Establishing an in vitro model of MR-T1ρ imaging technology to quantify nucleus pulposus tissue proteoglycans: a preliminary study

Introduction

Low back pain is common, and in severe cases, can affects people’s quality of life and ability to work (1). Worsening lumbar disc degeneration can cause lumbar disc herniation. Therefore, it important to detect and treat degeneration of the lumbar intervertebral disc as early as possible in the clinical setting to delay or even reverse the degeneration of the intervertebral disc and prevent further deterioration.

The accurate diagnosis of early degeneration of the lumbar intervertebral disc has important clinical value and significance. Previous studies have confirmed that the early degeneration of lumbar intervertebral disc is first reflected in the content change of proteoglycans in the nucleus pulposus tissue (2); however, the non-invasive imaging methods currently used cannot quantitatively detect proteoglycans. Due to the tissue structure of the lumbar intervertebral disc, the early diagnosis of degeneration cannot be confirmed by needle biopsy. Magnetic resonance imaging (MRI) is one of the most commonly used imaging techniques for diagnosing soft tissue diseases. Many scholars have reported that magnetic resonance T1rho (MR-T1ρ) imaging technology can quantify the changes of proteoglycan content in the nucleus pulposus tissue of early lumbar disc degeneration (3,4), and that it is more sensitive than conventional magnetic resonance T2 weighted image (MR-T2WI) imaging and T2-mapping imaging. In addition, some scholars have found that when studying degeneration of the lumbar intervertebral disc in humans and cadavers, MR-T1ρ imaging is more sensitive than MR-T2WI imaging in the early degeneration of the lumbar intervertebral disc (5). The proteoglycan concentration in the lumbar intervertebral disc nucleus pulposus tissue has been analyzed in many studies, which have found that the proteoglycan concentration has a good and positive correlation with the T1ρ relaxation time value (6-8).

However, when reviewing studies using MR-T1ρ imaging to quantitatively assess early lumbar disc degeneration, we found that in the lumbar disc nucleus pulposus tissue with the same Pfirrmann degeneration grade, significant differences in the MR-T1ρ relaxation time value have been reported (9). When we studied the correlation between the proteoglycan content in the nucleus pulposus tissue of the natural degeneration and the MR-T1ρ value of the rhesus monkeys in the early stage, we found significant differences in the MR-T1ρ value when the same lumbar disc nucleus pulposus tissue was detected on different magnetic resonance equipment. Therefore, if the MR-T1ρ value measured by the magnetic resonance equipment is directly used to quantitatively evaluate the proteoglycan content of the human lumbar intervertebral disc nucleus pulposus, the MR-T1ρ value data obtained by different examination equipment of intervertebral discs with similar degrees of degeneration will also have differences, which cannot meet the requirements of a clinically accurate diagnosis of early degeneration of lumbar intervertebral disc. Therefore, when using MR-T1ρ imaging technology to quantitatively evaluate the degeneration of the lumbar intervertebral disc, an in vitro calibration model is required on clinical application to solve the detection error of the magnetic resonance T1rho (MR-T1ρ) sequence on different machines. In the present study, we constructed a series of in vitro models of nucleus pulposus tissue with different proteoglycan concentrations to measure the MR-T1ρ relaxation time of the in vitro models. We determined if there were any similar characteristics between the in vitro nucleus pulposus tissue models and the human degenerated lumbar disc nucleus pulposus tissue, providing theoretical data for the establishment of a standardized mathematical model of degenerated lumbar intervertebral disc nucleus pulposus tissue and MR-T1ρ relaxation time value in the future. We present the following article in accordance with the MDAR reporting checklist (available at https://dx.doi.org/10.21037/atm-21-4297).

Methods

Proteoglycan concentration in the in vitro model of nucleus pulposus tissue

The 75% proportion was selected as the water content of the in vitro model, which was 2.25 g. The concentration of proteoglycan and the content of type II collagen fibers were set. Three groups were involved in the present study. In the first group, the water content and type II collagen fibers (West Asia Reagent Company, Chengdu, China) were set to 2.25 and 0.3 g, respectively. The concentration range of proteoglycan (West Asia Reagent Company) was set to 5–35%, with a concentration gradient of 5%. The 2 control groups (water with proteoglycan and water with type II collagen) and all of the in vitro models in this group were not sterilized after construction. To ensure uniformity in the in vitro model, the water content in the second group of the in vitro models was increased from 2.25 to 3 g, the content of type II collagen fiber remained 0.3 g, and the proteoglycan concentration range was set to 10–24%, with a concentration gradient of 1%. After the model was constructed, it needed to be sterilized by ultraviolet light and sealed and stored at a low temperature (10–15 °C). The water content and type II collagen fibers in the third group of in vitro models was the same as that in the first group; the proteoglycan concentration range was set to 10–29%, with a concentration gradient of 1%. After the in vitro model in the third experimental group was constructed, it also needed to be sterilized by ultraviolet light and sealed and stored at a low temperature (10–15 °C).

