Objectives The purpose was to implement a fast 3D glycosaminoglycan Chemical Exchange Saturation Transfer (gagCEST) sequence at 7 T, test stability and reproducibility in cartilage in the knee in healthy volunteers, and evaluate clinical applicability in cartilage repair patients. Methods Experiments were carried out on a 7-Tscanner using a volume transmit coil and a 32-channel receiver wrap-around knee coil. The 3D gagCEST measurement had an acquisition time of 7 min. Signal stability and reproducibility of the GAG effect were assessed in eight healthy volunteers. Clinical applicability of the method was demonstrated in five patients before cartilage repair surgery. Results Coefficient of variation of the gagCESTsignal was 1.9%. The reproducibility of the GAG effect measurements was good in the medial condyle (ICC = 0.87) and excellent in the lateral condyle (ICC = 0.97). GAG effect measurements in healthy cartilage ranged from 2.6%-12.4% compared with 1.3%-5.1% in damaged cartilage. Difference in GAG measurement between healthy cartilage and damaged cartilage was significant (p <0.05). Conclusions A fast 3D gagCEST sequence was applied at 7 T for use in cartilage in the knee, acquired within a clinically feasible scan time of 7 min. We demonstrated that the method has high stability, reproducibility and clinical applicability. Key Points � gagCEST measurements are stable and reproducible � A non-invasive GAG measurement with gagCEST can be acquired in 7 min � gagCEST is able to discriminate between healthy and damaged cartilage . . . . Keywords Cartilage Glycosaminoglycans Knee Magnetic resonance imaging Osteoarthritis Abbreviations CV Coefficient of variation GAG Glycosaminoglycans gagCEST Glycosaminoglycan chemical exchange Electronic supplementary material The online version of this article saturation transfer (https://doi.org/10.1007/s00330-017-5277-y) contains supplementary material, which is available to authorized users. ICC Intraclass correlation coefficient ICRS International Cartilage Repair Society * Sander Brinkhof OA Osteoarthritis email@example.com Department of Radiology, UMC Utrecht, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands Introduction Department of Orthopaedics, UMC Utrecht, Utrecht, The Netherlands With the ageing of our society, the prevalence of degen- erative diseases, such as osteoarthritis (OA), has increased MIRA institute for Biomedical Technology and Technical Medicine, University of Twente, Enschede, The Netherlands . OA is a degenerative whole-joint disease that affects the articular cartilage. Since cartilage tissue has a limited Department of Orthopaedics, Mayo Clinic, Rochester, Minnesota, USA ability to regenerate, early identification of cartilage Eur Radiol (2018) 28:2874–2881 2875 damage improves chances of successful treatment and Methods prognosis [2, 3]. Early-stage OA and early-stage cartilage damage in general are subject to delicate changes in bio- Numerical simulations chemical composition, i.e. a loss of glycosaminoglycans (GAG), on the surface of the cartilage ). The measure- The 3D gagCEST sequence implemented in this work is a ment of reductions in GAG is a promising approach for pseudo-steady-state pulsed 3D gradient echo CEST sequence the diagnosis and treatment monitoring of early stage OA. recently developed in our group . The sequence was MRI is an excellent modality to visualise cartilage mor- optimised through the Bloch-McConnell simulations . phology; however, a standard anatomical MRI is not suf- The following sequence parameters were investigated: the ficient to visualise early stage OA. number of saturation pulses, transmit field (B ) amplitude Quantitative MRI has been used increasingly over the and duty cycle. All other sequence parameters were fixed to past years to quantify GAG content in vivo in OA [5, 6]. the values that were eventually used for data acquisition . Initial studies focused on the application of delayed gad- Gradient and RF spoiling was simulated by setting the trans- olinium enhanced MRI of cartilage (dGEMRIC) [7, 8]to verse magnetisation components to zero. Two-pool (free water measure the GAG content indirectly. This technique is and GAG) Bloch-McConnell equations were solved numeri- based on the distribution of negatively charged ions of a cally  