T1rho and T2 relaxation times of the normal adult knee meniscus at 3T: analysis of zonal differences

The meniscus of the knee joint serves important functions for shock absorption, joint stability and joint lubrication [1‐3]. Patients with meniscus injury or degeneration may develop osteoarthritis of the knee [4].

The extracellular matrix of normal meniscus is composed of 72% water, 22% collagen and 0.8% glycosaminoglycans [5]. This composition may change with degeneration [5]. Recently, with advances in imaging software and magnetic resonance imaging (MRI) sequences, it is believed that quantitative analysis of T1rho (T1 relaxation time in the rotating frame) and T2 relaxation times of articular cartilage and meniscus can be examined to reveal the underlying composition of extracellular matrix. Previous studies have shown usefulness of quantitative evaluation using MR relaxation times of the knee meniscus [6‐15]. However, few studies about zonal differentiation of normal meniscal T1rho and T2 relaxation times have been published [7, 12, 15]. Previous studies have evaluated vertical differences in MR relaxation times with comparisons of the inner one-third, middle one-third and outer one-third. However, more detailed analysis including evaluation of horizontal zonal differences between superficial and deep zones has not been performed. Therefore, the purpose of this study is to evaluate vertical and horizontal zonal differentiation of T1rho and T2 relaxation times of the entire meniscus from volunteers without symptom and imaging abnormality. We also investigate inter-observer and intra-observer reproducibility of meniscus segmentation in this study.

There have been prior studies which describe differences in collagen and proteoglycan content in different meniscal zones. One study used electron microscopy to reveal three distinct meniscal layers: 1) superficial network: the tibial and femoral surfaces are covered by a meshwork of thin fibrils, 2) lamellar layer: a layer of lamellae of collagen fibrils beneath the superficial network, and 3) central main layer: bundles of collagen fibrils oriented in a circular manner [17]. A study using hematoxylin and eosin staining and immunohistochemical staining revealed vertical zonal differentiation: fibrocartilage without any vessels was demonstrated in the white zone and highly vascularized loose and areolar connective tissue was demonstrated in the red zone [18]. There have been several animal studies describing meniscal zonal contents. In an experiment with bovine subjects, the inner one-third of meniscus is composed of 60% type II and 40% type I collagen, while the predominant collagen of the outer two-thirds is type I with a trace amount of types III and V [19]. In a study with procine, collagen content in the central zone was lower than in the peripheral zone, and glycosaminoglycan content in the central zone was higher than that in the peripheral zone [20]. The surfaces of the femur and tibia did not contain glycosaminoglycans [20].

Several studies have described the correlation between T1rho and T2 relaxation times with collagen or proteoglycan composition in articular cartilage. Akella et al. reported that an increase in T1rho relaxation time has a strong correlation with loss of proteoglycan in articular cartilage [21]. Goodwin et al. reported that T2 relaxation time is related to collagen fiber orientation in articular cartilage [22]. Since the extracellular matrix of meniscus is composed mainly of water, collagen and proteoglycans (similar to articular cartilage), we can expand the same discussion about relaxation times in articular cartilage to the meniscus. In other words, higher T1rho relaxation time likely represents relatively lower proteoglycan content, and differences in T2 relaxation time may represent differences of water content and collagen fiber orientation in the meniscus.

In our study, there were zonal differences in T1rho and T2 relaxation times of adult meniscus without symptom and imaging abnormality. In most segments, T1rho and T2 relaxation times of the superficial zone were significantly higher than those of the deep zone. Likewise, T1rho and T2 relaxation times of the inner zone were significantly higher than those of the outer zone. Higher T1rho relaxation time in the superficial zone and inner zone may represent relatively lower proteoglycan content. Higher T2 relaxation time in these zones may represent differences in collagen fiber orientation and/or increase in water.

