- Email: [email protected]
Imaging biomarker with T1ρ and T2 mappings in osteoarthritis – In vivo human articular cartilage study
European Journal of Radiology 82 (2013) 647–650
Contents lists available at SciVerse ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Imaging biomarker with T1 and T2 mappings in osteoarthritis – In vivo human articular cartilage study Chun Sing Wong a,∗ , Chun Hoi Yan b , Nan Jie Gong c , Teng Li d , Queenie Chan e , Yiu Ching Chu f a
FHKAM (Radiology), Department of Diagnostic Radiology, The University of Hong Kong, Hong Kong FHKAM (Orthopedics), Department of Orthopedics and Traumatology, the University of Hong Kong, Hong Kong c Department of Diagnostic Radiology, the University of Hong Kong, Hong Kong d Department of Orthopedics and Traumatology, the University of Hong Kong, Hong Kong e Philips Electronics Hong Kong Ltd, Hong Kong f Tiffany, FHKAM(radiology), Department of Radiology, Kwong Wah Hospital, Hong Kong b
a r t i c l e
i n f o
Article history: Received 28 August 2012 Received in revised form 22 November 2012 Accepted 24 November 2012 Keywords: Osteoarthritis MRI T1 mapping T2 mapping Glycoaminoglycan
a b s t r a c t Introduction: Osteoarthritis (OA) of the knee is a common and disabling disease worldwide. Its prevalence is increasing in view of the aging population. Changes in collagen content, its orientation and GAG content in the articular cartilage with age are the major features in knee osteoarthritis. These changes in collagen and GAG contents show no manifestation in plain radiography and conventional magnetic resonance imaging (MRI). Nevertheless, early diagnosis of the knee osteoarthritis is of paramount importance clinically in view of the evolution of putative interventions in its early stage. The aim of this project is to identify the relationships between the two imaging biomarkers (i.e. T1 and T2 mappings) and the GAG concentration in living human symptomatic cartilage. Methodology: 28 patients with clinical diagnosis of knee osteoarthritis were enrolled. 7 males and 16 females were recruited and their mean age was 68.1 (ranges from 53 to 84). Conventional PD sequence, T1 and T2 mappings were performed for each subject within 1 week before total knee arthroplasty. Articular cartilage from the lateral tibial plateau was harvested and the GAG content in cartilage was determined by using dimethylmethylene blue method. T1 mean and T2 values were calculated and correlate with GAG concentration statistically. Results: The mean value T1 was 40.3 ± 13.5 ms, ranging from 15.3 to 69.3 ms and the mean value T2 was 31.0 ± 10.5 ms, ranging from 16.1 to 46.9 ms. The mean value of GAG content was 80.1 ± 33.3 mg, ranging from 24.9 to 166.0 mg while the mean value of GAG concentration was 267.4 ± 165.9 mg/cm3 , ranging from 91.3 to 760.5 mg/cm3 . T2 values were inversely correlated with GAG concentration with R2 = 0.375, p = 0.001 while T1 values were also inversely correlated with GAG concentration with R2 = 0.200, p = 0.025. Conclusion: This in vivo study confirmed that T1 and T2 values correlate with the GAG concentration in living human knee cartilages which corroborate with the previous works. The later (T2 values) is found more reliable in our study and less controversial in literatures. We postulate that T2 values can serve as a non-invasive imaging biomarker in the progress of knee osteoarthritis in terms of both disease diagnosis and treatment response monitoring. © 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Osteoarthritis (OA) of the knee is a common and disabling disease worldwide. Its prevalence is increasing in view of the aging population. A local study conducted in 2000 has demonstrated that
∗ Tel.: +852 22553308. E-mail addresses: [email protected] (C.S. Wong), [email protected] (C.H. Yan), [email protected] (N.J. Gong), [email protected] (T. Li), [email protected] (Q. Chan), [email protected] (Y.C. Chu). 0720-048X/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2012.11.036
the prevalence of knee osteoarthritis is 7% and 13% in male and female, respectively [1]. The result is in parallel with the data in United States [2]. Osteoarthritis of the knee is a degenerative process of the joint in which low grade inflammation occurs in the articular hyaline cartilage. This results in abnormal cartilage wearing and thus in turn causes joint pain and movement limitation. Functional loss and daily living deterioration are the unavoidable side effects of this disease. Patients suffering from severe OA knee will have an enormous detrimental impact to their quality of life and the disease is very costly to the medical system as a whole.