Methods of constructing an in vitro model of nucleus pulposus tissue

First, the reagent bottle was rinsed with deionized water and then sterilized and dried. The required proteoglycan and type II collagen were weighed in a sterile laboratory, and the materials were put into the reagent bottle. The required amount of water was added using a micropipette. The materials were stirred with a magnetic stirrer (Dalong Xingchuang, Beijing, China), by ultrasonic mixing, heating and shaking to mix the molding material into a gel-like material, and finally sterilized by ultraviolet light and stored at a low temperature. The model samples in the first group were modeled with a magnetic stirrer for 30 min, and the samples were not sterilized by ultraviolet rays or stored at low temperatures after modeling. The model samples in the second group were modeled by ultrasonic stirring for 30 min. After modeling, the samples were sterilized by ultraviolet light and stored at a low temperature (10–15 °C). The model samples in the third group were heated and shaken for 24 h (temperature: 60 °C; the temperature was automatically increased from room temperature to 60 °C within 24 h), and then ultrasonically stirred for 30 min for modeling. After modeling, the samples were sterilized by ultraviolet light and stored at a low temperature (10–15 °C). When the in vitro model failed to form a gel-like substance or there was obvious stratification, the in vitro model construction was considered to have failed.

Acquisition of magnetic resonance image

The clinical 3.0T GE magnetic resonance scanner (Discovery MR 750w 3.0T; GE Medical System, Boston, Massachusetts, USA) was used to scan the T1ρ sequence of the in vitro model. Two researchers who did not participate in the modeling selected the region of interest (ROI). Finally, the post-processing software provided by GE was employed to calculate the T1ρ relaxation time value. The human lumbar intervertebral disc was also scanned by T1ρ sequence with the same parameter, and its T1ρ relaxation time value was measured. The human lumbar intervertebral disc was scanned using the conventional T2WI sequence and then graded by a professional according to the Pfirrmann classification.

The main parameters of the MR-T1ρ imaging technology included pulse sequence, fast spin echo (10, 30, and 60 ms), repetition time: 80 ms, echo time: 8 ms, field of view: 160 mm × 100 mm, frequency (pixel bandwidth): 500 Hz, voxel size: 0.59 mm × 1.17 mm, Slice thickness: 3 mm, interslice gap: 0 mm, number of slices: 5–7, and acquisition time (min): 4 min. According to the volume of the in vitro model samples, the MR-T1ρ sequence scan in the first group was scanned by small size flexible-joint phased array coils with 8-channel. The second and third groups were scanned with small size flexible-joint phased array coils with 16-channel.

Image post-processing

A double-blind experimental method was adopted to obtain the ROI. Drawing of the ROI area was done by 2researchers. The basic conditions of selecting the ROI area were as follows (Figure 1A,1B): (I) the participator should choose at least 2 levels closing to the center. Edge area and defect part should be avoided to make the value of each region have a uniform representation; (II) the shape of the ROI was generally elliptical, and each ROI drawn was similar in size, and (III) the T1ρ relaxation time value of the ROI was measured and recorded 3 times for each sample in the in vitro model, and the average value was taken as the final statistical analysis data. The T1ρ value was calculated using a previously reported method (10). After being stored at a low temperature of 10–15 °C for 3 months, the third group of in vitro model samples was rescanned by MR-T1ρ with the same scanning parameters.

Figure 1 Schematic diagram of measurement of the in vitro model magnetic resonance T1rho (MR-T1ρ) relaxation time value. (A) Schematic diagram of the region of interest (ROI) of the coronal position of the in vitro model. (B) Schematic diagram of the horizontal position of the ROI of the in vitro model.