There have been a few studies about vertical zonal differentiation of MR relaxation times of normal meniscus. Calixto et al. reported that, for both osteoarthritis and control subjects, the MMPH and LMPH showed significantly higher T1rho and T2 relaxation time in the inner zone compared to the middle and outer zones [15]. They suggested that different patterns of zonal T1rho and T2 relaxation times were caused by differences in macromolecular organization such as collagen ultrastructure and possible collagen cross linking [15]. Our results of T1rho comparison between inner and middle zones in MMAH and MMPH are consistent with Calixto’s findings. Similarly, our results of T2 comparison between inner and middle zones in LMAH and MMPH are consistent with Calixto’s findings. On the other hand, Tsai et al. reported that the T2 relaxation time increased significantly from the inner zone to the outer zone of LMPH and MMPH [7]. They suggested that the highest T2 value of the outer zone reflected the abundant blood vessels within vascular zone, and lowest relaxation time of the inner zone was related to the main substance of circumferential fibers [7]. Our result is consistent to Tsai’s study in LMPH, but not in MMPH. This discrepancy in MMPH may be due to differences in the methods of vertical meniscus segmentation. In vertical segmentation, we included the layer of superficial zone which was not included in Tsai’s study. In our study, since T2 relaxation time of the superficial zone was higher than that of the deep zone, relatively larger proportions of superficial zone in our inner zone may contribute to higher average T2 relaxation times of the inner zone. In addition, the degree of vascularity in each subject may influence the results, especially related to age, although the mean age of the subjects in our study was similar to Tsai’s study (29 years vs. 26.5 years, respectively) [7].

To our knowledge, no studies have been published to date about horizontal zonal analysis of meniscal T1rho and T2 relaxation times. Horizontal zonal difference inT1rho relaxation time of the meniscus may represent the differences of proteoglycan composition which contributes to strength against compressive stress, since compressive stress in the superficial zone of the meniscus is greater than in the deep zone. This can be explained by the same mechanism and depth dependent T1rho changes of articular cartilage of the knee. Nozaki et al. reported that average T1rho values in the superficial layer of the femoral cartilage were higher than those in the deep layer in the most locations of the healthy subjects’ femoral condyles [23]. Horizontal zonal difference of T2 relaxation time of the meniscus may also be explained similarly to articular cartilage, which demonstrates higher T2 relaxation time in the superficial zone than in the deep zone. Kaneko et al. reported that T2 relaxation time in the superficial zone of healthy articular cartilage was higher than in the deep zone [24].

The results in this study show there are zonal variations in menisci with no signal abnormalities on conventional MRI sequence in asymptomatic knees. Our results will be good reference in quantitative analysis for detecting slight meniscus injury or early stage of meniscus degeneration. Further studies with histological analysis would be needed to confirm our findings.

Compared with SPGR, b-FFE T1 rho sequence showed more significant zonal differentiation with higher inter- and intra-observer reproducibility. The reason for this result may be that the advantages of b-FFE sequences include higher signal-to-noise ratio (SNR) and better contrast-to-noise ratio (CNR) as compared with SPGR sequences [25]. In our study, average T1rho relaxation times of the meniscus on SPGR sequence were higher among all zones than b-FFE sequence. However, Nozaki et al. reported that the average T1rho value of the entire femoral cartilage with b-FFE was significantly higher compared to SPGR, and they discussed that the difference may be due to difference in sensitivity to proteoglycan content between the two sequences [26].

There are some limitations to our study. First, the number of subjects was small. Second, subjects recruited for this study had no confirmation with arthroscopy or surgery that there was no meniscus injury or degeneration. However, no subjects in this study had any abnormalities of shape or signal intensity in the meniscus with proton density weighted MR imaging. Third, we have no histological or immunohistochemical correlation. Fourth, in this study, average T1rho relaxation times of the meniscus were between 72.0 ms and 97.2 ms on b-FFE T1rho, and 80.7 ms and 104.9 ms on SPGR T1rho in horizontal or vertical zonal analysis. Average T2 relaxation times were between 60.3 ms and 104.6 ms. These values were higher than other reports of T1rho or T2 relaxation times of healthy meniscus [6, 7, 10, 13, 15]. T1rho and T2 relaxation times may vary with different imaging sequences and vendors. However, some pixels in the ROIs were eliminated for MR relaxation time analysis in the present study because calculated relaxation time of these pixels are 0 due to misregistration or inadequate fitting. These pixels were more prominent on SPGR T1rho and T2 relaxation time sequence. This can be one of the reasons relaxation times of SPGR T1rho relaxation times are higher than those of b-FFE T1rho in the present study. This fact may affect the results of zonal MR relaxation time analysis. To address this problem, higher signal-to-noise sequences including TSL = 0 for T1rho mapping and shorter TE for T2 mapping may be obtained.

There were zonal differences in relaxation times of the normal meniscus. In most segments, T1rho and T2 relaxation times of superficial zone were higher than those of deep zone, and T1rho and T2 relaxation times of inner zone were higher than those of outer zone. These differences can be due to zonal variations of collagen and glycosaminoglycan composition, although this speculation needs to be confirmed with histological or immunohistochemical analysis. Compared with SPGR T1rho, b-FFE T1rho showed more significant zonal differentiation with higher inter- and intra-observer reproducibility.