Feasibility of Ultrashort Echo Time T2^* Mapping in Comparison With T2 Mapping for Quantitative Evaluation of Meniscal Degeneration

The meniscus protects and stabilizes the knee joint via shock absorption and load distribution [1]. While asymptomatic meniscal tears are commonly found on knee MRI [2], a diagnosis should be carefully evaluated as these tears can lead to knee osteoarthritis (OA); degenerative-type meniscal tear is strongly associated with both radiographic OA and symptomatic radiographic OA [3, 4].

However, little is known about meniscal degenerative processes and early OA pathogenesis [3, 4, 5]. During meniscal degeneration, some components in the meniscal tissue such as proteoglycans, water, and collagen can change, affecting the molecular environment of the degenerating tissue. Such changes in the molecular environment would affect the T1, T1 rho, T2, and T2^* relaxation times of degenerating tissue on quantitative MRI [5, 6].

The normal meniscus has very short intrinsic T2 relaxation times due to abundant type I collagen and highly ordered collagen structure [5, 7]. The ultrashort echo time (UTE) sequence uses very short echo time (TE) for imaging, so it can capture rapidly decaying signals from short T2 tissues such as the meniscus [5]. With UTE sequences, the internal structure of the meniscus can be assessed using mapping techniques, and the quantification of T2 and T2^* values is also possible [7]. Quantitative MRI using T2 mapping or UTE-T2^* mapping can demonstrate the integrity of the collagen matrix, enabling assessment of the meniscal tissue composition using quantitative values that correlate well with the degree of meniscal degeneration [8, 9, 10].

Recent studies have suggested that T2 mapping and comparison with histological grades of degeneration as a reference standard is the most robust method for the characterization of meniscal tissue in early OA [11]. However, compared to T2^*, T2 mapping is more time-consuming and T2^* mapping has a better signal-to-noise ratio [12]. Nonetheless, T2^* mapping has a few technical limitations; the meniscus shows considerable changes in T2^* values based on orientation [13] and T2^* decay reflecting the anisotropic properties of the meniscal collagen fibers is strong, due to its highly organized structure [13]. Thus far, T2 mapping and UTE-T2^* mapping in the human meniscus have not been compared in vivo and therefore it is not clear whether UTE-T2^* mapping could replace T2 mapping.

Our study aimed to assess the feasibility of UTE-T2^* mapping for quantitative evaluation of meniscal degeneration compared with the standard T2 mapping of the meniscus and hyperintensity subregion within the meniscus. Our final goal was to investigate whether UTE-T2^* mapping could be a suitable substitute for standard T2 mapping, because UTE-T2^* mapping has advantages over standard T2 mapping such as shorter acquisition time.

In our study T2 values and T2^* values of the meniscus showed a strong correlation. T2m, T2^*m, T2h, and T2^*h values were significantly different among the morphological groups. All T2 and T2^* values increased with an increase in morphological grades.

T2 mapping is a technique of pixelwise illustrations of absolute T2 transverse relaxation times on a map [15]. T2 mapping allows direct T2 quantification using the grayscale intensity of each voxel, which is the output of a calculation performed independently at each corresponding spatial pixel from a series of input images [15]. T2^* mapping is a parametric imaging technique of T2^* relaxation time, which is the time of inherent decay of transverse magnetization caused by a combination of spin-spin relaxation (T2) and magnetic field inhomogeneity (T2^*) [15, 16]. In general, T2^* imaging requires a strong magnetic field and radiofrequency signals from a gradient [16].

T2 relaxation time has been used widely to evaluate meniscal tissue composition and is regarded as a reliable reference test for meniscal degeneration [11]. We investigated the T2^* relaxation times of meniscal tissue with a reference to T2 relaxation times and the correlation between T2 values and T2^* values. The results were statistically significant and reliably correlated. Therefore, we concluded that T2^* relaxation time was also a feasible method for evaluating meniscal tissue.

