10,14,15 (with the longest times corresponding to 20 slice sets). The 3Dimage sets have been acquired with isotropic resolutions of 0.6–0.9 mm in imaging times of 5–20 min10 and with in-plane resolutions of 0.28 or 0.14 mm and slice thicknesses of 2 or 3 mm in scan times of 5 or 10 min7. Contrast between short T2 tissues and surrounding tissue, which may or may not have short T2 relaxation times, can be improved through several other imaging-based techniques in UTE. Some sequences use conventional fat saturation or suppression techniques7,14 while others use long T2 suppression (cartilage, fluid) methods9,10,14. Spectroscopy can also be used to separate fat and water components9. Finally, intravenous contrast agents can be administered to enhance the signal within the short T2 tissues and improve contrast14. As with conventional MRI techniques, UTE can be used for quantitative, regional mapping of the tissue. To date, T2, T2* and T1ρ relaxation time maps of the meniscus have been created using UTE. There is some evidence to suggest that T2 and T2* are related to collagen content and T1ρ is related to proteoglycan content; however, this has been shown in cartilage16,17 and not the meniscus. Similar to T2, T2* describes the transverse relaxation of the tissue; however, in this case additional factors, such as magnetic field inhomogeneity and tissue susceptibility, affect the relaxation process. These factors cause the spins to relax more quickly, therefore, T2* is always shorter than T2. T1ρ relaxation, or spin-lattice relaxation in the rotating frame, is similar to T2 relaxation except a spin-lock pulse is applied in the transverse plane which affects the relaxation process. Using conventional MRI, T1ρ and T2 are most commonly measured3; however, T2* may provide us with different information about the tissue than T2 and T1ρ and therefore measuring all of these quantities may be important. 2D UTE, 2D UTE spectroscopy, and 3D UTE have all been used for T2* mapping of the meniscus. To measure T2* relaxation a series of images with different TEs must be acquired. TEs between 0.008 and 40 ms have been used to date7,15 Ideally the TEs chosen sample a range that captures the entire T2* relaxation process. Mono-exponential curves can then be fit to the series of images on a voxel-by-voxel basis to estimate T2*. Mean T2* relaxation times have been reported as 4–10 ms for the meniscus14,15; the mean T2* relaxation time of 9.8 ms reported by Williams et al. falls within this range. T2* has also been shown to vary by region in the meniscus; one study showed that the mean T2* relaxation times in the inner (white), middle (transition) and outer (red) zones were 7.8 ± 1.2, 6.5 ± 1.4, and 12.6 ± 1.9 ms, respectively14. The study by Williams et al. is the first to compare quantitative MRI maps to histological tissue sections in the meniscus. They found that menisci with lower T2* relaxation times have lower histologic scores and vice versa. The authors suggest that heterogeneity in the T2* maps are indicative of collagen disorganization. These findings highlight the need for further work that aims to understand the relationship between quantitative MRI measures (such as T2*, T2 and T1ρ relaxation) and meniscal structural components. Williams et al. used an ACL injury model to determine the sensitivity of T2* relaxation times to meniscal degeneration. This is a particularly useful model for studying meniscal degeneration because up to 82% of injured individuals have radiographic degenerative changes in the joint and up to 50% of injured individuals eventually develop radiographic OA, even with ACL reconstruction18. Williams et al. very nicely demonstrated the sensitivity of their technique by showing increased T2* relaxation times in the posterior meniscal horns of individuals with ACL injury with or without a concomitant mensical tear when compared to asymptomatic individuals. These results compliment previous findings from studies using conventional MRI techniques which have shown that T1ρ was greater 1–3 months post-ACL surgery when compared to controls19. T2 and T1ρ have also been shown to distinguish between individuals with and without OA3. UTE shows great promise for improving our understanding of early degenerative changes in short T2 tissues, such as the meniscus. The results by Williams et al. in the ACL population suggest that the technique has the potential to be helpful in developing early therapies for meniscal degeneration prior to tearing or frank tissue loss. Further, UTE can be used to study other short T2 tissues including tendons, ligaments and the synovium, whose importance is increasingly being recognized in the OA disease process.