Delayed gadolinium-enhanced MRI of meniscus (dGEMRIM) and cartilage (dGEMRIC) in healthy knees and in knees with different stages of meniscus pathology

Osteoarthritis (OA) affects approximately 10 % of the elderly and is a leading cause of disability [1]. OA is suggested to be a whole-joint disease involving cartilage, ligaments, periarticular muscles, subchondral bone and menisci [2]. The meniscus has a critical protective role for knee joint integrity by absorbing impact and distributing load [3, 4]. Overall, it is estimated that the knee meniscus carries 45 to 75 % of the total joint load [5]. Loss of meniscal function has been strongly associated with development and progression of radiographic OA [6, 7]. However, OA-related changes observed by radiography are late events in the degenerative process [8, 9]. In OA, the cartilage matrix undergoes pathological molecular cleavage with loss of type II collagen and glycosaminoglycans (GAG) [1, 10]. To increase our understanding of pathogenic mechanisms in the early stage of the disease, anterior cruciate ligament and meniscus injured patients, with increased risk for OA, may serve as models [6, 11]. Delayed gadolinium enhanced magnetic resonance imaging of cartilage (dGEMRIC) is an imaging technique that can identify pre-radiographic degenerative changes in articular cartilage, and potentially also in the meniscus [12, 13]. When this technique is applied to the latter tissue, it is designated gadolinium enhanced MRI of meniscus (dGEMRIM). In dGEMRIC, the fixed charged density of GAG is studied in vivo through quantitative measurement of the longitudinal relaxation time (T1) of the articular cartilage in the presence of the negatively charged contrast agent Gd-DTPA²⁻ (gadolinium diethylene triaminepentaacetic acid). Recently, we have shown in healthy subjects that the contrast medium distributes also into the meniscus after intravenous injection (dGEMRIM) [13]. The purpose of our study is to evaluate dGEMRIM as a diagnostic tool in menisci with MRI verified pathological changes. The literature has indicated conflicting results concerning the GAG content in pathological meniscus [14‐18]. Since these results may differ depending on the grade of pathological changes we have two separate main hypotheses. Firstly, we assume that meniscus with pathological grade 2 changes will display values suggesting an increased content of GAG compared to healthy meniscus and secondly, we assume that meniscus with pathological grade 3 changes will display values suggesting a decreased content of GAG compared to healthy meniscus. The specific aims of the study were first to investigate whether there is a measurable uptake of contrast agent in pathological menisci and when the maximum level of contrast agent was obtained, as measured in minutes after contrast injection. Second, we investigate whether there is a difference between the uptake of contrast agent in the posteromedial pathological meniscus and the posteromedial healthy meniscus in a previously examined knee without symptoms.

In our study we show that dGEMRIM can differentiate between pathological (grade 2) and healthy meniscal tissue. The decreased distribution of Gd-DPTA²⁻ in the pathological menisci grade 2 suggests more negative charge and an increased GAG concentration in this meniscus pathology. These results are consistent with previous meniscus studies that showed increased GAG concentration by histology and biochemical analysis [14, 15]. In addition, it has also been reported that proteoglycan content (μg/mg dry weight) in menisci with meniscal tears increase in relation to the severity of the meniscal degeneration [16]. Interestingly, the proteoglycan content in the meniscus does not seem to increase in rheumatoid arthritis [22].

A previous dGEMRIM study of 17 subjects with knee symptoms and MRI verified pathology in cartilage and menisci, showed a trend towards decreased T1_Gd90 values (i.e. values obtained 90 min after injection of contrast agent) in pathologically changed meniscus (posteromedial grade 2; n = 5) compared to healthy meniscus within the same knee joint [17]. All patients in that study were diagnosed with OA (Kellgren & Lawrence grade 1 or 2) and the results are in agreement with the results for the meniscus tissue with pathological changes grade 3 in our study. ∆R1 values, not presented in the study by van Tiel et al., may have enabled a more thorough comparison with our study. It can be speculated that increased GAG synthesis, reflected by longer T1_Gd and decreased ΔR1 values for pathological meniscus grade 2 in knee joints without cartilage damages in our study, is an early event in meniscus pathology. In contrast, in later stages of degenerative disease as presented in our study by pathological meniscus grade 3 and in the study by van Tiel et al. (knee joints with meniscus changes grade 2 and diagnosed OA, Kellgren & Lawrence grade 1 or 2), a more extensive damage to the knee joint results in a facilitated diffusion of contrast medium into the meniscus with corresponding shorter T1_Gd and increased ΔR1.