T2 mapping has a limitation in meniscal imaging due to the very short T2 components and the heterogeneity of the meniscal tissue itself [8]. Multi-TEs for T2 mapping sequences are between 10 ms and 100 ms in a usual knee MRI; however, the mean T2 value of a normal meniscus has been reported as 11 ± 4 ms [17]. Therefore, quantitative MRI techniques with extremely short TEs, such as UTE-T2^* or zeroTE-T2^*, have been suggested as better methods of quantifying menisci than standard spin-echo-based T2 mapping [8, 9, 18]. Meniscal T2^* can also respond to the extracellular matrix of collagen, glycoproteins, and proteoglycans, and be affected by myxoid changes, fibrocartilaginous separation of the matrix, extensive fraying, and tears that increase the heterogeneity of meniscal tissue [12].

UTE-T2^* imaging has several advantages over T2 imaging; however, it also has a few disadvantages. UTE-T2^* mapping has been reported to show different results among meniscal zones depending on the circumferential fibers [13]. Using UTE sequences with sub-millisecond TEs, the fibrillar network of the meniscus substructures is better depicted [19]. However, UTE-T2^* imaging studies have revealed significant artifacts in images and the orientation dependency of T2^* values: the magic angle orientation showed the highest T2^* values and at a fiber-to-field angle of 0° the T2^* values were lowest [13]. The relation between mono-exponential decay and bi-exponential decay in terms of fiber-to-field angle is still unclear [13]. This study was performed with standard 2D fast spin echo T2-weighted images and 3D UTE-T2^* images with variable TEs.

UTE-T2^* values have been higher in torn or degenerated menisci and significantly different between normal meniscus, degenerated meniscus, and torn meniscus [9]. Biexponentially fitted T2^* relaxation time has demonstrated a greater ability to distinguish normal and degenerated menisci in other investigations [20]. Another study revealed that both T2 and T2^* values were higher in abnormally thick meniscal lamellar layers compared with the values in normal lamellar layers, and only the T2^* value was statistically significant [21]. Interestingly, in an ex vivo study, there were significant increases in T2 and UTE-T2^* with increasing histology grades and high sensitivity and variable specificity but not in T2^* [6].

In this study, we assessed the feasibility of UTE-T2^* mapping to detect meniscal degeneration in comparison with T2 mapping. Both mean T2m and T2^*m and mean T2h and T2^*h, were significantly different among all the grades (p < 0.001) except for T2h grade 1 vs. T2h grade 2 (p = 0.002) and tended to be higher in the meniscus that was more degraded (p < 0.001 for both). These results suggest that UTE-T2^* mapping was feasible in the evaluation of meniscal degeneration as a standard T2 mapping, and the focal hyperintensity subregions within the menisci depicted higher grades of degeneration. The results of our study suggest that both T2 mapping and UTE-T2^* mapping are effective noninvasive diagnosing and monitoring methods for early degenerative changes within the meniscal tissue.

In previous studies, quantification of MRI mapping demonstrated a highly heterogeneous distribution of T2 and T2^* values with focal regions of elevated values in degenerated or torn menisci [6, 9, 21]. These findings implied an inadequacy in measuring mean T2 and T2^* values of menisci calculated by averaging across all voxel values within the whole meniscus. Therefore, in this study, we measured T2 and UTE-T2^* values of hyperintensity subregion within the menisci (T2h, T2^*h) as well as the cross-sectional area of menisci (T2m, T2^*m). However, both T2m and T2^*m, and T2h and T2^*h values were statistically significant.

The mean values of meniscus reported in previous studies with different imaging acquisition parameters were 8–12 ms for the T2 value [17, 22], and 5–8 ms for the UTE-T2^* value [7, 20], respectively. In this study, our values were in excellent agreement with previous values.

This study has several limitations. First, there was no histologic correlation of the meniscus. Our meniscal degeneration grading relied fully on the evaluation of morphologic MRI. Histologic validation as a marker for biochemical structures of the meniscus would be confirmative to verify T2 and UTE-T2^* values. Second, we analyzed the entire meniscus without considering zonal variations. Degenerative changes within the meniscus can occur unevenly according to the zonal distribution. Further studies analyzing T2 and T2^* values in each anatomic zone of the menisci are required. Third, we did not consider the possible confounding factors such as age, sex, and articular cartilage degeneration. Moreover, we did not correlate with OA clinical findings. Fourth, because only one radiologist analyzed the images, we could not evaluate the reproducibility of our measurements.

In conclusion, UTE-T2^* relaxation time is feasible for the quantitative evaluation of meniscus degeneration. With a few improvements UTE-T2^* mapping is a potential substitute for the standard T2 mapping, with improved efficacy.