Biochemical MRI in Musculoskeletal Applications

Technology Tissue function Biochemical MRI in Musculoskeletal Applications

Siegfried Trattnig; Štefan Zbýň; Vladimir Juras; Pavol Szomolanyi; Stephan Domayer; Iris-Melanie Noebauer-Huhmann; Goetz Welsch; Benjamin Schmitt

MR Centre – High field MR, Department of Radiology, Medical University of Vienna/Vienna General Hospital, Vienna, Austria

1. Biochemical MRI assessment 1.1) Proteoglycan-sensitive MRI exchange-dependent saturation transfer of cartilage and cartilage techniques (CEST) [6]. Of these techniques, only repair Although no MRI sequence is really dGEMRIC has so far become clinically use- To visualize the constitution of articular 100% specific for only proteoglycans or ful in cartilage repair imaging. cartilage and cartilage repair tissue, collagens, there are MRI methods that a variety of different MRI methods are reportedly focus mainly on one compo- a) Delayed gadolinium-enhanced MRI available. Considering the composition nent of articular cartilage. The nega- of cartilage (dGEMRIC*) of healthy hyaline articular cartilage, tively-charged proteoglycan, composed GAG are important to the cartilage tissue’s these methods can depict relatively spe- of a central core protein to which GAG biochemical and biomechanical func- cifically either one component of carti- are bound, can be visualized by delayed tion. GAG are the main source of fixed lage or a combination of different com- gadolinium-enhanced MRI of cartilage charge density (FCD) in cartilage, and ponents. Although some more recent (dGEMRIC*) [3], sodium MR imaging are often decreased in the early stages techniques, in particular CEST, have not [4, 5], and, very recently, chemical of cartilage degeneration [7], or in repar- yet been validated sufficiently, studies have shown promising initial results. Articular cartilage is complex, dense, 1A 1 (1A) Shows a sagittal connective tissue that relies on the dif- 3D GRE image in a patient fusion of solutes for nutrition [1]. after microfracturing at 3T. Responsible for the biomechanical prop- In the area of microfracture erties of articular cartilage is the extra- only slight thinning of the cartilage layer can be seen. cellular matrix, mainly composed of (1B) Precontrast* T1-map water (~75%), collagen (~20%), and shows no difference in proteoglycan aggregates (~5%) [1, 2]. T1 relaxation times in the Water either freely moves throughout repair tissue; however, in the matrix or is bound to macromole- the postcontrast* image (1C) a significant drop in cules. Collagen in hyaline cartilage is T1 values in the repair largely type II, which creates a stable tissue can be seen, which network throughout the cartilage. The corresponds to a low gly- negatively-charged proteoglycans are cosaminoglycan content. composed of a central core protein to which glycosaminoglycans (GAG) are bound.

86 MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world Tissue function Technology

ative cartilage after cartilage repair [8]. requires a break in between the two b) 23Na (Sodium) imaging of cartilage Intravenously administered gadolinium MRI scans, where the contrast agent The major advantage of sodium MRI in diethylenetriamine pentaacetate anion, must be administered and a delay of at musculoskeletal applications is that it (Gd-DTPA2-) penetrates the cartilage through least 90 minutes after injection is needed is highly specific for glycosaminoglycan both the articular surface and the sub- for penetration of the contrast agent content, and, since the sodium from chondral bone. The contrast equilibrates into the cartilage. Scan time reduction, surrounding structures in the joint is low in inverse relation to the FCD, which is, compared to the standard inversion (< 50 mM), cartilage can be visualized in turn, directly related to the GAG con- recovery (IR) evaluation, has been with very high contrast without the centration. Therefore, T1, which is deter- achieved with a new approach using fast requirement for any exogenous contrast mined by the Gd-DTPA2- concentration, T1-mapping based on GRE-technique agent, such as with dGEMRIC*. becomes a specific measure of tissue [12]. Although the 90-minute delay is The recent proliferation of 7T1 whole- GAG concentration, suggesting that still required, this might increase the body MRI scanners in clinical research Gd-DTPA 2- enhanced MRI has the poten- clinical applicability of the dGEMRIC* centers has had a significant impact on tial to monitor the GAG content of carti- technique. Other drawbacks of dGEMRIC* sodium MRI and its potential for clinical lage in vivo [9]. Thus, T1-mapping, are the use of i. v. contrast agent admin- use. Since signal-to-noise (SNR) scales enhanced by delayed administration of istration, the necessity of a double dose linearly with increasing field strength, 2- Gd-DTPA (T1 dGEMRIC*), can be con- of contrast agent, and the fact that and because of the lack of B1 penetration

sidered the most widely used methodol- uptake and distribution of contrast agent and B0 susceptibility that pose problems ogy for detecting proteoglycan depletion is not only dominated by GAG. with proton imaging, sodium MRI can in articular cartilage (Fig. 1), and has Since GAG content is responsible for be particularly advantageous at higher shown promising results [10, 11]. cartilage function, particularly its tensile fields. The low gyromagnetic ratio of As differences in pre-contrast values strength, the monitoring of the develop- sodium also means significantly lower between cartilage repair tissue and nor- ment of GAG content in cartilage repair power deposition compared with proton mal hyaline cartilage are larger compared tissues may provide information about imaging, which thus reduces SAR prob- to early cartilage degeneration, in carti- the quality of the repair tissue. A study lems at 7T1. It is, therefore, very likely lage repair tissue, the pre-contrast T1 by our group showed that dGEMRIC* that the improved SNR sodium MRI at values must also be calculated [8]. The was able to differentiate between differ- 7T1 can provide a robust tool for quan- concentration of GAG is represented by ent cartilage repair tissues with higher titative imaging of MSK structures in delta ΔR1, i.e., the difference in relax- delta ΔR1 values, and thus, lower GAG particular cartilage. Although sodium

ation rate (R1=1/T1) between T1precontrast content, in cartilage repair tissue after MRI has high specificity and does not

and T1postcontrast. Thus, the sequence must MFX, compared to MACT [13]. The appli- require any exogenous contrast agent, be performed twice, for pre-contrast cability of this technique has also been it does require special hardware capa- and delayed post-contrast T1-mapping. shown in regions other than the knee bilities (multinuclear) and specialized This increases the total scan time, and joint [14–16]. RF coils [5].