In our study the maximum level of contrast medium in healthy femoral cartilage occurred approximately 120 min after injection. This is consistent with earlier results [23].

Comparing the femoral cartilage between healthy knee joints and knee joints with pathological grade 2 changes in the posteromedial meniscus, we found no differences with regard to the T1_Pre, T1_Gd and ΔR1 values on the lateral side. On the medial side, there was a difference in T1_Gd but not in the T1_Pre or ΔR1 values. These findings strengthen the results from the diagnostic MRI, suggesting that the posteromedial meniscus is the only tissue in the examined knee joint with pathological meniscus changes grade 2 that has pathological changes.

It is important to acknowledge that other factors than GAG content may contribute to the distribution of contrast agent into a specific tissue. The fact that the wet weight GAG concentration in the meniscus is 0.3 % as opposed to 2.0 % in articular cartilage [24] and that collagen network differs between the two tissues [25, 26], suggests that GAG content is not the only factor that determines the contrast distribution into the meniscus. Previous studies [18, 27] addressed this issue in more detail using both an ionic (inversely related to the GAG content) and a non-ionic (no known interaction with GAG content) contrast agent when investigating meniscus and femoral cartilage in knee joints with OA. Data from these studies demonstrated similar differences in T1_Gd values in meniscus and cartilage when comparing subjects with OA and healthy volunteers, with both kinds of contrast agents, indicating that GAG is not the sole factor influencing the uptake of contrast agent. The authors of these studies speculated that the integrity of the collagen network may play a role, as well as increased diffusion due to degenerative changes of both the meniscus and cartilage.

Using dGEMRIC, Hawezi et al. [28] showed a depth-wise variation in contrast distribution in femoral knee cartilage. The maximum concentration of the contrast agent was observed at different time points in different cartilage layers: at 120–180 min in the superficial layer and at 240 min or more in the deeper layers. Furthermore, the differences increased with cartilage thickness. Although the GAG content and the composition of different collagen types differ when comparing different parts of the meniscus [23‐25], it is unlikely that there is a larger depth-wise alteration of the concentration of the contrast agent in the meniscus since the contrast agent diffuses into the meniscus from two sides. These authors further discussed a wash out effect of contrast agent occurring after approximately 120 min. It has previously been demonstrated that following an intravenous Gd-DTPA²⁻ injection, the gadolinium concentration in blood plasma will reach its maximum and then decrease within 120 min [29]. Assuming similar concentration decay in synovial fluid, the gadolinium concentration of articular cartilage will be higher than that in synovial fluid and a wash out of gadolinium from the cartilage will start approximately after 120 min. The wash out effect needs to be considered in order to choose an optimal time point post contrast to investigate pathological menisci with dGEMRIM. From our study as well as previous studies, it can be concluded that 120 min post contrast seems optimal for the evaluation of both meniscus and cartilage tissue [13].

It may be argued that the exclusion of grade 1 meniscus lesions is a limitation of the present study. However, grade 1 changes are very difficult to depict when drawing ROIs on a specific slice. It is therefore a significant risk that the measurements would be calculated on the wrong slice and resulting in incorrect values.

The results of menisci with pathological changes grade 3 may be an issue for discussion. Firstly, these changes are often considered as an advanced stage of OA and an examination with contrast enhanced MRI would in clinical use be of less interest. Furthermore, meniscal tears might fill with contrast agent directly from joint cavity, causing erroneously T1_Gd relaxation times. Therefore, this must be taken in consideration when evaluating the obtained values using dGEMRIM to examine menisci with pathological grade 3 changes.

An issue using gadolinium as contrast agent is the possible occurrence of nephrogenic systemic fibrosis (NSF). This has mainly been observed in patients with severe chronic or acute renal failure following exposure to gadolinium-based contrast agents [30]. However, after the introduction of clinical guidelines restricting the use of gadolinium in these patients, reports from several academic medical centers have been published indicating that no new cases of NSF have been observed among this category of patients [31, 32].

dGEMRIM and dGEMRIC may be feasible to combine in vivo, preferably with one examination before and one 2 h after contrast injection. Possible different dGEMRIM patterns at different stages of meniscus lesions must be taken into account when evaluating meniscus pathology.