Statistical analysis

All statistical data were analyzed using SPSS version 25.0 (SPSS Inc. Chicago, IL, USA). Measurement data were expressed as mean ± standard deviation. To test whether the T1ρ relaxation time value of the in vitro model in line with normal distribution, the Pearson correlation coefficient was used to analyze the correlation between the in vitro model proteoglycan concentration and the T1ρ value. The slope of the regression equation between the T1ρ value and Pfirrmann grading of the in vitro model and volunteer lumbar disc nucleus pulposus tissue was analyzed by covariance analysis. The consistency test and evaluation of the T1ρ value of the same batch of in vitro model samples at different detection times were analyzed by intraclass correlation coefficient (ICC). P<0.05 indicated statistical difference.

Results

MR-T1ρ imaging of the in vitro model and its T1ρ relaxation time value

Of the 10 in vitro model samples in the first experimental group, sample 8 failed to form, so the success rate of the model making was 90%. The MR-T1ρ images of the in vitro model are shown in Figure 2. The biochemical compositions of each sample successfully modeled and their corresponding T1ρ values are shown in Table 1. Of these, in vitro model group 1 was used as the control group. Type II collagen was not added to this group, and its T1ρ value was 200 ms. Proteoglycan was not added to control group 10, with its T1ρ value was 306.1 ms. T1ρ values of in vitro models 2–9 in the experimental group deviated significantly from the control group.

Figure 2 Magnetic resonance T1rho (MR-T1C) pseudo-color image of in vitro model samples in the first experimental group in the coronal position. There were defective areas in small amount of the samples, as indicated by the uneven color distribution.

Of the 15 in vitro model samples in the second experimental group, the success rate of model making was 100%. The MR-T1ρ images of the in vitro model sample are shown in Figure 3, and small, spot-like or patchy defects could be seen in the model sample. The biochemical compositions of each sample and their corresponding T1ρ values are shown in Table 2.

Figure 3 Magnetic resonance T1rho (MR-T1ρ) sequence pseudo-color image of in vitro model samples in the second experimental group in the horizontal position.

Of the 20 in vitro model samples in the third experimental group, 5 samples were not modeled, so the success rate of the model-making process was 75%. The samples modeled successfully were tested twice. The MR-T1ρ images of the model samples tested for the first time are shown in Figure 4. No defect area was seen inside the model. The model samples tested for the second time were stored at 10–15 °C for 3 months, and then MR-T1ρ imaging was re-performed, with MR-T1ρ images shown in Figure 5. Of these, 4 model samples had small patchy defects inside. The biochemical components and their T1ρ values of 15 successful in vitro model samples tested twice are shown in Table 3.

Figure 4 Magnetic resonance T1rho (MR-T1ρ) sequence pseudo-color image of in vitro model samples in the third experimental group in the horizontal position.

Figure 5 Magnetic resonance T1rho (MR-T1ρ) sequence pseudo-color image after 3 months modeling of the in vitro model samples in the third experimental group in the horizontal position.

The ICC was further used to test the consistency of the 2 T1ρ values before and after the cold storing of each sample. The results showed that the ICC of the T1ρ values of the 2 scans was 0.980, which was greater than 0.75, suggesting that the biochemical components of the in vitro model samples in the third experimental group were still relatively stable after being stored for 3 months.

Correlation between in vitro model proteoglycan concentration and T1ρ value of the 3 groups

Plotting was done by Pearson linear regression analysis, and the T1ρ relaxation time values of samples 2–9 in the first group were found to be normally distributed. The results showed that there was no correlation between the in vitro model proteoglycan concentration and T1ρ values (Y=1.49X+34.43, R²=0.518, P>0.05) (Figure 6A). The T1ρ values of the 15 samples in the second group were normally distributed, as plotted by Pearson linear regression analysis. The results showed that there was a distinctly negative correlation between the proteoglycan concentration of the in vitro model samples and their T1ρ values (Y=–3.02X+131.8, R²=0.852, P<0.05) (Figure 6B).

Figure 6 Correlation analysis between the proteoglycan concentration and T1rho (T1ρ) values of the model samples in the experimental groups. (A) Scatterplot and regression line of the proteoglycan concentration of the in vitro model samples in the first experimental group and their T1ρ values; (B) scatterplot and regression line of the proteoglycan concentration of the model samples in the second experimental group and their T1ρ values; (C) scatterplot and the regression line of the proteoglycan concentration and T1ρ values for the first and second measurements of the model samples in the third experimental group.