assuming the parameters in Table 1. GAG effect size gadolinium-based contrast agent in cartilage, which is in- was quantified by the pool difference method: versely proportional to the GAG content . Another GAG ¼S0ðÞ :9ppm; M ¼ 0 −S0ðÞ :9ppm; M ¼ 1ð1Þ A A technique to assess GAG content, without the use of a contrast agent, is T1ρ mapping. In this technique, spinlock pulses of different durations are applied to assess where S(Δω,M ) is the simulated signal in the z-spectrum the T1ρ relaxation time, which is lower in water associ- at Δω = 0.9 ppm, and M is the simulated amplitude of the ated with large macromolecules, such as GAG, as com- GAG compartment. The saturation parameters were chosen to pared to free water. However, it is still disputed whether achieve an optimal GAG effect size, but with as low as pos- T1ρ is directly correlated with the GAG content . sible acquisition time and within the limitations of the RF Alternatively, sodium ( Na) MRI measures the sodium amplifier duty cycle. ions in the interstitial fluid in the cartilage [11–13]. Sodium counterbalances the negative charge of the sul- phate and carboxyl groups of GAG. A lower fixed charge MRI data acquisition density (FCD), and thus a loss of GAG, causes a loss of sodium ions from cartilage . However, for Na MRI Eight healthy volunteers without a history of knee pain or dedicated MRI coils are required, which are highly exper- trauma and five patients undergoing arthroscopy for repair imental and not widely available. of a focal knee cartilage defect were included in this study In contrast to these assessments of GAG, chemical ex- (approved by the medical ethics committee). Patients were change saturation transfer (CEST) directly quantifies the selected within our specialised knee clinic of the University GAG content based on the chemical exchange of its labile Medical Centre Utrecht. Patients undergoing an arthroscopy hydroxyl (-OH) protons with the bulk water [14–16]. for cartilage repair on the femoral condyle were included for a These exchangeable protons resonate at a different fre- pre-operative MRI. Exclusion criteria were as follows: history quency compared with bulk water protons and are satu- of cartilage repair, history of cruciate ligament tears or repair rated via selective radiofrequency (RF) irradiation. and/or trochlear/patellar cartilage damage. Informed consent Because of the exchange, the saturation is transferred to was acquired from all the subjects after explaining the study the bulk water pool, which ultimately results in large con- procedures. MRI experiments were carried out on a 7.0-T trast enhancement factors [14, 17–19]. The quantification whole-body scanner (Achieva; Philips Healthcare, Best, of GAG in articular cartilage with the use of CEST, i.e. The Netherlands), using an in-house developed and built vol- gagCEST, has a high potential for the examination of ume transmit coil and a dedicated 32-channel receiver wrap- cartilage degeneration and hence diagnosis of early stage around knee coil (MR Coils BV, Zaltbommel, OA. However, in previous applications gagCEST data The Netherlands). were mostly acquired in 2D because 3D sequences are The 3D gagCEST sequence included a pre-saturation mod- very time consuming [16, 20, 21]. The purpose of this ule consisting of a train (n = 20) of sinc-shaped pulses (B =2 study was to implement a fast 3D gagCEST sequence at μT, pulse length = 25 ms, duty cycle = 70%, based on simu- 7 T with a clinically feasible scan time and to evaluate the lations). The readout parameters were as follows: five-shot stability, reproducibility and clinical applicability of this turbo field echo (TFE), TFE factor of 370, SENSE factor of method in articular cartilage in the knee. 2, TR/TE/FA = 2.75 ms/1.4 ms/5 degrees, field of view = 140 2876 Eur Radiol (2018) 28:2874–2881 Table 1. Overview of interest (ROI): the medial condyle, the trochlear groove and Water GAG parameters for Bloch- the lateral condyle. These regions were also used for the quan- McConnell equation tification of the reproducibility of the GAG effect. T (s) 1.2 1* simulations 1 CEST spectra were B corrected using WASSR and T (ms) 40 10 0 were normalised using the high off-resonant gagCEST images Δw(ppm) 0 0.9 at ± 100 kHz. The B corrected and normalised spectra were M (%) - 0.27 0 fitted pixel-wise using a sum of three Lorentzians to account R (Hz) - 1000 for the GAG, water and magnetisation transfer pools . *Fixed in simulation The GAG effect is expected to be around 270 Hz, for which we chose three offsets to represent that point (200, 275, 350 3 3 ×150 ×135 mm ,resolution = 1×1 × 3mm , inter-shot T1 Hz). The amplitude of the GAG pool is averaged over these recovery time = 2 s, k-space centre-weighted acquisition, two three offsets to avoid outliers in the fit. dummy scans, and total acquisition time = 6 min 59 s. A 3D segmentation of the cartilage was used to evaluate the The gagCEST images were acquired at 17 saturation off- GAG effect in the patient group. Both weight-bearing con- sets ranging from -900 Hz to 900 Hz (-3 ppm to 3 ppm), i.e. - dyles were divided into four regions (medial/lateral/superior/ 900, -600, -425, -350, -275, -200, -75, -25, 0, 25, 75, 200, 275, inferior) and the regions where a defect was present, according 350, 425, 600 and 900 Hz. In addition, gagCEST images were to the surgeon’s notes, were used for the analyses. These de- acquired at offsets of -100 kHz and +100 kHz to normalise the fect regions also include the defect rim, which is of great CEST spectrum. The expected resonance frequency of the interest for treatment planning. These defect regions were hydroxyl side groups of GAG is 0.9 ppm, which is 270 Hz compared with the same regions on the healthy contralateral at 7 T . condyle. A detailed explanation of the analysis workflow and Signal stability tests were carried out in five healthy volun- the described regions of interest are shown in the Appendix. teers (mean age: 26 years, age range: 21 to 35 years, two males and three females). Each subject was scanned twice and dur- Statistical analysis ing each session 19 gagCEST images were acquired at a single saturation offset of 0.9 ppm (270 Hz). Nineteen acquisitions Stability of the signal (i.e. the value of the CEST spectrum at were chosen to represent the same scan duration as for the 275 Hz, where the GAG effect is expected) is expressed with gagCEST experiment with 19 different offsets. the coefficient of variation (CV), which was calculated by The reproducibility of the measurement of the GAG effect dividing the standard deviation by the mean of the signal. was assessed in eight healthy volunteers (mean age: 24 years, The coefficient of variation was calculated in the three afore- age range: 21 to 30 years, three males and five females). Each mentioned ROIs and was calculated for both acquired stability subject was scanned twice within the same scan session. assessments. The reproducibility of the GAG effect size was The clinical applicability of the method was demonstrated assessed by means of Bland-Altman plots and correlation by comparing the GAG effect size in healthy cartilage versus plots with corresponding intraclass correlation coefficients damaged cartilage in five patients before cartilage repair (age (ICC), i.e. the degree of absolute agreement among measure- range: 21 to 41 years, all male, no significant/obvious varus or ments (criterion-referenced reliability). To evaluate differ- valgus leg axis). These patients were scanned up to 24 h prior ences between healthy cartilage and damaged cartilage in the to surgery. During surgery, cartilage defects were graded with patients, a Wilcoxon signed rank test was applied. the International Cartilage Repair Society (ICRS) grading scale (grade 0 to 4, 0 = no damage; 4 = full thickness cartilage defect) . The ICRS grade was graded in the femoral car- Results tilage because we solely included patients with defects in the cartilage of the femoral condyles. The cartilage on the healthy Simulation data condyle was graded with ICRS grade 0. Figure 1 shows results of Bloch-McConnell simulations for Image analysis the applied sequence. Twenty pulses were chosen as the num- ber in the train, close to the maximum effect size for GAG but Data analysis was performed in MATLAB (R2016b, the still within a clinically feasible acquisition time. A duty cycle MathWorks, Natick, MA, USA) with in-house developed pro- of 70% was used to stay within RF amplifier duty cycle limits; cessing scripts. The signal stability measurements were nor- 2 μT was chosen to approach the optimal effect