648
C.S. Wong et al. / European Journal of Radiology 82 (2013) 647–650
Articular hyaline cartilage is a hierarchical structure which is divided into layers with different amount of type II collagen fibrils oriented in the glycosaminoglycan (GAG) matrix [3]. Changes in collagen content or its orientation and GAG content in the articular cartilage with age are the major features in knee osteoarthritis [4]. Loss of GAG is characteristic in early stage of OA which is followed by alteration in collagen structures [5,6]. These changes in collagen and GAG contents show no manifestation in plain radiography and conventional magnetic resonance imaging (MRI). Nevertheless, early diagnosis of the knee osteoarthritis is of paramount importance clinically in view of the evolution of putative interventions in its early stage [7–9]. Researchers had therefore spent tremendous efforts to explore a non-invasive imaging biomarker to achieve the above mentioned intuitive goal. New and advanced MRI sequences such as T1 or T2 mappings have been widely studied recently for this purpose [10–12]. In T1 sequence, a “spin lock” pulse is applied to the magnetization transversely and T1 relaxation time is thought to be sensitive to protons on macromolecules such as GAG. The drop of GAG content will increase the mobile proton density in bulk water and thus increases the T1 value. Yet, previous literatures in T1 mapping were not straight forward and many of them had demonstrated conflicting results. Menezes et al. and Mlynarik et al. had negated the correlation with T1 with GAG content [13,14] while Nishioka et al. and Duvvuri et al. had made an opposite statement [16,15]. In 2004, Menezes et al. even reported correlation between T1 values and collagen concentration in his in vitro study using three human articular cartilage-bone specimens [13]. Therefore the general consensus has regarded T1 sequence as a research tool with minimal clinical application. T2 parameter is a spin–spin relaxation method which reflects the cartilage water content. It is believed to represent the collagen anisotropy as protons in water molecules are arranged within collagen fibers [17]. However, recent projects had found that the T2 value also correlate well with the GAG content [15,18,24]. Previous works were predominantly in vitro studies and had made a preliminary conclusion that T1 values represent GAG content while T2 values represent collagen content. However, their implications in living human cartilage are not yet completely elucidated. In vivo study concerning this issue is sparse in the literatures and most of the previous works were small sample cohorts. Therefore, we have devised this project to identify the relationships between the two imaging biomarkers (i.e. T1 and T2 mappings) and the GAG concentration in living human symptomatic cartilage. Up to our knowledge, this is by far the study with the biggest cohort in literatures answering this clinical question. 2. Methodology 2.1. Patients Twenty-eight patients with clinical diagnosis of knee osteoarthritis were enrolled. 7 males and 16 females were recruited and their mean age was 68.1 (ranges from 53 to 84). MRI was performed for each subject within 1 week before total knee arthroplasty is carried out (range from 3 to 6 days). Articular cartilage from the lateral tibial plateau was harvested and the GAG content in cartilage was determined by using dimethylmethylene blue method. T1 mean and T2 values were calculated and correlate with GAG concentration statistically. Two patients had both knees scanned and three patients were excluded due to complete cartilaginous denudation in MR images, making measurement impossible. Twenty-five data points were calculated at the end. This study was approved by the institutional review board in Hong Kong and informed consent was obtained from every patient.
2.2. Harvesting of human articular cartilage All patients underwent total knee arthroplasty (TKA) after the MRI scans. The TKA was performed using the Install approach. After the soft tissue release and balancing, the proximal tibial bone cut was performed perpendicular to the mechanical axis, using the extramedullary cutting guide. Approximately 10 mm bone was resected, measuring from the most proximal point of lateral tibia plateau. Attention was paid not to damage the articular cartilage on the lateral side. The removed specimens were immersed in normal saline and delivered to the laboratory immediately. The articular cartilage over the lateral tibial plateau was harvested carefully [19]. All specimens were immersed in normal saline and kept in −30 ◦ C until experiments started. 2.3. GAG content analysis A stable solution of 1,9-dimethylmethylene blue (DMMB) was prepared as follows: DMMB (21 mg) was stirred with 5 ml ethanol; sodium hydroxide (1.72 g) and formic acid (1.5 ml) were added, and the volume made up to 1 l with distilled water. The reagents were stored in bottle sealed completely to avoid light at room temperature [20]. Cartilage sample solutions were prepared by digestion with proteinase K (2 mg) in 30 ml TE buffer (pH 8.0) containing 100 mM NaCl, 10 mM TRIS, 1 mM EDTA at 60 ◦ C. Sample solutions were diluted with TE buffer to adjust GAG concentration to a proper value. Cartilage sample solution (40 l) was pipetted into a 96 wells plate. DMMB solution (200 l) was added with good mixing. 40 l of chondroitin sulfate A of different concentration (0, 8, 16, 32, 64, 128 g/ml) was mixed with 200 l DMMB solution to prepare the standard curve. Absorbance was read at 595 nm and the concentrations of the GAG content in the samples were estimated based on the chondroitin sulfate A standard curve. 2.4. MRI sequence All MRI examinations were performed using a 3T Achieva scanner (Philips Healthcare, Best, The Netherlands) and an eight channel SENSE knee coil. The knee was positioned at 15◦ flexion and in a neutral position of rotation so as to minimize the magic angle artifact. Conventional PD sagittal and coronal scan were performed. Coronal T1 weighted images were obtained with a 3D balanced turbo field echo (B-TFE) sequence with the following parameters: TR/TE = 5.0 ms/2.3 ms, flip angle = 50◦ , FOV = 140 mm × 140 mm, slice thickness = 4 mm. Five subsequent T1 weighted scans were performed with spin lock durations 0, 10, 20, 30, 40 ms and with scan time of 2 min 33 s for each scan. Images were fitted on a pixel by pixel basis to the exponentially decaying T1 function using IDL (Research Systems, Inc., USA) to generate T1 relaxation map. Quantitative T2 measurement in identical geometry as that of the T1 image was performed using standard multi-echo spin-echo sequence with TE = 16, 32, 48, 64, 80, 96, 112 ms. Scan time = 11 min 25 s. Image analyses were conducted by using Image J software (NIH, Bethesda, MD). Regions of interest (ROIs) were drawn manually by two specialist radiologists (CSW, YCC) on the T1 and T2 mappings with reference to the PD images (Fig. 1). To calculate the volume of each cartilage, the volume corresponding to every drawn ROI was first computed. Slice thickness plus slice gap was regarded as apparent slice thickness. Products of area included in ROI and apparent slice thickness were calculated as volume of ROI. Then, sum of all the ROIs was calculated as volume of cartilage. To be more accurate, the final volume was an averaged volume of measurements from T2 and T1 images. To account for the volume effect in GAG measurement, we calculated
C.S. Wong et al. / European Journal of Radiology 82 (2013) 647–650
649
Fig. 2. The correlation between T2 and extracellular matrix composition: T2 values were negatively correlated with GAG concentration (R2 = 0.375, p = 0.001).