The T1ρ values of the 15 in vitro model samples in the third group were normally distributed, as plotted by Pearson linear regression analysis. The results showed that there was a distinctly positive correlation between the proteoglycan concentration in the first in vitro model samples and the T1ρ values (Y=3.05X+11.99, R²=0.834, P<0.05) (Figure 6C). The comparison of the slope of the regression equation between the proteoglycan concentration in the third experimental group in the in vitro model sample and the first measurement, as well as the second measurement, showed that there was no significant difference between them (F=1.27, P=0.270), with a good degree of fit. This finding indicated that the biochemical components of the model samples were still relatively stable after being sealed and cryopreserved.

MR-T1ρ correlation analysis of the third experimental group in the in vitro model and lumbar intervertebral disc of the volunteers

The MR-T1ρ imaging of the in vitro model samples were compared with the human lumbar intervertebral discs of volunteers. In this experiment, the MR-T1ρ sequence with the same parameters was used to scan the lumbar intervertebral discs of 4 adult volunteers with no low back pain, and the T1ρ values of the nucleus pulposus tissue of their intervertebral discs were measured (Table 4). T2WI sequence scanning was applied to obtain T2WI images and Pfirrmann grading evaluation to establish the regression equation between the Pfirrmann grading of the lumbar disc nucleus pulposus tissues of volunteers and their T1ρ values (Y=–10.43X+75.60, R²=0.731) (Table 5, Figures 7,8).

Figure 7 Sequence images. (A) Magnetic resonance T2 weighted image (MR-T2WI) sequence images of volunteers' lumbar intervertebral discs; (B) Magnetic resonance T1rho (MR-T1ρ) sequence pseudo-color images.

Figure 8 Scatterplot and regression line of the regression equation between the Pfirrmann grading and T1rho (T1ρ) value of the model samples in the third experimental group and volunteers.

To evaluate the fit degree of the T1ρ values between the third experimental group model samples and the lumbar intervertebral disc nucleus pulposus tissues of volunteers, the in vitro model sample proteoglycan concentration of 8–10% was set as Pfirrmann III, 11–16% as Pfirrmann II, and 17–20% as Pfirrmann I. The regression equation between the Pfirrmann grading and T1ρ values of the in vitro model samples in the third experimental group was obtained (Y=–8.69X+133.6, R²=0.765) (Figure 8). The covariance comparison of the slope of the regression equation between the Pfirrmann grading and T1ρ values between the in vitro model samples in the third experimental group and the lumbar intervertebral disc nucleus pulposus tissues of volunteers indicated no significant differences between them (F=0.54, P=0.814), suggesting that the slope of regression equation between the Pfirrmann grading and T1ρ value in the in vitro model samples and lumbar disc nucleus tissues of volunteers was a good fit.

Discussion

Approximately 80% of people worldwide will experience low back pain during their lifetime, and 5% will develop chronic low back pain, which affects quality of life and ability to work (11). Degeneration of the lumbar intervertebral disc is the main cause of low back pain. China has a large, aging population, so early diagnosis and the prevention and treatment of this disease it is of great economic.

The most accurate diagnostic method for early degeneration of lumbar intervertebral disc is to directly obtain the nucleus pulposus tissue to quantitatively measure the biochemical components, but this method is associated with invasive injury to the intervertebral disc, which will aggravate the degeneration of the intervertebral disc, rendering it unsuitable for clinical application (12). Studies have shown that the main biochemical change in the early degeneration of the intervertebral disc is the reduction of proteoglycan in the nucleus pulposus tissue (13). Therefore, there is a clinical need for a non-invasive, quantitative method to detect the proteoglycan content in the nucleus pulposus tissue of the intervertebral disc.

Lumbar X-rays and computed tomography scans cannot diagnose early disc degeneration, and conventional magnetic resonance T2WI sequencing cannot quantitatively evaluate the degree of lumbar disc degeneration. At present, only MR-T1ρ and T2-mapping imaging technology can be used to quantitatively measure early lumbar disc degeneration without invasive injury (12). Studies have shown that using MR-T1ρ sequence to quantitatively evaluate nucleus pulposus tissue proteoglycan is superior to other magnetic resonance sequences (14). Because MR-T1ρ imaging technology is sensitive to the low-frequency movement of macromolecular substances, and the proteoglycan molecules in the nucleus pulposus tissue of the intervertebral disc are macromolecular substances, MR-T1ρ technology can be used to quantitatively evaluate the proteoglycan content in the nucleus pulposus tissue of the intervertebral disc (15,16). Previous studies have shown that the T1ρ relaxation time is significantly and positively correlated with the proteoglycan content of human or animal lumbar intervertebral discs (17,18). Our research group has researched the natural degeneration of lumbar intervertebral discs in rhesus monkeys, and confirmed that the proteoglycan content of the lumbar disc nucleus pulposus of rhesus monkeys is positively correlated with the T1ρ value. However, we also found that the T1ρ value of the lumbar intervertebral disc from the same rhesus monkey was significantly different when scanned by different MRI equipment. Therefore, the T1ρ value cannot be directly used to accurately diagnose the early degeneration of the human lumbar intervertebral disc without calibration.