size within the malised with respect to the signal intensity of the first mea- desired acquisition time. The combination of both leads to a surement. The signal stability was quantified from the aver- maximum effect size of roughly 8 percent, which was in line aged signal over all pixels in each of the three regions of with the chosen number of pulses in the pre-pulse train. Eur Radiol (2018) 28:2874–2881 2877 Fig. 1. (A) The simulated GAG effect size (%) as a function of the number of pulses in the CEST pre-pulse. (B) The simulated 3D plot of GAG effect size (%) as a function of the RF duty cycle (of the CEST pre-pulse) and B field amplitude 1+ Stability and reproducibility show that there is no proportional bias between the two measurements. The coefficients of variation of the signal stability assessments are reported in Table 2. The average CV in the medial condyle Clinical applicability was 2.00%, the average CV in the lateral condyle was 2.25%, and the average CV in the trochlea was 1.40%. A 3D segmented model of the knee cartilage of a patient is Figure 2 shows an example of a fitted CESTspectrum, with shown in Fig. 5. A difference in GAG effect in this patient was the GAG, water and MT pools visualised in purple, light blue observed between the medial and the lateral sides. This spe- and dark blue, respectively. The GAG effect can be observed cific patient had an ICRS grade IV defect in the medial con- at the expected offset around 0.9 ppm. dyle, which corresponds with the gagCEST findings. An The correlation plots in Fig. 3 show strong reproducibility artroscopic view of this patient and the corresponding 3D in the lateral condyle (ICC = 0.97, p < 0.01) and the medial gagCEST map is shown in Fig. 6. The ICRS grades and condyle (ICC = 0.87, p < 0.01). The ICC for the trochlear GAG effects of all patients are summarised in Table 3.The groove was weak (0.064, p = 0.43). Bland-Altman plots of ICRS grade ranged from 3 to 4 (> 50% thickness defects to the medial condyle and the lateral condyle are shown in Fig. 4. full thickness cartilage defects). The GAG effect of healthy Bland-Altman analysis was not carried out in the trochlear cartilage ranged from 2.6% to 12.4% and the GAG effect of groove because of the poor ICC. The Bland-Altman analyses damaged cartilage ranged from 1.3 to 5.1%. The GAG effect Table 2. Stability assessments of Subject Age Gender Scan Medial CV (%) Trochlea CV (%) Lateral CV (%) GAG effect at 0.9 ppm in healthy volunteers 1 21 F 1 1.61 0.88 1.23 2 1.67 2.07 1.11 2 29 M 1 0.89 0.52 2.89 2 1.25 0.52 0.73 3 35 M 1 3.2 1.96 1.74 2 5.44 2.96 3.34 4 21 F 1 1.64 1.55 6.57 2 1.54 0.57 0.88 5 25 F 1 1.08 1.54 1.49 2 1.67 1.38 3.54 Mean coefficient of variation: 2.00 1.40 2.25 2878 Eur Radiol (2018) 28:2874–2881 Fig. 2. An example of the CEST spectrum and its three-pool Lorentzian decomposition. The black line shows the multi- Lorentzian fit of the three pools; acquired data are represented with black dots in damaged cartilage was significantly different (p < 0.05) a scan time of 19 min, albeit with better resolution compared from that in healthy cartilage. with our study . Note that a higher resolution reduces the signal to noise and is more prone to artefacts related to motion of the knee. All sequences published used the same or a com- Discussion parable number of offsets and comparable field of view. We chose to implement an in-plane resolution of 1 × 1 mm to This study presents a fast 3D gagCEST sequence with full minimise partial volume effects in the directions with the most curvature of the cartilage. This came with the drawback that cartilage coverage, which can quantify the GAG effect in healthy volunteers and patients. The data were acquired within the slice thickness needed to be 3 mm to achieve a sufficient SNR. Several other 3D gagCEST studies also implemented a seven minutes and shown to be stable and reproducible. Moreover, the method could differentiate healthy from dam- comparable slice thickness of 3 mm [15, 16] or 5 mm . An isotropic voxel size would be more ideal for 3D visualisation aged cartilage in patients before their cartilage repair surgery. The main goal of this study was to present a fast 3D purposes, but this can only be achieved with a lower in-plane resolution or with much longer scan times. gagCEST