1B 1C

MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world 87 Technology Tissue function

2A 2B

With the application of a 7T whole-body MACT (p = .002) repair tissue, compared is transplanted, and in patients after system and a modified 3D GRE opti- to reference cartilage. These differences patella dislocation, have demonstrated mized for sodium imaging and dedicated were influenced by surgery type, but not the potential of sodium imaging in the multi-element sodium coils, we per- by age and follow-up interval. Although detection of early stages of cartilage formed a series of clinical studies. sodium SNR was not different between degeneration [19]. In a small group of 12 patients after the reference cartilage in the MACT and matrix-associated autologous chondro- BMS patients (p = .528), it was signifi- c) Chemical Exchange Saturation cyte transplantation (MACT), sodium cantly higher in MACT than in BMS repair Transfer (CEST) imaging enabled the differentiation tissue (p = .002). There was no differ- Chemical exchange between bulk water between sodium content, and thus GAG ence between the magnetic resonance protons and protons bound to solutes in the transplants, compared to native, observation of cartilage repair tissue can be visualized with MR imaging [20] healthy cartilage. In all patients, the (MOCART) scores for MACT and BMS by using the saturation transfer (ST) sodium SNR was lower in the repair tis- patients (p = .889). No correlation was technique, a common approach in sue compared to healthy cartilage. observed between the MOCART score nuclear magnetic resonance (NMR) [21]. A good correlation between sodium and repair tissue SNR with sodium imag- Consequently, the corresponding MR imaging and dGEMRIC* in the quantifi- ing (R = .111). Our results suggest that imaging scheme is referred to as CEST cation of GAG content was found in MACT provides higher GAG content and, MRI [22, 23]. In principle, CEST MRI patients after MACT [17]. therefore, higher-quality repair tissue relies on a reduction of bulk water MR In another study, 18 patients who had compared to the BMS techniques. Sodium signal as a consequence of selective sat- different cartilage repair surgeries (nine MR imaging at 7T1 (Fig. 2) might be uration of off-resonant (solute) spins bone marrow stimulation (BMS) and beneficial in the non-invasive evaluation and subsequent chemical transfer of the nine MACT patients), age-matched, with of cartilage repair procedure efficacy in saturation to spins of bulk water [24]. similar postoperative intervals and the knee [18]. A prerequisite for observation of CEST defect locations, were examined with Similar studies in patients after long-term phenomena within a chemical environ- sodium imaging. Sodium SNR was signif- autologous osteochondral transplanta- ment is the abundance of sufficiently icantly lower in BMS (p < .001) and tion (AOT) in which hyaline cartilage labile solute protons that are in contact

88 MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world Tissue function Technology

2C 2 Sagittal images of the knee of a 43-year-old woman obtained 42 months after 400 MFX cartilage repair procedure. The proton-density- weighted 2D-TSE image with fat suppression (2A), sodium 3D-GRE image (2B), and color- 300 coded sodium 3D-GRE image (2C). The cartilage repair tissue is situated between the two arrows. Color scale repre- sents the sodium signal inten- sity values. 200

100

with bulk water. The offset of the reso- a sensitive biomarker for cartilage scales linearly with the ratio of GAG nance frequency of the labile protons GAG content in in vitro studies of bovine protons to bulk water protons. (Δω) in Hz must be greater than the cartilage at 11.7T1 [25]. gagCEST imaging (Fig. 3) was shown to exchange rate (k) to provide CEST Since the introduction of the gagCEST be a valuable tool for the non-invasive effects that can be distinguished from imaging, more studies have been per- assessment of GAG content in vivo. For direct water saturation. For example, formed on animal and human subjects example, we could demonstrate that hydroxyl and amide protons of GAG to assess the feasibility of gagCEST gagCEST can be used to reliably detect were shown to be suited for CEST experi- imaging in vivo [27–29], and a large pal- GAG in the knee cartilage of patients ments (gagCEST) due to their chemical ette of imaging sequences for fast and after cartilage repair surgery [33]. In this exchange properties [25]: GAG –OH reliable detection of CEST effects have study, which was performed at 7T1, protons resonate at frequency offsets of been proposed [27, 30–32]. The vast a 3D GRE-based gagCEST measurement 1 and 1.5–2 ppm downfield from bulk majority of techniques is based on repet- technique was used, and results from water, and k can be on the order of up itive image acquisition with pre-satura- 23Na MRI served as a reference for GAG to 1000 s-1 [26]. This means that mag- tion at different offset frequencies (Δω), measurements. Moreover, the potential netic field strengths greater than 3 Tesla which allows plotting of the remaining of gagCEST for GAG evaluation in inter-

are beneficial for gagCEST experiments, bulk water signal (MSat), normalized to vertebral discs (IVD) at 3T has been

as the condition Δω>k will be increas- a reference (MRef), against the RF pre- demonstrated in healthy volunteers ingly fulfilled. At higher fields, i.e., with saturation offset (z-spectrum). In the [34, 35] and in patients with lower back increasing frequency differences, the z-spectrum, CEST effects appear asym- pain [36]. Given the results from the confounding direct water saturation metric with respect to the water reso- latter studies on IVDs from our group, it (RF spillover) decreases. In combination nance and, thus, calculation of the is reasonable to assume that gagCEST with longer T1 relaxation times, this asymmetry of the magnetization transfer can also be used to assess cartilage GAG

factor makes ultra high field strengths, ratio, (Eq. 1) (MTRasym(Δω )=(MSat(-Δω)- content at 3T. Thus, the indicated poten- 1 such as 7T , potentially favorable for MSat(+Δω))/MRef), provides a positive mea- tial of gagCEST for the clinical routine CEST experiments. Consequently, gagCEST sure of CEST effects. In first order, the is tremendous because it also offers

imaging was first demonstrated to be magnitude of MTRasym values in cartilage advantages compared to other GAG-sen-

MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world 89 3A

3 High-spatial- sitive imaging techniques, such as resolution (3A) dGEMRIC* and sodium imaging. These morphologic, advantages include the relatively short (3B) gagCEST, acquisition time, which covers the entire and (3C) 23 Na volume of a knee joint in ~10 minutes MR images of the knee joint carti- [33], no administration of contrast lage of a patient agent, and simple implementation into after MFX in the a standard imaging protocol. medial femoral condyle. Color bars on (3B) and 1.2) Collagen and water-sensitive (3C) represent MRI techniques MTR asym values Although the differentiation of particu- summed over off- lar components of articular cartilage sets from 0 to 1.3 that are visualized by a specific sequence ppm (gagCEST) and sodium SNRs, is not entirely possible, the classic bio- respectively. Both chemical MR methodology that focuses techniques show on the collagen content of articular car- decreased signals 3B 20 tilage is T2-mapping [37]. In addition to in repair tissue compared to sur- the transverse relaxation time (T2) of 15 rounding native articular cartilage, recently, T2* relax- tissue. ation is being discussed for the depiction 10 of the collagen matrix [38]. Magnetiza- tion transfer contrast (MTC) might also 5 play a more important role in future approaches [39]. 0 T2 relaxation time mapping The transverse relaxation time (T2) of cartilage is a sensitive parameter for the evaluation of changes in water and collagen content and tissue anisotropy [37]. Cartilage T2 reflects the interaction of water and the extracellular matrix on a molecular level. The collagen fiber orientation defines the layers of articular cartilage. Thus, the three-dimensional organization and curvature of the collagen network, influenced by water mobil-

3C ity, the proteoglycan orientation, and 18 the resulting magic angle at 55º (with

16 respect to the main magnetic field B0) 14 influence the appearance of T2 [40, 41]. In healthy articular cartilage, an increase 12 in T2 values from deep to superficial car- 10 tilage layers can be observed, based on the anisotropy of collagen fibers running 8 perpendicular to cortical bone in the 6 deep layer of cartilage [42]. Histologically validated animal studies have shown this zonal increase in T2 values as a marker of hyaline or hyaline-like cartilage structure after cartilage repair procedures in the knee [43, 44]. To visualize this zonal variation in vivo, high spatial resolution Tissue Function Technology

is essential, which can already be In addition to the knee joint, studies 20 cases after microfracture and MACT achieved with high-field MR, together have been performed to analyze differ- of the ankle joint cartilage. The analysis with dedicated multi-channel coils in ent repair surgery techniques in the confirmed that there was no significant clinical approaches (Fig. 4). In cartilage ankle joint. The initial results at 3T were difference between the average T2 val- repair tissue, global (bulk) T2 values, highly surprising; the average T2 value ues of the repair tissue after either tech- as well as line profiles, have shown an of the repair tissue after microfracture nique. The major advance was, however, increase in the early post-operative fol- did not differ significantly from the repair the possibility of a zonal analysis of the low-up period, which might enable visu- tissue after matrix associated autologous healthy cartilage in asymptomatic volun- alization of cartilage repair maturation chondrocyte transplantation (MACT). teers, and in patients after cartilage [45]. Another study by our group further These results were very different from repair [47]. showed the ability of zonal T2 evalua- the analyses in the knee, but did, how- The values of the superficial layer of tion to differentiate cartilage repair tissue ever, agree with the clinical evidence the repair tissue were comparable to after microfracture (MFX) and MACT in the literature. those of articular cartilage; however, [46]. Whereas cartilage repair tissue after A major drawback of T2-mapping in the there was a substantial difference in the MFX – histologically seen as fibrocarti- ankle was that the SNR even at 3T was deep layer. A significantly higher T2 lage – shows no zonal increase from deep not sufficient to perform a zonal analysis was demonstrated for the repair tissue to superficial cartilage aspects, repair of the very thin cartilage layer of the when compared to the healthy refer- tissue after MACT – histologically reported talus. Based on the results in the knee, ence. Again, there was no difference as hyaline-like – shows a significant this might be crucial for characterizing between the cartilage repair techniques. stratification. the quality of cartilage repair tissue, as T2-mapping yielded repair tissue with The current evidence shows that the the zonal T2 variation correlates with comparable properties in the MR analysis, quality of the repair tissue is the key to the degree of organization of the colla- which further substantiates the notion success in the long term. Patients with gen network, and therefore, with the that cartilage repair of the ankle cannot repair tissue similar to the native articu- general quality of the repair tissue. be compared to the knee. Bone marrow lar cartilage are more likely to have good The optimization of the multi-echo spin stimulation techniques are less invasive mid-term results, and in the knee, cell- echo T2-mapping algorithm for 7T yielded and less costly than MACT, and might based techniques result more often in a high SNR, better resolution within a yield a comparable or even better out- hyaline-like repair tissue than does clinically feasible scan time, and enabled come for this joint. It should be noted microfracturing. us to perform an initial comparison of that the evidence does not allow for

4A 4B

100 40

75 30

50 20

25 10

0 0 T2 (ms) T2*(ms)

4 Axial T2-maps (4A) and T2*-map (4B) of the cartilage layer of the patella at 7 Tesla1 nicely demonstrate the zonal variation in the collagen fiber network in cartilage in this volunteer.

MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world 91 Technology Tissue function

a conclusion at this time; however, 5A 5B biochemical MRI can be of substantial help in determining future treatment algorithms in this field. In osteoarthritis, however, T2-mapping has shown varying results [48–50] and the role of T2 as an absolute quantification parameter must be further analyzed. Thus, it is not yet totally clear whether a slight increase or decrease in T2 relaxation times can be correlated to morphological changes. Nevertheless, T2-mapping seems to offer potential in this area as well. Conclusions would be more easily derived, however, from a longitudinal evaluation of the same sub- ject, with MRI performed at the same time of day. 2. Biochemical MRI assessment of the intervertebral disc For biochemical MR imaging of the intervertebral disc (IVD), the following MR methods and techniques were optimized by our group: T2-mapping; T2*-mapping; diffusion-weighted imaging at 3 Tesla; and recently, sodium imaging at 7 Tesla; and Chemical Exchange Satura- tion Transfer (CEST) at 3 Tesla (Fig. 5). In an initial trial on 34 patients with lower back pain but without radicular symp- toms we compared standard morpholog- 5 Comparison of morphological and biochemical imaging of the intervertebral discs ical MRI at 3T with T2-mapping [51, 52]. of the lumbar spine in a healthy volunteer using T2 FSE (5A), T1 FSE (5B), T2-map (5C), For T2-mapping, all discs of the lumbar and T2*-map (5D). spine were analyzed by classification into five equally-sized regions-of-interest (ROI) (20% each) in the sagittal plane, where the anterior and posterior 20% tional MRI findings of the spine at 3T 3 Tesla demonstrated significant differ- were defined anatomically as the annu- was performed [53]. In a total of 265 ences in T2 values between different lus fibrosus (AF) and the 60% area in discs, there were 39 focal herniations, dislocations of the discs of the lumbar between the anterior and posterior rep- 10 annular tears, 123 disc bulgings, and spine. resented the nucleus pulposus (NP) of 103 normal discs. We found a statisti- In addition, methodologically, T2-map- the IVD. We found a good correlation cally significant difference between the ping of the IVD was compared to T2*- between the morphological Pfirrmann nucleus pulposus T2 values in discs with mapping at 3 Tesla in 30 patients with classification and T2 relaxation time val- annular tears and all other groups with lower back pain [54]. The highest varia- ues. To further increase the sensitivity the discs with annular tears, showing tion with an increase of T2 and T2* of the analysis, the posterior annulus significantly lower T2 values compared values from the AF to NP was found in fibrosus was further subdivided into two to discs without annular tears. The dif- Pfirrmann grade I and decreased with 10% ROIs to consider the axial load on ference in NP T2 values between discs higher Pfirrmann grades II-IV. In addition the posterior column of the spine. This with focal herniations and normal discs to T2-mapping, T2*-mapping revealed analysis provided a possible marker for was also significant. However, no signifi- further changes in the posterior annulus the early stage of disc degeneration cant difference was found between NP fibrosus. The correlation between T2 [51]. T2 values between disc bulging and and T2*-mapping showed a moderate Furthermore, a comparison of quantita- disc herniation. Thus, for the first time, correlation of 0.21 to 0.356. The clear tive T2 values of the IVD with conven- quantitative T2-mapping of the NP at differentiation between different stages

92 MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world Tissue function Technology

5C 5D scopic changes [57]. Glycosaminoglycan (GAG) content was observed to be dou- bled in pathologic tendons in biochemical studies [58]. In a study our group, the feasibility of sodium magnetic resonance imaging at 7T1 for the diagnosis of Achilles tendinopathy was investigated [59]. The cohort comprised twenty healthy volunteers with no history of pain in the Achilles tendon and eight patients with clinical findings of chronic Achilles tendinopathy. We used a three- dimensional gradient-echo sequence optimized for sodium imaging. The parameters were as follows: TR 17 msec, TE 8.34 msec, FOV 199 x 199 mm, sec- tion thickness 3 mm, flip angle 50º; 12 signals acquired, acquisition matrix 224 x 224 pixels. The total measurement time for sodium imaging, including flip angle calibration and localizers, was about 32 minutes. Moreover, five fresh human cadaver lower legs from four different subjects (mean age of 48 ± 8 years) were used as a reference for sodium signal and GAG content relation. These samples were analyzed biochemi- cally to obtain GAG and water content. The study found that the mean bulk sodium SNR was 4.9 ± 2.1 in healthy control subjects and 9.3 ± 2.3 in patients with Achilles tendinopathy, and that the difference between the means was statistically significant. The correlation between sodium SNR values acquired from three regions of five cadaver Achil- les tendons and GAG content was found of disc degeneration and the possibility 3. Biochemical MRI assessment to be high, with a Pearson coefficient of quantification by T2 and T2*-map- of tendons of 0.71. ping may provide a new tool for a more This study showed a statistically signifi- sensitive monitoring of different thera- 3.1) Sodium imaging of tendons cant increase in sodium SNR in patients pies in patients with lower back pain. Tendons, which are dense, fibrous con- with Achilles tendinopathy, compared In a preliminary study, sodium imaging of nective tissues, are often affected by with healthy tendons (Fig. 6C). More- the IVD of the lumbar spine at 7 Tesla was tendinopathy. This clinical condition is over, the study also revealed abnormal compared with T2-mapping and morpho- defined as a syndrome of tendon pain, sodium signal values in the whole ten- logical grading by the Pfirrmann classifica- tenderness, and swelling that affects don, although morphologically focal tion. No correlation was found between tendon function. Morphologically, degen- abnormalities with focal thickening were sodium imaging and T2-mapping, con- eration of the Achilles tendon, for exam- seen. This suggests that the Achilles firming that both methods detect different ple, leads to thickening of the tendon tendon is affected diffusely in tendinop- components of the IVD [55]. [56]. Tendinopathy is accompanied by athy rather than focally. Sodium signal a disaggregation of the microfibrillar values mostly correspond to GAG content bundles due to the greater quantities of in the Achilles tendon, which, as shown water and proteoglycan macromole- in in vitro studies, is increased with ten- cules. In chronic Achilles tendinopathy, dinopathy [59]. It has been shown that, biochemical alterations presage macro- despite the relatively challenging techni-

MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world 93 Technology Tissue function

6A 6B 6C T2* [ms]

2.50 1408

0.00

6 Morphological image (6A), T2*-map measured by 3D-UTE multi-echo sequence (6B) acquired at different echo times (minimal TE 0.07 ms, maximal TE 9.0 ms), sodium image (6C) of a patient with chronic Achilotendinitis. Sodium image shows increased sodium SNR and thus higher GAG content.

cal requirements of sodium MRI, it is degeneration, such as T2, T2*, T1ρ, of using this parameter as a marker for possible to clinically use sodium MRI for ADC, and magnetization transfer [61– Achilles tendinopathy [69]. An SNR diagnostic purposes in the Achilles 65]. It was shown spectroscopically that increase between 3 and 7T1 was tendon in vivo. T2* decay in the Achilles tendon demon- observed in this study as well. T2* was strates a multi-component nature [66]. acquired by fitting bi-component expo- 3.2) Quantitative analysis of T2* The MR signals from the second, third, nential functions, and both short (T2*s) in the Achilles tendon at 7 Tesla1 and fourth components are difficult to and long (T2*l) components were evalu- With the recent hardware and software acquire on conventional clinical MR sys- ated. With regard to the comparison developments, MRI is still more often tems, since these have a relatively small of patients and volunteers at 7T1, the used clinically for the evaluation of the component ratio compared to the first bulk T2*s was significantly higher in Achilles tendon. Modern MRI methods component. The first component has patients, although the bulk T2*l differ- provide direct access to signal from fast- the highest ratio, but it is in the submilli- ence was not statistically significant. relaxing tissues, such as tendons. Since second range and can therefore hardly This study suggests that ultra-short bi- the relaxation times in the AT are on be acquired with conventional echo component T2* measurements in the the order of 1 ms, a very short echo time times (~5 ms). Recent sequence devel- human Achilles tendon in vivo are feasi- (TE) must be used to acquire signal from opment allows the acquisition of a sig- ble using a 3D-UTE sequence, with high the tendon. These methods were suc- nal from these tissues. The most impor- accuracy at 7T1 (Fig. 6B). Although it is cessfully used to detect partial or total tant sequences used in MR imaging of difficult to separate the water and colla- tendon rupture or even the degenerative highly organized tissues are two- and gen fibers in the Achilles tendon; T2* processes in the tendon tissue [56]. three-dimensional ultra-short echo time reflects the overall condition of AT as an Pathological and consecutive repair pro- sequences (2D- and 3D-UTE), variable interplay between the amount of water cesses may lead to collagen fiber altera- echo time sequences (VTE), and an and the organization of collagen macro- tions, and thus, relaxation properties acquisition-weighted stack of spirals molecules. The observed differences may be changed [60]. This results in sig- (AWSOS) sequence. Du and colleagues between T2*s in healthy and pathologi- nal alterations and changes in relaxation observed a T2* of 0.78 ± 0.07 ms at 3T, cal tendons suggest that advanced properties of the bound and free water using a 2D-UTE sequence [67]. Regional quantitative imaging of the human molecules. Quantitative analysis of the dependencies of T2* within the Achilles Achilles tendon may provide additional relaxation and/or diffusion constants of tendon were described by Robson et al. information to standard clinical imaging the AT, such as T1, T2*, T1ρ, or ADC, [61]. Ultra high field MRI provides a in reasonably short MR data acquisition may provide additional information about substantial increase in the SNR ratio of times. Quantitative analysis of T2* may the tendon condition. On conventional fast-relaxing tissue types [68]. In a recent be a promising marker for the diagnosis high-field MR systems, several parameters study, we used a 3D-UTE sequence at of early tendinopathy in the Achilles ten- have been investigated in the past as 7T1 to estimate T2* in tendons in order don [69]. prospective markers for Achilles tendon to investigate the potential feasibility

94 MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world Tissue function Technology

In conclusion, ultra high field MRI at 7T1 5 Borthakur A, Shapiro EM, Beers J, et al (2000) 18 Zbyn S, Stelzeneder D, Welsch GH, et al Evalua- offers the advantage of higher resolu- Sensitivity of MRI to proteoglycan depletion in tion of native hyaline cartilage and repair tissue cartilage: comparison of sodium and proton after two cartilage repair surgery techniques tion imaging for morphological informa- MRI. Osteoarthritis and Cartilage 8, 288-293. with (23)Na MR imaging at 7 T: initial experi- tion, and improves spatial resolution in 6 Ling W, Regatte RR, Navon G, et al (2008b) ence. Osteoarthritis Cartilage 20, 837-45. T2-mapping, which enables an evaluation Assessment of glycosaminoglycan concentration 19 Krusche-Mandl I, Schmitt B, Zak L, et al of the zonal variation of even thin carti- in vivo by chemical exchange-dependent satura- Long-term results 8 years after autologous lage layers and significantly improves tion transfer (gagCEST). Proceedings of the osteo-chondral transplantation: 7 T gagCEST National Academy of Sciences of the United and sodium magnetic resonance imaging sensitivity for other nuclei, such as 23Na. States of America 105, 2266-2270. with morphological and clinical correlation. This plays an important role in MSK 7 Burstein D, Bashir A, Gray ML (2000) MRI Osteoarthritis Cartilage 20, 357-63. imaging, since sodium correlates directly techniques in early stages of cartilage disease. 20 Guivel-Scharen V, Sinnwell T, Wolff SD, et al with GAG content. In addition, the GAG- Investigative Radiology 35, 622-638. (1998) Detection of proton chemical exchange specific CEST technique also benefits 8 Watanabe A, Wada Y, Obata T, et al (2006) between metabolites and water in biological Delayed gadolinium-enhanced MR to determine tissues. Journal of Magnetic Resonance 133, from the ultra high field. This will provide glycosaminoglycan concentration in reparative 36-45 DOI: S1090-7807(98)91440-9 [pii] new insights into the normal and abnor- cartilage after autologous chondrocyte implanta- 10.1006/jmre.1998.1440.˙ mal physiology of musculoskeletal tis- tion: Preliminary results. Radiology 239, 201-208. 21 Forsen S, Hoffman RA (1963) Study of Moder- sues, and will, therefore, provide new in 9 Bashir A, Gray ML, Boutin RD, et al (1997) ately Rapid Chemical Exchange Reactions by vivo clinical applications. Glycosaminoglycan in articular cartilage: Means of Nuclear Magnetic Double Resonance. In vivo assessment with delayed Gd(DTPA)(2-)- Journal of Chemical Physics 39, 2892. enhanced MR imaging. Radiology 205, 551-558. 22 Ward KM, Aletras AH, Balaban RS (2000) A new Acknowledgment 10 Tiderius CJ, Olsson LE, Leander P, et al (2003) class of contrast agents for MRI based on proton The authors would like to thank the Delayed gadolinium-enhanced MRI of cartilage chemical exchange dependent saturation Editorial office of European Radiology, (dGEMRIC) in early knee osteoarthritis. transfer (CEST). Journal of Magnetic Resonance Magnetic Resonance in Medicine 49, 488-492. Osteoarthritis and Cartilage and Radi- 143, 79-87 DOI: 10.1006/jmre.1999.1956 11 Williams A, Gillis A, McKenzie C, et al (2004) S1090-7807(99)91956-0 [pii].˙ ology for permissions to use figures Glycosaminoglycan distribution in cartilage as 23 Ward KM, Balaban RS (2000) Determination of from previously published articles. determined by delayed gadolinium-enhanced pH using water protons and chemical exchange MRI of cartilage (dGEMRIC): Potential clinical dependent saturation transfer (CEST). Magnetic * A licensed physician may choose to use FDA- applications. American Journal of Roentgen- Resonance in Medicine 44, 799-802 DOI: approved contrast agents in conjunction with an MRI ology 182, 167-172. 10.1002/1522-2594(200011)44:5<799::AID- exam, based on his/her medical opinion and discre- 12 Trattnig S, Marlovits S, Gebetsroither S, et al MRM18>3.0.CO;2-S [pii].˙ tion and in accordance with the instructions for use (2007) Three-dimensional delayed gadolinium- 24 Zhou JY, van Zijl PCM (2006) Chemical exchange and indications for use supplied by the pharmaceuti- cal manufacturer for the contrast agents. enhanced MRI of cartilage (dGEMRIC) for in vivo saturation transfer imaging and spectroscopy. evaluation of reparative cartilage after matrix- Progress in Nuclear Magnetic Resonance 1 Works in Progress in the USA. The information about associated autologous chondrocyte transplan- Spectroscopy 48, 109-136 DOI: DOI 10.1016/j. this product is preliminary. The product is under tation at 3.0T: Preliminary results. Journal of pnmrs.2006.01.001.˙ development and is not commercially available in the Magnetic Resonance Imaging 26, 974-982. 25 Ling W, Regatte RR, Schweitzer ME, et al (2008) U.S. and its future availability cannot be ensured. 13 Trattnig S, Mamisch TC, Pinker K, et al (2008) Characterization of bovine patellar cartilage The 7T system is an investigational device. Differentiating normal hyaline cartilage from by NMR. Nmr in Biomedicine 21, 289-295 DOI: Limited by U.S. federal law to investigational use. This research system is not cleared, approved or post-surgical repair tissue using fast gradient Doi 10.1002/Nbm.1193.˙ licensed in any jurisdiction for patient examinations. echo imaging in delayed gadolinium-enhanced 26 Ling W, Regatte RR, Navon G, et al (2008) This research system is not labelled according to MRI (dGEMRIC) at 3 Tesla. European Radiology Assessment of glycosaminoglycan concentration applicable medical device law and therefore may 18, 1251-1259. in vivo by chemical exchange-dependent only be used for volunteer or patient examinations 14 Kim YJ, Jaramillo D, Millis MB, et al (2003) saturation transfer (gagCEST). Proc Natl Acad Sci in the context of clinical studies according to Assessment of early osteoarthritis in hip U S A 105, 2266-70 DOI: 0707666105 [pii] applicable law. dysplasia with delayed gadolinium-enhanced 10.1073/pnas.0707666105.˙ References magnetic resonance imaging of cartilage. 27 Schmitt B, Bock M, Stieltjes B, et al. A new, 1 Buckwalter JA, Mankin HJ (1998) Articular carti- Journal of Bone and Joint Surgery-American 3D GRE based CEST imaging method for clinical lage: Degeneration and osteoarthritis, repair, Volume 85A, 1987-1992. application and verification with gagCEST in regeneration, and transplantation. Instructional 15 Vaga S, Raimondi MT, Caiani EG, et al (2008) articular cartilage. in Proceedings 18th Scientific Course Lectures, Vol 47 – 1998 47, 487-504. Quantitative assessment of intervertebral disc Meeting, International Society for Magnetic 2 Poole AR, Kojima T, Yasuda T, et al (2001) Compo- glycosaminoglycan distribution by gadolinium- Resonance in Medicine. 2010. Stockholm. sition and structure of articular cartilage – A enhanced MRI in orthopedic patients. 28 Vinogradov E, Ivanishev A, Grant AK, et al. CEST template for tissue repair. Clinical Orthopaedics Magnetic Resonance in Medicine 59, 85-95. and Sodium Imaging of Glycosaminoglycans and Related Research, S26-S33. 16 Williams A, Shetty SK, Burstein D, et al (2008) in-vivo in the 3T: Preliminary Results. in Pro- 3 Burstein D, Velyvis J, Scott KT, et al (2001) Proto- Delayed gadolinium enhanced MRI of cartilage ceedings 18th Scientific Meeting, International col issues for delayed Gd(DTPA)(2-)-enhanced (dGEMRIC) of the first carpometacarpal (1CMC) Society for Magnetic Resonance in Medicine. MRI: (dGEMRIC) for clinical evaluation of articu- joint: a feasibility study. Osteoarthritis and 2010. Stockholm. lar cartilage. Magnetic Resonance in Medicine Cartilage 16, 530-532. 29 Fenty M, Kassey V, Kogan F, et al. Feasibility of 45, 36-41. 17 Trattnig S, Welsch GH, Juras V, et al (2010) (23) CEST imaging on the guinea pig stifle at 9.4 T. 4 Bashir A, Gray ML, Burstein D (1996) Gd-DTPA(2-) Na MR Imaging at 7 T after Knee Matrix-associ- in Proceedings 19th Scientific Meeting, Inter- as a measure of cartilage degradation. ated Autologous Chondrocyte Transplantation: national Society for Magnetic Resonance in Magnetic Resonance in Med. 36, 665-673. Preliminary Results. Radiology 257, 175-184. Medicine. 2011. Montreal.

MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world 95 Technology Tissue function

30 Vinogradov E, Lenkinski RE. Detection of Glycos- 44 White LM, Sussman MS, Hurtig M, et al (2006) human patellar tendinopathy. Clin J Sport Med 17, aminoglycans using Positive CEST approach. Cartilage T2 assessment: Differentiation of 129-34. in Proceedings 18th Scientific Meeting, Interna- normal hyaline cartilage and reparative tissue 59 Juras V, Zbyn S, Pressl C, et al (2012) Sodium tional Society for Magnetic Resonance in after arthroscopic cartilage repair in equine MR Imaging of Achilles Tendinopathy at 7 T: Medicine. 2010. Stockholm. subjects. Radiology 241, 407-414. Preliminary Results. Radiology 262, 199-205. 31 Varma G, Alsop DC, Lenkinski RE, et al. Optimiza- 45 Trattnig S, Mamisch TC, Welsch GH, et al (2007) 60 Gelberman RH, Manske PR, Vande Berg JS, et al tion of pulsed-gagCEST at 3.0T. in Proceedings Quantitative T-2 mapping of matrix-associated (1984) Flexor tendon repair in vitro: a compara- 19th Scientific Meeting, International Society for autologous chondrocyte transplantcation tive histologic study of the rabbit, chicken, dog, Magnetic Resonance in Medicine. 2011. at 3 Tesla – An in vivo cross-sectional study. and monkey. J Orthop Res 2, 39-48. Montreal. Investigative Radiology 42, 442-448. 61 Robson MD, Benjamin M, Gishen P, et al (2004) 32 Varma G, Lenkinski RE, Vinogradov E. Keyhole 46 Welsch GH, Mamisch TC, Domayer SE, et al Magnetic resonance imaging of the Achilles ten- Chemical Exchange Saturation Transfer. in Pro- (2008) Cartilage T2 assessment at 3-T MR imag- don using ultrashort TE (UTE) pulse sequences. ceedings 19th Scientific Meeting, International ing: In vivo differentiation of normal hyaline Clin Radiol 59, 727-35. Society for Magnetic Resonance in Medicine. cartilage from reparative tissue after two 62 Hodgson RJ, Evans R, Wright P, et al (2011) 2011. Montreal. cartilage repair procedures – Initial experience. Quantitative magnetization transfer ultrashort 33 Schmitt B, Zbyn S, Stelzeneder D, et al (2011) Radiology 247, 154-161. echo time imaging of the Achilles tendon. Cartilage quality assessment by using glycos- 47 Domayer SE, Apprich S, Stelzeneder D, et al Magn Reson Med 65, 1372-6. aminoglycan chemical exchange saturation Cartilage repair of the ankle: first results of T2 63 Du J, Carl M, Diaz E, et al (2010) Ultrashort TE transfer and (23)Na MR imaging at 7 T. Radiolo- mapping at 7.0 T after microfracture and matrix T1rho (UTE T1rho) imaging of the Achilles gy 260, 257-64 DOI: 10.1148/radiol.11101841.˙ associated autologous cartilage transplanta- tendon and meniscus. Magn Reson Med 64, 34 Ling W, Saar G, Regatte R, et al. Assessing the tion. Osteoarthritis Cartilage 20, 829-36. 834-42. Inververtebral Disc via gagCEST. in Proceedings 48 David-Vaudey E, Ghosh S, Ries M, et al (2004) 64 Fechete R, Demco DE, Eliav U, et al (2005) 17th Scientific Meeting, International Society for T-2 relaxation time measurements in osteo- Self-diffusion anisotropy of water in sheep Magnetic Resonance in Medicine. 2009. Honolulu. arthritis. Magnetic Resonance Imaging 22, Achilles tendon. NMR Biomed 18, 577-86. 35 Kim M, Chan Q, Anthony MP, et al (2011) As- 673-682. 65 Henkelman RM, Stanisz GJ, Kim JK, et al (1994) sessment of glycosaminoglycan distribution in 49 Dunn TC, Lu Y, Jin H, et al (2004) T2 relaxation Anisotropy of NMR properties of tissues. human lumbar intervertebral discs using chemi- time of cartilage at MR imaging: Comparison Magn Reson Med 32, 592-601. cal exchange saturation transfer at 3 T: feasi- with severity of knee osteoarthritis. Radiology 66 Peto S, Gillis P (1990) Fiber-to-field angle depen- bility and initial experience. NMR in Biomedicine 232, 592-598. dence of proton nuclear magnetic relaxation in 24, 1137-44 DOI: 10.1002/nbm.1671.˙ 50 Burstein D, Gray ML (2006) Is MRI fulfilling its collagen. Magn Reson Imaging 8, 705-12. 