The quantification and calibration of T1ρ values are technical difficulties in the clinical setting (19), and some methods have been reported to calibrate T1ρ values to increase the detection rate of diseases (20-22). In the present study, we set up in vitro model samples with different proteoglycan content or water content to simulate different degenerated nucleus pulposus tissues. In the in vitro model in the first experimental group, the results showed that more defective areas were seen in the MR-T1ρ image of the modeled samples. After changing the experimental conditions, some spots or flaky defect areas still could still be seen from the MR-T1ρ image of the in vitro model in the second experimental group. After another conditional improvement, the MR-T1ρ image showed no obvious defect area in the modeled samples in the third experimental group, and the proteoglycan concentration was positively correlated with the T1ρ value. The in vitro model samples in the third group underwent MR-T1ρ imaging re-scanning after being sterilized by ultraviolet rays and sealed at a low temperature (10–15 °C) for 3 months. According to the ICC value, the biochemical components of the model samples were still stable after a long period of storage. The results of these 3 groups confirmed that the ideal in vitro model with stable biochemical components needs a higher water content and the cryopreserved samples should be aseptically seal.

We selected 4 adult volunteers with no low back pain to perform the T1ρ sequence scan and T2WI sequence scan with the same parameters, and evaluated the Pfirrmann grading of their lumbar intervertebral discs. The results showed that there were no obvious differences between the slope of the regression equation between the Pfirrmann grading and T1ρ value of the in vitro model samples in the third experimental group and the lumbar intervertebral disc tissue of volunteers, suggesting that this model has a good degree of fit with the degeneration of the nucleus pulposus tissue of human lumbar intervertebral disc.

Conclusions

In this experiment, we successfully constructed an in vitro model of MR-T1ρ to quantitatively evaluate the proteoglycans of nucleus pulposus tissue. However, to accurately quantify the proteoglycan content of human lumbar intervertebral disc nucleus pulposus by clinical MR-T1ρ imaging technology, large samples are required to establish an in vitro model database and post-processing platform to accurately detect theMR-T1ρ value of human lumbar intervertebral disc nucleus pulposus tissue, and to provide a scientific method for the accurate diagnosis of early degeneration of the lumbar intervertebral disc.

Acknowledgments

Funding: None.

Footnote

Reporting Checklist: The authors have completed the MDAR reporting checklist. Available at https://dx.doi.org/10.21037/atm-21-4297

Data Sharing Statement: Available at https://dx.doi.org/10.21037/atm-21-4297

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://dx.doi.org/10.21037/atm-21-4297). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Benzakour T, Igoumenou V, Mavrogenis AF, et al. Current concepts for lumbar disc herniation. Int Orthop 2019;43:841-51. [Crossref] [PubMed]
2. Yaltirik CK, Timirci-Kahraman Ö, Gulec-Yilmaz S, et al. The Evaluation of Proteoglycan Levels and the Possible Role of ACAN Gene (c.6423T>C) Variant in Patients with Lumbar Disc Degeneration Disease. In Vivo 2019;33:413-7. [Crossref] [PubMed]
3. Zuo J, Joseph GB, Li X, et al. In vivo intervertebral disc characterization using magnetic resonance spectroscopy and T1ρ imaging: association with discography and Oswestry Disability Index and Short Form-36 Health Survey. Spine (Phila Pa 1976) 2012;37:214-21. [Crossref] [PubMed]
4. Zhang SF, Zhang B, Wang GH, et al. The application value of 3.0 T T1ρ and T2 mapping imaging in the study of cervical disc degeneration in patients. Chinese Journal of Radiology 2019;53:732-6.
5. Johannessen W, Auerbach JD, Wheaton AJ, et al. Assessment of human disc degeneration and proteoglycan content using T1rho-weighted magnetic resonance imaging. Spine (Phila Pa 1976) 2006;31:1253-7. [Crossref] [PubMed]
6. Pietrosimone B, Pfeiffer SJ, Harkey MS, et al. Quadriceps weakness associates with greater T1ρ relaxation time in the medial femoral articular cartilage 6 months following anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc 2019;27:2632-42. [Crossref] [PubMed]
7. Van Rossom S, Smith CR, Zevenbergen L, et al. Knee Cartilage Thickness, T1ρ and T2 Relaxation Time Are Related to Articular Cartilage Loading in Healthy Adults. PLoS One 2017;12:e0170002 [Crossref] [PubMed]
8. Hänninen NE, Nykänen O, Prakash M, et al. Orientation anisotropy of quantitative MRI parameters in degenerated human articular cartilage. J Orthop Res 2021;39:861-70. [Crossref] [PubMed]
9. Xiong X, Zhou Z, Figini M, et al. Multi-parameter evaluation of lumbar intervertebral disc degeneration using quantitative magnetic resonance imaging techniques. Am J Transl Res 2018;10:444-54. [PubMed]
10. Zhou Z, Jiang B, Zhou Z, et al. Intervertebral disk degeneration: T1ρ MR imaging of human and animal models. Radiology 2013;268:492-500. [Crossref] [PubMed]
11. Becker BA, Childress MA. Nonspecific Low Back Pain and Return To Work. Am Fam Physician 2019;100:697-703. [PubMed]
12. Melrose J, Ghosh P, Taylor TK. A comparative analysis of the differential spatial and temporal distributions of the large (aggrecan, versican) and small (decorin, biglycan, fibromodulin) proteoglycans of the intervertebral disc. J Anat 2001;198:3-15. [Crossref] [PubMed]
13. Sampara P, Banala RR, Vemuri SK, et al. Understanding the molecular biology of intervertebral disc degeneration and potential gene therapy strategies for regeneration: a review. Gene Ther 2018;25:67-82. [Crossref] [PubMed]
14. Duvvuri U, Goldberg AD, Kranz JK, et al. Water magnetic relaxation dispersion in biological systems: the contribution of proton exchange and implications for the noninvasive detection of cartilage degradation. Proc Natl Acad Sci U S A 2001;98:12479-84. [Crossref] [PubMed]
15. Belavý DL, Quittner MJ, Ridgers N, et al. Running exercise strengthens the intervertebral disc. Sci Rep 2017;7:45975. [Crossref] [PubMed]
16. Chen L, Liao J, Klineberg E, et al. Small leucine-rich proteoglycans (SLRPs): characteristics and function in the intervertebral disc. J Tissue Eng Regen Med 2017;11:602-8. [Crossref] [PubMed]
17. Peixoto-Santos JE, Kandratavicius L, Velasco TR, et al. Individual hippocampal subfield assessment indicates that matrix macromolecules and gliosis are key elements for the increased T2 relaxation time seen in temporal lobe epilepsy. Epilepsia 2017;58:149-59. [Crossref] [PubMed]
18. Ying J, Han Z, Zeng Y, et al. Evaluation of intervertebral disc regeneration with injection of mesenchymal stem cells encapsulated in PEGDA-microcryogel delivery system using quantitative T2 mapping: a study in canines. Am J Transl Res 2019;11:2028-41. [PubMed]
19. Chen W. Errors in quantitative T1rho imaging and the correction methods. Quant Imaging Med Surg 2015;5:583-91. [PubMed]
20. Qi H, Bustin A, Kuestner T, et al. Respiratory motion-compensated high-resolution 3D whole-heart T1ρ mapping. J Cardiovasc Magn Reson 2020;22:12. [Crossref] [PubMed]
21. Andronesi OC, Nicholson K, Jafari-Khouzani K, et al. Imaging Neurochemistry and Brain Structure Tracks Clinical Decline and Mechanisms of ALS in Patients. Front Neurol 2020;11:590573 [Crossref] [PubMed]
22. Gram M, Seethaler M, Gensler D, et al. Balanced spin-lock preparation for B1 -insensitive and B0 -insensitive quantification of the rotating frame relaxation time T1ρ. Magn Reson Med 2021;85:2771-80. [Crossref] [PubMed]

(English Language Editor: R. Scott)