sequence. The acquisition time for the gagCEST This sequence was optimised using Bloch-McConnell sim- sequence used in this study was 6 min 59 s, because a pseudo-steady state sequence was applied with an optimised ulations. Our goal was to minimise the scan time, which could lead to a sub-optimal CEST effect size. The number of pulses number of saturation pulses. Other 3D sequences were pub- lished with scan times ranging from 11 min to almost in the pre-pulse train could be increased to 60 for optimal effect size, as shown in Fig. 1A. However, this would increase 15 min . The latest study of the group of Trattnig reported Fig. 3. The correlation graphs of three assessed locations (medial lateral condyle: 0.97 (p < 0.001). Measurement 1 refers to the condyle, lateral condyle and trochlear groove). ICC medial condyle: amplitude of the GAG fit in the first measurement; measurement 2 0.87 (p = 0.0049), ICC trochlear groove: 0.063 (p = 0.43) and ICC refers to the amplitude of the GAG fit in the second measurement Eur Radiol (2018) 28:2874–2881 2879 Fig. 4. Bland-Altman plots of GAG effects in the medial and lateral condyle Fig. 5. The 3D segmented GAG map of articular cartilage in the knee of a patient with an ICRS grade IV defect on the medial side of the knee the shot time to 5.4 s, which increases the acquisition time per B field amplitude of 2 μT and DC of 70% to obtain the 1+ offset by 8 s. This increase of 20% in effect size (8% to 10%) maximum achievable effect size of 8%. would lead to a 40% increase in total scan time (6:59 to 9:54). We applied a Lorentzian fitting algorithm for quantifi- In our study we did not increase the scan time and selected a cation of the GAG effect. We chose Lorentzian fitting to Fig. 6. Comparison of the arthroscopic view and gagCEST map. Left upper corner shows the arthroscopic view of the knee of patient 1. The defect (red)and corresponding defect rim (orange) are highlighted in the image on the upper right.The lower left shows the gagCEST map of this patient, where the defect is clearly visualised. The same regions are highlighted again, with the defect in red and the defect rim in orange 2880 Eur Radiol (2018) 28:2874–2881 Table 3. Comparison of GAG effect in cartilage repair patients: comparison of cartilage on damaged side of the knee versus cartilage on the healthy side of the knee. Difference between groups is statistically significant (p <0.05) No. Age BMI (kg/m ) ICRS Defect Defect size Defect origin GAG effect GAG effect (years) grade location (cm ) healthy condyle defective condyle 1 38 21.1 4 MFC 3 No trauma, gradual increase of pain 12.0 (5.7 – 21.2) 5.1 (0.1 – 11.8) 2 21 22.5 4 LFC 2 Distortion trauma 12.4 (5.0 – 21.6) 1.3 (0 – 7.5) 3 25 23.0 3 LFC 1.5 Cartilage damage after removal of 9.3 (2.2 – 20.1) 1.8 (0 – 8.8) meniscal lesion 4 41 29.5 4 MFC 4 Distortion trauma 2.6 (0 – 11) 2.5 (0 – 9.7) 5 26 22.9 4 LFC 1.5 Rotational trauma 3.7 (0.2 – 10.8) 1.4 (0 – 7.3) MFC = medial femoral condyle; LFC = lateral femoral condyle; GAG effect is expressed as a median and interquartile range achieve better discrimination between the water peak and volunteers (1.6% to 13.9%), which raises the question metabolite peak, in this case GAG. Because we expect whether every volunteer had completely healthy cartilage. GAG to resonate at 0.9 ppm , whichisonly270 Hz These ranges could indicate that there are underlying fac- upfield from the water peak, Lorentzian fitting was cho- tors that affect the GAG effect, for instance age, gender or sen. Lorentzian fitting also decreases the influence of B BMI, as has been suggested in other studies [31, 32]. Due inhomogeneities . Previous literature used MTR to possible confounding effects of these factors, we chose asymmetry as a method for quantification, which is prone not to compare the gagCEST values of patients with to these B inhomogeneities [15, 16]. We used WASSR to healthy volunteers. correctly centre all CEST spectra as recommended by pre- Detection of the range of GAG effect values could be an vious gagCEST studies . interesting tool for osteoarthritis research, for monitoring of The reproducibility in the lateral and medial femoral disease but also for earlier diagnosis. Therefore, a next step in condyles was very good, which is promising for imple- this research