36 Haneder S, Apprich S, Schmitt B, et al (2012) promise for molecular imaging of cartilage 67 Du J, Chiang AJT, Chung CB, et al (2010) Orien- Glycosaminoglycan content in intervertebral in arthritis? Osteoarthritis and Cartilage 14, tational analysis of the Achilles tendon and discs using chemical exchange saturation trans- 1087-1090. enthesis using an ultrashort echo time spectro- fer at 3T MRI: preliminary results in patients 51 Stelzeneder D, Welsch GH, Kovacs BK, et al scopic imaging sequence. Magnetic Resonance with low back pain. European Radiology Accept- Quantitative T2 evaluation at 3.0T compared to Imaging 28, 178-184. ed for Publication. morphological grading of the lumbar interverte- 68 Regatte RR, Schweitzer ME (2007) Ultra-high- 37 Mosher TJ, Dardzinski BJ (2004) Cartilage MRI T2 bral disc: a standardized evaluation approach in field MRI of the musculoskeletal system at 7.0T. relaxation time mapping: Overview and appli- patients with low back pain. Eur J Radiol 81, J Magn Reson Imaging 25, 262-9. cations. Seminars in Musculoskeletal Radiology 324-30. 69 Juras V, Zbyn S, Pressl C, et al (2012) Regional 8, 355-368. 52 Stelzeneder D, Messner A, Vlychou M, et al Variations of T2* in Healthy and Pathologic 38 Welsch GH, Mamisch TC, Hughes T, et al (2008) Quantitative in vivo MRI evaluation of lumbar Achilles Tendon In Vivo at 7 Tesla: Preliminary In vivo biochemical 7.0 Tesla magnetic reso- facet joints and intervertebral discs using axial Results. Magnetic resonance in Medicine In nance - Preliminary results of dGEMRIC, zonal T2 mapping. Eur Radiol 21, 2388-95. Press. T2, and T2* mapping of articular cartilage. 53 Trattnig S, Stelzeneder D, Goed S, et al Lumbar Investigative Radiology 43, 619-626. intervertebral disc abnormalities: comparison 39 Welsch GH, Trattnig S, Scheffler K, et al (2008) of quantitative T2 mapping with conventional Contact Magnetization transfer contrast and T2 map- MR at 3.0 T. Eur Radiol 20, 2715-22. Prof.Siegfried Trattnig, M.D. ping in the evaluation of cartilage repair tissue 54 Welsch GH, Trattnig S, Paternostro-Sluga T, et al MR Center – High field MR with 3T MRI. Journal of Magnetic Resonance Parametric T2 and T2* mapping techniques Department of Radiology Imaging 28, 979-986. to visualize intervertebral disc degeneration in Medical University of Vienna/Vienna 40 Goodwin DW, Zhu HQ, Dunn JF (2000) In vitro patients with low back pain: initial results on General Hospital MR imaging of hyaline cartilage: Correlation the clinical use of 3.0 Tesla MRI. Skeletal Radiol Lazarettgasse 14 with scanning electron microscopy. American 40, 543-51. A-1090 Vienna Journal of Roentgenology 174, 405-409. 55 Noebauer-Huhmann IM, Juras V, Pfirrmann C, Austria 41 Goodwin DW, Wadghiri YZ, Dunn JF (1998) et al (in press) Sodium Imaging of the Lumbar Phone: +43 1 40 400 6460 Micro-imaging of articular cartilage: T2, proton Intervertebral Disc at 7T: Correlation with T2 Fax.: +43 1 40 400 6475 density, and the magic angle effect. Academic Mapping and Modified Pfirrmann Score at 3T. [email protected] Radiology 5, 790-798. Preliminary results. Radiology. 42 Smith HE, Mosher TJ, Dardzinski BJ, et al (2001) 56 Schweitzer ME, Karasick D (2000) MR imaging Spatial variation in cartilage T2 of the knee. of disorders of the Achilles tendon. AJR Am J Journal of Magnetic Resonance Imaging 14, Roentgenol 175, 613-25. 50-55. 57 Samiric T, Parkinson J, Ilic MZ, et al (2009) 43 Watrin-Pinzano A, Ruaud JP, Cheli Y, et al (2004) Changes in the composition of the extracellular Evaluation of cartilage repair tissue after bio- matrix in patellar tendinopathy. Matrix Biol 28, material implantation in rat patella by using T2 230-6. mapping. Magnetic Resonance Materials in 58 Fu SC, Chan KM, Rolf CG (2007) Increased Physics Biology and Medicine 17, 219-228. deposition of sulfated glycosaminoglycans in

96 MAGNETOM Flash · 2/2012 · www.siemens.com/magnetom-world