3. Results and discussion The mean value T1 was 40.3 ± 13.5 ms, ranging from 15.3 to 69.3 ms and the mean value T2 was 31.0 ± 10.5 ms, ranging from 16.1 to 46.9 ms. The mean of GAG content was 80.1 ± 33.3 mg, ranging from 24.9 to 166.0 mg while the mean of GAG concentration was 267.4 ± 165.9 mg/cm3 , ranging from 91.3 to 760.5 mg/cm3 . T2 values were inversely correlated with GAG concentration with R2 = 0.375, p = 0.001 (Fig. 2) while T1 values were also inversely correlated with GAG concentration with R2 = 0.200, p = 0.025 (Fig. 3).
Fig. 1. Knees were imaged using a 3T Achieva scanner. T1 relaxation time map was generated by fitting on a pixel by pixel basis using IDL (Research Systems, Inc., USA). Quantitative T2 measurement was performed using standard multi-echo spin-echo sequence. (a) PD coronal, (b) corresponding T2 and (c) corresponding T1 images.
GAG concentration of each cartilage. It was calculated by dividing the total GAG content with cartilage volume. 2.5. Data analysis and statistics Associations between T1, T2 and GAG measurements were tested using Pearson correlation. All p values were two tailed and 0.05 was chosen as significant level. All statistical analyses were performed using SPSS software (v.19.0.0; Chicago, IL).
Fig. 3. The correlation between T1 and extracellular matrix composition: T1 values were negatively correlated with GAG concentration (R2 = 0.200, p = 0.025).
650
C.S. Wong et al. / European Journal of Radiology 82 (2013) 647–650
Our results demonstrated that the T1 and T2 values significantly and inversely correlate with the GAG concentration (R2 = 0.200, p = 0.025; R2 = 0.375, p = 0.001, respectively). The result about the T1 corroborates with one end of the pendulum. For instance, Duvvuri et al. had successfully drawn the conclusion that T1 correlate negatively with the GAG content by using bovine cartilage [15] while Meneszes et al. found no such correlation by using human specimen [17]. These conflicting results can be partly due to the different methods used for the GAG measurement and the potential errors in the dilution process. Another possibility would be the different enzymes used in different groups to degrade the cartilage and thus initiate controversies. Moreover, most previous works were conducted in in vitro fashions with controlled environment and thus application of their results in living human condition is questionable, not to mention that most works were performed with small sample cohorts. Our results in this in vivo study support the school of thought that T1 correlates inversely with the GAG concentration and coherent with the in vivo study by Nishioka et al. [16] in 2011 but with a weaker correlation (R2 = 0.200 vs R2 = 0.410). One reason for that is the difference in methodology. They collected various samples within the same subject and that could induce a sampling error as GAG content and its concentration can vary significantly even within the same block of cartilage. Therefore we tried to eliminate this error by measuring the entire block of cartilage as a whole. Also, the recruited subjects in our study were all in a late stage which T1 value is not only affected by loss of GAG as in early stage but also by other types of macromolecule change such as collagen loss. We have tried to lessen this effect by harvesting the non-weight bearing side of cartilage which is less worn out. The relationship between T2 value and GAG concentration has remained unclear in literatures. Proteoglycan or collagen depletion would increase the free space inside the extra-cellular matrix in cartilage and thus increase the influx of water and therefore T2 value will increase [21]. Therefore T2 value has been considered representing not only collagen anisotropy but also GAG concentration. Our results on T2 values found significant inverse correlation with the cartilaginous GAG concentration which reiterates the current literatures. The correlation is stronger than that of the T1 values (T1: R2 = 0.200; T2: R2 = 0.375). T2 value has been well described in its correlation with collagen content and with that of GAG content has also been revealed recently. Unlike T1, none of them negate the correlation between T2 value and GAG concentration, and therefore, T2 value is thought to be a less controversial imaging biomarker in knee cartilaginous degeneration. Moreover, less scan time is required for T2 mapping and its post processing technique is more robust and well established. All these would render it a more feasible imaging biomarker in clinical application. There are several limitations of our study which need to be addressed. The high exclusion rate (n = 5) would limit the sample size and then hence the results’ accuracy. Moreover, T2 values can be affected by magic angle artifact [22]. Although the magic angle artifact is thought to be small in in vivo situation [23], care has already been executed so that the articular surface was positioned perpendicular to the magnetic field to minimize the artifact. Moreover, as the entire knee cartilage (lateral tibial plateau) was assessed, zonal delineation of the cartilage can be performed to obtain a more detailed structural categorization. Lastly, late stage diseases were recruited but specimens usually cannot be obtained from early disease stage in view of the ethical issue. 4. Conclusion This in vivo study confirmed that T1 and T2 values correlate with the GAG concentration in living human knee cartilages which