would be an analysis of the GAG effect in pa- mentation in clinical practice. However, one should notice tients with cartilage defects, ranging from small focal defects the poor reproducibility in the trochlear groove. The re- to osteoarthritic knees. This will reveal the value of gagCEST producibility in the trochlea was much lower compared to sequences in clinical practice and the patient characteristics the condyles, which was also shown at 7 T in a study affecting the GAG effect. In conclusion, this study presents from Schreiner and colleagues . The area around the a fast gagCEST sequence that is stable and reproducible and trochlear groove is prone to movement of the patella. We shows clinical value. speculate that this movement could be the cause of the Funding The authors state that this work has not received any funding. poor reproducibility of the CEST spectra and their respec- tive fits. Larger muscles could lead to more muscle Compliance with ethical standards twitches, ultimately leading to movement of the structures attached to the muscle, in this case the patella. The poor Guarantor The scientific guarantor of this publication is Daniël B.F. reproducibility could possibly be explained by this phe- Saris. nomenon. In addition, a 3-mm slice thickness could lead to volume averaging with surrounding tissue, especially in Conflict of interest The authors of this manuscript declare no relation- ships with any companies, whose products or services may be related to tissue with a high curvature such as the trochlea. Another the subject matter of the article. limitation of this study is that measurements were only done on severe defects (ICRS grade III or IV) and healthy Statistics and biometry No complex statistical methods were necessary cartilage (ICRS grade 0). Because of the small population for this paper. and the inclusion criteria for cartilage repair surgery in this study, no other defects were observed and gagCEST Informed consent Written informed consent was obtained from all sub- jects (patients) in this study. values of mild cartilage defects (ICRS grade I-II) are absent. Ethical approval Institutional Review Board approval was obtained. The GAG effect value varied across the included healthy volunteers and patients. The range of GAG effect Methodology values is rather large in patients, healthy cartilage ranging � prospective from 2.6% to 12.4%, compared with 1.3% to 5.1% for � diagnostic or prognostic study � performed at one institution damaged cartilage. A similar range is observed in healthy Eur Radiol (2018) 28:2874–2881 2881 Open Access This article is distributed under the terms of the Creative and sodium magnetic resonance imaging with morphological and Commons Attribution 4.0 International License (http:// clinical correlation. Osteoarthr Cartil 20:357–363 creativecommons.org/licenses/by/4.0/), which permits unrestricted use, 16. Schmitt B, Zbýn S, Stelzeneder D et al (2011) Cartilage quality distribution, and reproduction in any medium, provided you give assessment by using glycosaminoglycan chemical exchange satu- appropriate credit to the original author(s) and the source, provide a link ration transfer and 23 Na MR imaging at 7 T 1. Radiology 260:257– to the Creative Commons license, and indicate if changes were made. 264 17. Ward K, Aletras A, Balaban R (2000) A new class of contrast agents for MRI based on proton chemical exchange dependent saturation transfer (CEST). J Magn Reson 143:79–87 References 18. Vinogradov E, Sherry A, Lenkinski R (2013) CEST: From basic principles to applications, challenges and opportunities. J Magn Reson 229:155–172 1. Madry H, Luyten FP, Facchini A (2012) Biological aspects of early 19. van Zijl P, Yadav N (2011) Chemical exchange saturation transfer osteoarthritis. Knee Surgery, Sport Traumatol Arthrosc 20:407–422 (CEST): What is in a name and what isn’t? Magn Reson Med 65: 2. Fox AJ, Bedi A, Rodeo SA (2009) The basic science of articular 927–948 cartilage: structure, composition, and function. Sports Health 1: 20. Krishnamoorthy G, Nanga RPR, Bagga P et al (2017) High quality 461–468 three-dimensional gagCEST imaging of in vivo human knee carti- 3. Bekkers JEJ, Creemers LB, Dhert WJA, Saris DBF (2010) lage at 7 Tesla. Magn Reson Med 77:1866–1873 Diagnostic modalities for diseased