corroborate with the previous works. The later (T2 values) is found more reliable in our study and less controversial in literatures. We postulate that T2 values can serve as a non-invasive imaging biomarker in the progress of knee osteoarthritis in terms of both disease diagnosis and treatment response monitoring. This result laid a foundation on further related studies with a larger cohort. References [1] Lau EC, Cooper C, Lam D, et al. Factors associated with osteoarthritis of the hip and knee in Hong Kong Chinese: obesity, joint injury, and occupational activities. American Journal of Epidemiology 2000;152(November (9)):855–62. [2] Dillon CF, Rasch EK, Gu Q, et al. Prevalence of knee osteoarthritis in the United States: arthritis data from the Third National Health and Nutritional Examination Survey 1991–94. Journal of Rheumatology 2006;33:2271–9. [3] Xia Y, Moody JB, Alhadlaq H. Orientational dependence of T2 relaxation in articular cartilage: a microscopic MRI (microMRI) study. Magnetic Resonance in Medicine 2002;48:460–9. [4] Arokoski JP, Jurvelin JS, Väätäinen U, et al. Normal and pathological adaptations of articular cartilage to joint loading. Scandinavian Journal of Medicine and Science in Sports 2000;10:186–98. [5] Burstein D, Bashir A, Gray ML. MRI techniques in early stages of cartilage disease. Investigative Radiology 2000;35:622–38. [6] Buckwalter JA, Martin J. Degenerative joint disease. Clinical Symposia 1995;47:1–32. [7] Sarzi-Puttini P, Cimmino MA, Scarpa R, et al. Osteoarthritis: an overview of the disease and its treatment strategies. Seminars in Arthritis and Rheumatism 2005;35(Suppl. 1):1–10. [8] Eshed I, Trattnig S, Sharon M, et al. Assessment of cartilage repair after chondrocyte transplantation with a fibrin–hyaluronan matrix—correlation of morphological MRI, biochemical T2 mapping and clinical outcome. European Journal of Radiology 2012;81(June (6)):1216–23. [9] Kon E, Di Martino A, Filardo G, et al. Second-generation autologous chondrocytetransplantation: MRI findings and clinical correlations at a minimum 5-year follow-up. European Journal of Radiology 2011;Sep 79(3):382–8, http://dx.doi.org/10.1016/j.ejrad.2010.04.002. [10] Zilkens C, Miese F, Kim YJ, et al. Three-dimensional delayed gadoliniumenhanced magnetic resonance imaging of hip joint cartilage at 3T: a prospective controlled study. European Journal of Radiology 2012;81(November (11)):3420–5, http://dx.doi.org/10.1016/j.ejrad.2012.04.008. [11] Juras V, Apprich S, Pressl C, et al. Histological correlation of 7T multi-parametric MRI performed in ex-vivo Achilles tendon. European Journal of Radiology 2011 [Epub ahead of print]. [12] Zilkens C, Miese F, Bittersohl B, et al. Delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC), after slipped capital femoral epiphysis. European Journal of Radiology 2011;79(3):400–6 [Epub 2010 May 26]. [13] Menezes NM, Gray ML, Hartke JR, et al. T2 and T1 MRI in articular cartilage systems. Magnetic Resonance in Medicine 2004;51:503–9. [14] Mlynarik V, Trattnig S, Huber M, et al. The role of relaxation times in monitoring proteoglycan depletion in articular cartilage. Journal of Magnetic Resonance Imaging 1999;10:497–502. [15] Duvvuri U, Reddy R, Patel SD, et al. T1 rho-relaxation in articular cartilage: effects of enzymatic degradation. Magnetic Resonance in Medicine 1997;38:863–7. [16] Nishioka H, Hirose J, Nakamura E, et al. T1 rho and T2 mapping reveal the in vivo extracellular matrix of articular cartilage. Journal of Magnetic Resonance Imaging 2012;35:147–55. [17] Nieminen MT, Töyräs J, Rieppo J, et al. Quantitative MR microscopy of enzymatically degraded articular cartilage. Magnetic Resonance in Medicine 2000;43:676–81. [18] Astrid Watrin P, Ruaud JP, Oliver P, et al. Effect of proteoglycan depletion on T2 mapping in rat patellar cartilage. Radiology 2005;234:162–70. [19] Altman R, Asch E, Bloch D, et al. Development of criteria for the classification and reporting of osteoarthritis classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis and Rheumatism 1986;29(8):1039–49. [20] Juras V, Bittsansky M, Majdisova Z, et al. In vitro determination of biomechanical properties of human articular cartilage in osteoarthritis using multi-parametric MRI. Journal of Magnetic Resonance 2009;197(1):40–7. [21] Xia Y. Magic angle effect in magnetic resonance imaging of articular cartilage: review. Investigative Radiology 2000;35:602–21. [22] Akella SV, Regatte RR, Wheaton AJ, et al. Reduction of residual dipolar interaction in cartilage by spin lock technique. Magnetic Resonance in Medicine 2004;52:1103–9. [23] Mosher TJ, Smith H, Dardzinski BJ, et al. Imaging and T2 mapping of femoral cartilage: in vivo determination of the magic angle effect. American Journal of Roentgenology 2001;177:665–9. [24] Keenan KE, Besier TF, Pauly JM, et al. Prediction of glycosaminoglycan content in human cartilage by age T1 rho and T2 MRI. Osteoarthritis and Cartilage 2011;19:171–9.
Contents lists available at SciVerse ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Imaging biomarker with T1 and T2 mappings in osteoarthritis – In vivo human articular cartilage study Chun Sing Wong a,∗ , Chun Hoi Yan b , Nan Jie Gong c , Teng Li d , Queenie Chan e , Yiu Ching Chu f a
FHKAM (Radiology), Department of Diagnostic Radiology, The University of Hong Kong, Hong Kong FHKAM (Orthopedics), Department of Orthopedics and Traumatology, the University of Hong Kong, Hong Kong c Department of Diagnostic Radiology, the University of Hong Kong, Hong Kong d Department of Orthopedics and Traumatology, the University of Hong Kong, Hong Kong e Philips Electronics Hong Kong Ltd, Hong Kong f Tiffany, FHKAM(radiology), Department of Radiology, Kwong Wah Hospital, Hong Kong b
a r t i c l e
i n f o
Article history: Received 28 August 2012 Received in revised form 22 November 2012 Accepted 24 November 2012 Keywords: Osteoarthritis MRI T1 mapping T2 mapping Glycoaminoglycan
a b s t r a c t Introduction: Osteoarthritis (OA) of the knee is a common and disabling disease worldwide. Its prevalence is increasing in view of the aging population. Changes in collagen content, its orientation and GAG content in the articular cartilage with age are the major features in knee osteoarthritis. These changes in collagen and GAG contents show no manifestation in plain radiography and conventional magnetic resonance imaging (MRI). Nevertheless, early diagnosis of the knee osteoarthritis is of paramount importance clinically in view of the evolution of putative interventions in its early stage. The aim of this project is to identify the relationships between the two imaging biomarkers (i.e. T1 and T2 mappings) and the GAG concentration in living human symptomatic cartilage. Methodology: 28 patients with clinical diagnosis of knee osteoarthritis were enrolled. 7 males and 16 females were recruited and their mean age was 68.1 (ranges from 53 to 84). Conventional PD sequence, T1 and T2 mappings were performed for each subject within 1 week before total knee arthroplasty. Articular cartilage from the lateral tibial plateau was harvested and the GAG content in cartilage was determined by using dimethylmethylene blue method. T1 mean and T2 values were calculated and correlate with GAG concentration statistically. Results: The mean value T1 was 40.3 ± 13.5 ms, ranging from 15.3 to 69.3 ms and the mean value T2 was 31.0 ± 10.5 ms, ranging from 16.1 to 46.9 ms. The mean value of GAG content was 80.1 ± 33.3 mg, ranging from 24.9 to 166.0 mg while the mean value of GAG concentration was 267.4 ± 165.9 mg/cm3 , ranging from 91.3 to 760.5 mg/cm3 . T2 values were inversely correlated with GAG concentration with R2 = 0.375, p = 0.001 while T1 values were also inversely correlated with GAG concentration with R2 = 0.200, p = 0.025. Conclusion: This in vivo study confirmed that T1 and T2 values correlate with the GAG concentration in living human knee cartilages which corroborate with the previous works. The later (T2 values) is found more reliable in our study and less controversial in literatures. We postulate that T2 values can serve as a non-invasive imaging biomarker in the progress of knee osteoarthritis in terms of both disease diagnosis and treatment response monitoring. © 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Osteoarthritis (OA) of the knee is a common and disabling disease worldwide. Its prevalence is increasing in view of the aging population. A local study conducted in 2000 has demonstrated that
∗ Tel.: +852 22553308. E-mail addresses: [email protected] (C.S. Wong), [email protected] (C.H. Yan), [email protected] (N.J. Gong), [email protected] (T. Li), [email protected] (Q. Chan), [email protected] (Y.C. Chu). 0720-048X/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2012.11.036
the prevalence of knee osteoarthritis is 7% and 13% in male and female, respectively [1]. The result is in parallel with the data in United States [2]. Osteoarthritis of the knee is a degenerative process of the joint in which low grade inflammation occurs in the articular hyaline cartilage. This results in abnormal cartilage wearing and thus in turn causes joint pain and movement limitation. Functional loss and daily living deterioration are the unavoidable side effects of this disease. Patients suffering from severe OA knee will have an enormous detrimental impact to their quality of life and the disease is very costly to the medical system as a whole.
648
C.S. Wong et al. / European Journal of Radiology 82 (2013) 647–650
Articular hyaline cartilage is a hierarchical structure which is divided into layers with different amount of type II collagen fibrils oriented in the glycosaminoglycan (GAG) matrix [3]. Changes in collagen content or its orientation and GAG content in the articular cartilage with age are the major features in knee osteoarthritis [4]. Loss of GAG is characteristic in early stage of OA which is followed by alteration in collagen structures [5,6]. These changes in collagen and GAG contents show no manifestation in plain radiography and conventional magnetic resonance imaging (MRI). Nevertheless, early diagnosis of the knee osteoarthritis is of paramount importance clinically in view of the evolution of putative interventions in its early stage [7–9]. Researchers had therefore spent tremendous efforts to explore a non-invasive imaging biomarker to achieve the above mentioned intuitive goal. New and advanced MRI sequences such as T1 or T2 mappings have been widely studied recently for this purpose [10–12]. In T1 sequence, a “spin lock” pulse is applied to the magnetization transversely and T1 relaxation time is thought to be sensitive to protons on macromolecules such as GAG. The drop of GAG content will increase the mobile proton density in bulk water and thus increases the T1 value. Yet, previous literatures in T1 mapping were not straight forward and many of them had demonstrated conflicting results. Menezes et al. and Mlynarik et al. had negated the correlation with T1 with GAG content [13,14] while Nishioka et al. and Duvvuri et al. had made an opposite statement [16,15]. In 2004, Menezes et al. even reported correlation between T1 values and collagen concentration in his in vitro study using three human articular cartilage-bone specimens [13]. Therefore the general consensus has regarded T1 sequence as a research tool with minimal clinical application. T2 parameter is a spin–spin relaxation method which reflects the cartilage water content. It is believed to represent the collagen anisotropy as protons in water molecules are arranged within collagen fibers [17]. However, recent projects had found that the T2 value also correlate well with the GAG content [15,18,24]. Previous works were predominantly in vitro studies and had made a preliminary conclusion that T1 values represent GAG content while T2 values represent collagen content. However, their implications in living human cartilage are not yet completely elucidated. In vivo study concerning this issue is sparse in the literatures and most of the previous works were small sample cohorts. Therefore, we have devised this project to identify the relationships between the two imaging biomarkers (i.e. T1 and T2 mappings) and the GAG concentration in living human symptomatic cartilage. Up to our knowledge, this is by far the study with the biggest cohort in literatures answering this clinical question. 2. Methodology 2.1. Patients Twenty-eight patients with clinical diagnosis of knee osteoarthritis were enrolled. 7 males and 16 females were recruited and their mean age was 68.1 (ranges from 53 to 84). MRI was performed for each subject within 1 week before total knee arthroplasty is carried out (range from 3 to 6 days). Articular cartilage from the lateral tibial plateau was harvested and the GAG content in cartilage was determined by using dimethylmethylene blue method. T1 mean and T2 values were calculated and correlate with GAG concentration statistically. Two patients had both knees scanned and three patients were excluded due to complete cartilaginous denudation in MR images, making measurement impossible. Twenty-five data points were calculated at the end. This study was approved by the institutional review board in Hong Kong and informed consent was obtained from every patient.
2.2. Harvesting of human articular cartilage All patients underwent total knee arthroplasty (TKA) after the MRI scans. The TKA was performed using the Install approach. After the soft tissue release and balancing, the proximal tibial bone cut was performed perpendicular to the mechanical axis, using the extramedullary cutting guide. Approximately 10 mm bone was resected, measuring from the most proximal point of lateral tibia plateau. Attention was paid not to damage the articular cartilage on the lateral side. The removed specimens were immersed in normal saline and delivered to the laboratory immediately. The articular cartilage over the lateral tibial plateau was harvested carefully [19]. All specimens were immersed in normal saline and kept in −30 ◦ C until experiments started. 2.3. GAG content analysis A stable solution of 1,9-dimethylmethylene blue (DMMB) was prepared as follows: DMMB (21 mg) was stirred with 5 ml ethanol; sodium hydroxide (1.72 g) and formic acid (1.5 ml) were added, and the volume made up to 1 l with distilled water. The reagents were stored in bottle sealed completely to avoid light at room temperature [20]. Cartilage sample solutions were prepared by digestion with proteinase K (2 mg) in 30 ml TE buffer (pH 8.0) containing 100 mM NaCl, 10 mM TRIS, 1 mM EDTA at 60 ◦ C. Sample solutions were diluted with TE buffer to adjust GAG concentration to a proper value. Cartilage sample solution (40 l) was pipetted into a 96 wells plate. DMMB solution (200 l) was added with good mixing. 40 l of chondroitin sulfate A of different concentration (0, 8, 16, 32, 64, 128 g/ml) was mixed with 200 l DMMB solution to prepare the standard curve. Absorbance was read at 595 nm and the concentrations of the GAG content in the samples were estimated based on the chondroitin sulfate A standard curve. 2.4. MRI sequence All MRI examinations were performed using a 3T Achieva scanner (Philips Healthcare, Best, The Netherlands) and an eight channel SENSE knee coil. The knee was positioned at 15◦ flexion and in a neutral position of rotation so as to minimize the magic angle artifact. Conventional PD sagittal and coronal scan were performed. Coronal T1 weighted images were obtained with a 3D balanced turbo field echo (B-TFE) sequence with the following parameters: TR/TE = 5.0 ms/2.3 ms, flip angle = 50◦ , FOV = 140 mm × 140 mm, slice thickness = 4 mm. Five subsequent T1 weighted scans were performed with spin lock durations 0, 10, 20, 30, 40 ms and with scan time of 2 min 33 s for each scan. Images were fitted on a pixel by pixel basis to the exponentially decaying T1 function using IDL (Research Systems, Inc., USA) to generate T1 relaxation map. Quantitative T2 measurement in identical geometry as that of the T1 image was performed using standard multi-echo spin-echo sequence with TE = 16, 32, 48, 64, 80, 96, 112 ms. Scan time = 11 min 25 s. Image analyses were conducted by using Image J software (NIH, Bethesda, MD). Regions of interest (ROIs) were drawn manually by two specialist radiologists (CSW, YCC) on the T1 and T2 mappings with reference to the PD images (Fig. 1). To calculate the volume of each cartilage, the volume corresponding to every drawn ROI was first computed. Slice thickness plus slice gap was regarded as apparent slice thickness. Products of area included in ROI and apparent slice thickness were calculated as volume of ROI. Then, sum of all the ROIs was calculated as volume of cartilage. To be more accurate, the final volume was an averaged volume of measurements from T2 and T1 images. To account for the volume effect in GAG measurement, we calculated
C.S. Wong et al. / European Journal of Radiology 82 (2013) 647–650
649
Fig. 2. The correlation between T2 and extracellular matrix composition: T2 values were negatively correlated with GAG concentration (R2 = 0.375, p = 0.001).
3. Results and discussion The mean value T1 was 40.3 ± 13.5 ms, ranging from 15.3 to 69.3 ms and the mean value T2 was 31.0 ± 10.5 ms, ranging from 16.1 to 46.9 ms. The mean of GAG content was 80.1 ± 33.3 mg, ranging from 24.9 to 166.0 mg while the mean of GAG concentration was 267.4 ± 165.9 mg/cm3 , ranging from 91.3 to 760.5 mg/cm3 . T2 values were inversely correlated with GAG concentration with R2 = 0.375, p = 0.001 (Fig. 2) while T1 values were also inversely correlated with GAG concentration with R2 = 0.200, p = 0.025 (Fig. 3).
Fig. 1. Knees were imaged using a 3T Achieva scanner. T1 relaxation time map was generated by fitting on a pixel by pixel basis using IDL (Research Systems, Inc., USA). Quantitative T2 measurement was performed using standard multi-echo spin-echo sequence. (a) PD coronal, (b) corresponding T2 and (c) corresponding T1 images.
GAG concentration of each cartilage. It was calculated by dividing the total GAG content with cartilage volume. 2.5. Data analysis and statistics Associations between T1, T2 and GAG measurements were tested using Pearson correlation. All p values were two tailed and 0.05 was chosen as significant level. All statistical analyses were performed using SPSS software (v.19.0.0; Chicago, IL).
Fig. 3. The correlation between T1 and extracellular matrix composition: T1 values were negatively correlated with GAG concentration (R2 = 0.200, p = 0.025).
650
C.S. Wong et al. / European Journal of Radiology 82 (2013) 647–650
Our results demonstrated that the T1 and T2 values significantly and inversely correlate with the GAG concentration (R2 = 0.200, p = 0.025; R2 = 0.375, p = 0.001, respectively). The result about the T1 corroborates with one end of the pendulum. For instance, Duvvuri et al. had successfully drawn the conclusion that T1 correlate negatively with the GAG content by using bovine cartilage [15] while Meneszes et al. found no such correlation by using human specimen [17]. These conflicting results can be partly due to the different methods used for the GAG measurement and the potential errors in the dilution process. Another possibility would be the different enzymes used in different groups to degrade the cartilage and thus initiate controversies. Moreover, most previous works were conducted in in vitro fashions with controlled environment and thus application of their results in living human condition is questionable, not to mention that most works were performed with small sample cohorts. Our results in this in vivo study support the school of thought that T1 correlates inversely with the GAG concentration and coherent with the in vivo study by Nishioka et al. [16] in 2011 but with a weaker correlation (R2 = 0.200 vs R2 = 0.410). One reason for that is the difference in methodology. They collected various samples within the same subject and that could induce a sampling error as GAG content and its concentration can vary significantly even within the same block of cartilage. Therefore we tried to eliminate this error by measuring the entire block of cartilage as a whole. Also, the recruited subjects in our study were all in a late stage which T1 value is not only affected by loss of GAG as in early stage but also by other types of macromolecule change such as collagen loss. We have tried to lessen this effect by harvesting the non-weight bearing side of cartilage which is less worn out. The relationship between T2 value and GAG concentration has remained unclear in literatures. Proteoglycan or collagen depletion would increase the free space inside the extra-cellular matrix in cartilage and thus increase the influx of water and therefore T2 value will increase [21]. Therefore T2 value has been considered representing not only collagen anisotropy but also GAG concentration. Our results on T2 values found significant inverse correlation with the cartilaginous GAG concentration which reiterates the current literatures. The correlation is stronger than that of the T1 values (T1: R2 = 0.200; T2: R2 = 0.375). T2 value has been well described in its correlation with collagen content and with that of GAG content has also been revealed recently. Unlike T1, none of them negate the correlation between T2 value and GAG concentration, and therefore, T2 value is thought to be a less controversial imaging biomarker in knee cartilaginous degeneration. Moreover, less scan time is required for T2 mapping and its post processing technique is more robust and well established. All these would render it a more feasible imaging biomarker in clinical application. There are several limitations of our study which need to be addressed. The high exclusion rate (n = 5) would limit the sample size and then hence the results’ accuracy. Moreover, T2 values can be affected by magic angle artifact [22]. Although the magic angle artifact is thought to be small in in vivo situation [23], care has already been executed so that the articular surface was positioned perpendicular to the magnetic field to minimize the artifact. Moreover, as the entire knee cartilage (lateral tibial plateau) was assessed, zonal delineation of the cartilage can be performed to obtain a more detailed structural categorization. Lastly, late stage diseases were recruited but specimens usually cannot be obtained from early disease stage in view of the ethical issue. 4. Conclusion This in vivo study confirmed that T1 and T2 values correlate with the GAG concentration in living human knee cartilages which
corroborate with the previous works. The later (T2 values) is found more reliable in our study and less controversial in literatures. We postulate that T2 values can serve as a non-invasive imaging biomarker in the progress of knee osteoarthritis in terms of both disease diagnosis and treatment response monitoring. This result laid a foundation on further related studies with a larger cohort. References [1] Lau EC, Cooper C, Lam D, et al. Factors associated with osteoarthritis of the hip and knee in Hong Kong Chinese: obesity, joint injury, and occupational activities. American Journal of Epidemiology 2000;152(November (9)):855–62. [2] Dillon CF, Rasch EK, Gu Q, et al. Prevalence of knee osteoarthritis in the United States: arthritis data from the Third National Health and Nutritional Examination Survey 1991–94. Journal of Rheumatology 2006;33:2271–9. [3] Xia Y, Moody JB, Alhadlaq H. Orientational dependence of T2 relaxation in articular cartilage: a microscopic MRI (microMRI) study. Magnetic Resonance in Medicine 2002;48:460–9. [4] Arokoski JP, Jurvelin JS, Väätäinen U, et al. Normal and pathological adaptations of articular cartilage to joint loading. Scandinavian Journal of Medicine and Science in Sports 2000;10:186–98. [5] Burstein D, Bashir A, Gray ML. MRI techniques in early stages of cartilage disease. Investigative Radiology 2000;35:622–38. [6] Buckwalter JA, Martin J. Degenerative joint disease. Clinical Symposia 1995;47:1–32. [7] Sarzi-Puttini P, Cimmino MA, Scarpa R, et al. Osteoarthritis: an overview of the disease and its treatment strategies. Seminars in Arthritis and Rheumatism 2005;35(Suppl. 1):1–10. [8] Eshed I, Trattnig S, Sharon M, et al. Assessment of cartilage repair after chondrocyte transplantation with a fibrin–hyaluronan matrix—correlation of morphological MRI, biochemical T2 mapping and clinical outcome. European Journal of Radiology 2012;81(June (6)):1216–23. [9] Kon E, Di Martino A, Filardo G, et al. Second-generation autologous chondrocytetransplantation: MRI findings and clinical correlations at a minimum 5-year follow-up. European Journal of Radiology 2011;Sep 79(3):382–8, http://dx.doi.org/10.1016/j.ejrad.2010.04.002. [10] Zilkens C, Miese F, Kim YJ, et al. Three-dimensional delayed gadoliniumenhanced magnetic resonance imaging of hip joint cartilage at 3T: a prospective controlled study. European Journal of Radiology 2012;81(November (11)):3420–5, http://dx.doi.org/10.1016/j.ejrad.2012.04.008. [11] Juras V, Apprich S, Pressl C, et al. Histological correlation of 7T multi-parametric MRI performed in ex-vivo Achilles tendon. European Journal of Radiology 2011 [Epub ahead of print]. [12] Zilkens C, Miese F, Bittersohl B, et al. Delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC), after slipped capital femoral epiphysis. European Journal of Radiology 2011;79(3):400–6 [Epub 2010 May 26]. [13] Menezes NM, Gray ML, Hartke JR, et al. T2 and T1 MRI in articular cartilage systems. Magnetic Resonance in Medicine 2004;51:503–9. [14] Mlynarik V, Trattnig S, Huber M, et al. The role of relaxation times in monitoring proteoglycan depletion in articular cartilage. Journal of Magnetic Resonance Imaging 1999;10:497–502. [15] Duvvuri U, Reddy R, Patel SD, et al. T1 rho-relaxation in articular cartilage: effects of enzymatic degradation. Magnetic Resonance in Medicine 1997;38:863–7. [16] Nishioka H, Hirose J, Nakamura E, et al. T1 rho and T2 mapping reveal the in vivo extracellular matrix of articular cartilage. Journal of Magnetic Resonance Imaging 2012;35:147–55. [17] Nieminen MT, Töyräs J, Rieppo J, et al. Quantitative MR microscopy of enzymatically degraded articular cartilage. Magnetic Resonance in Medicine 2000;43:676–81. [18] Astrid Watrin P, Ruaud JP, Oliver P, et al. Effect of proteoglycan depletion on T2 mapping in rat patellar cartilage. Radiology 2005;234:162–70. [19] Altman R, Asch E, Bloch D, et al. Development of criteria for the classification and reporting of osteoarthritis classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis and Rheumatism 1986;29(8):1039–49. [20] Juras V, Bittsansky M, Majdisova Z, et al. In vitro determination of biomechanical properties of human articular cartilage in osteoarthritis using multi-parametric MRI. Journal of Magnetic Resonance 2009;197(1):40–7. [21] Xia Y. Magic angle effect in magnetic resonance imaging of articular cartilage: review. Investigative Radiology 2000;35:602–21. [22] Akella SV, Regatte RR, Wheaton AJ, et al. Reduction of residual dipolar interaction in cartilage by spin lock technique. Magnetic Resonance in Medicine 2004;52:1103–9. [23] Mosher TJ, Smith H, Dardzinski BJ, et al. Imaging and T2 mapping of femoral cartilage: in vivo determination of the magic angle effect. American Journal of Roentgenology 2001;177:665–9. [24] Keenan KE, Besier TF, Pauly JM, et al. Prediction of glycosaminoglycan content in human cartilage by age T1 rho and T2 MRI. Osteoarthritis and Cartilage 2011;19:171–9.