articular cartilage—From defect 21. Schreiner MM, Zbýn S, Schmitt B et al (2016) Reproducibility and to degeneration: A review. Cartilage 1:157–164 regional variations of an improved gagCEST protocol for the 4. Maroudas A, Venn M (1977) Chemical composition and swelling in vivo evaluation of knee cartilage at 7T. Magn Reson Mater of normal and osteoarthrotic femoral head cartilage. II. Swelling. Physics, Biol Med 29:513–521 Ann Rheum Dis 36:399–406 22. Khlebnikov V, Geades N, Klomp DWJ et al (2016) Comparison of 5. Guermazi A, Alizai H, Crema MD et al (2015) Compositional MRI pulsed three-dimensional CEST acquisition schemes at 7 tesla: techniques for evaluation of cartilage degeneration in osteoarthritis. steady state versus pseudosteady state. Magn Reson Med 77: Osteoarthr Cartil 23:1639–1653 2280–2287 6. Trattnig S, ZbýňŠ, Schmitt B et al (2012) Advanced MR methods 23. McConnell H (1958) Reaction rates by nuclear magnetic resonance. at ultra-high field (7 Tesla) for clinical musculoskeletal applications. JChem Phys 28:430–431 Eur Radiol 22:2338–2346 24. Woessner D, Zhang S, Merritt M, Sherry A (2005) Numerical solu- 7. Bashir A, Gray ML, Hartke J, Burstein D (1999) Nondestructive tion of the Bloch equations provides insights into the optimum design imaging of human cartilage glycosaminoglycan concentration by of PARACEST agents for MRI. Magn Reson Med 53:790–799 MRI. Magn Reson Med 41:857–865 25. Brittberg M, Peterson L (1998) Introduction of an articular cartilage 8. Trattnig S, Mlyńarik V, Breitenseher M et al (1999) MRI visualiza- classification. ICRS Newsl 5:8 tion of proteoglycan depletion in articular cartilage via intravenous 26. Kim M, Gillen J, Landman BA et al (2009) Water saturation shift administration of Gd-DTPA. Magn Reson Imaging 17:577–583 referencing (WASSR) for chemical exchange saturation transfer 9. Bekkers JEJ, Bartels LW, Benink RJ et al (2013) Delayed gadolin- (CEST) experiments. Magn Reson Med 61:1441–1450 ium enhanced MRI of cartilage (dGEMRIC) can be effectively 27. Zaiss M, Schmitt B, Bachert P (2011) Quantitative separation of applied for longitudinal cohort evaluation of articular cartilage re- CEST effect from magnetization transfer and spillover effects by generation. Osteoarthr Cartil 21:943–949 Lorentzian-line-fit analysis of z-spectra. J Magn Reson 211:149– 10. Li X, Benjamin Ma C, Link TM et al (2007) In vivo T(1rho) and T(2) mapping of articular cartilage in osteoarthritis of the knee 28. Singh A, Haris M, Cai K et al (2012) Chemical exchange saturation using 3 T MRI. Osteoarthritis Cartilage 15:789–797 transfer magnetic resonance imaging of human knee cartilage at 3 T 11. Reddy R, Insko EK, Noyszewski EA et al (1998) Sodium MRI of and 7 T. Magn Reson Med 68:588–594 human articular cartilage in vivo. Magn Reson Med 39:697–701 29. Windschuh J, Zaiss M, Meissner JE et al (2015) Correction of B1- 12. Shapiro EM, Borthakur A, Dandora R et al (2000) Sodium visibility inhomogeneities for relaxation-compensated CEST imaging at 7T. and quantitation in intact bovine articular cartilage using high field NMR Biomed 28:529–537 Na-23 MRI and MRS. J Magn Reson 142:24–31 30. Schleich C, Bittersohl B, Miese F et al (2016) Glycosaminoglycan 13. Wheaton AJ, Borthakur A, Shapiro EM et al (2004) Proteoglycan chemical exchange saturation transfer at 3T MRI in asymptomatic loss in human knee cartilage: quantitation with sodium MR imag- knee joints. Acta radiol 57:627–632 ing—feasibility study. Radiology 231:900–905 31. Müller-Lutz A, Schleich C, Schmitt B et al (2016) Gender, BMI and 14. Ling W, Regatte RR, Navon G, Jerschow A (2008) Assessment of T2 dependencies of glycosaminoglycan chemical exchange satura- glycosaminoglycan concentration in vivo by chemical exchange- tion transfer in intervertebral discs. Magn Reson Imaging 34:271– dependent saturation transfer (gagCEST). Proc Natl Acad Sci U S A 105:2266–2270 32. Müller-Lutz A, Schleich C, Pentang G et al (2015) Age- 15. Krusche-Mandl I, Schmitt B, Zak L et al (2012) Long-term results 8 dependency of glycosaminoglycan content in lumbar discs: A 3T years after autologous osteochondral transplantation: 7 T gagCEST gagCEST study. J Magn Reson Imaging 42:1517–1523
European Radiology – Springer Journals
Published: Jan 30, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera