Detecting Early Changes in ACL-Reconstructed Knee Cartilage: Cluster Analysis of T2 Relaxation Times in Superficial and Deep Cartilage and ADC Analysis
Marianne Black1,2, Daehyun Yoon2, Kate Young2, Akshay S. Chaudhari2, Feliks Kogan3, Garry Evan Gold2,4,5, Marc Elliot Levenston1,2,5, and Brian Hargreaves2,5,6

1Mechanical Engineering, Stanford University, Stanford, CA, United States, 2Radiology, Stanford University, Stanford, CA, United States, 3Stanford University, Stanford, CA, United States, 4Orthopedic Surgery, Stanford University, Stanford, CA, United States, 5Bioengineering, Stanford University, Stanford, CA, United States, 6Electrical Engineering, Stanford University, Stanford, CA, United States


ACL-injured subjects are at an increased risk of developing osteoarthritis. There is a need to detect early osteoarthritic changes for the development of treatments that can slow or stop osteoarthritis progression. T2 and ADC are considered reflective of the structure and composition of cartilage, and may be valuable for detecting early osteoarthritis. This study used two qDESS acquisitions to obtain T2 and ADC maps in 10 ACL-reconstructed subjects and 10 controls 3-weeks, 3-months and 9-months post-surgery. Our results show that T2 cluster analysis was able to detect changes to the ACL-reconstructed knee as early as 3-months post-surgery.


Individuals who have experienced anterior cruciate ligament (ACL) tears have a significantly elevated risk of developing osteoarthritis, whether surgically reconstructed or not1,2. Quantitative MRI measures such as T2 relaxation time, considered reflective of collagen content and organization, have been used to measure changes in cartilage following ACL-injury3. Cluster analysis of cartilage T2, a method that involves projecting the 3D cartilage surface onto a 2D plane to detect locally elevated T2 clusters in cartilage, has been shown to detect differences between healthy and ACL-injured knees as early as 6 months post-surgery4. Superficial and deep layers of cartilage exhibit different properties and are affected differently during OA development; thus, separate analysis of both layers may allow for additional sensitivity to changes that may otherwise be undetected using full-thickness cartilage analysis. Apparent Diffusion Coefficient (ADC) measurements in cartilage have been correlated to proteoglycan loss5 and may provide additional information of cartilage health in early osteoarthritis. The objective of this study was to evaluate (1) if there are changes in deep and superficial cartilage following ACL-reconstruction surgery using cluster analysis to detect regions of elevated T2 and (2) if there are changes to ADC values following ACL-reconstruction surgery.


Twenty subjects (30 knees) were included in this study: 10 ACL-injured subjects undergoing ACL-reconstruction surgery (5 women, 5 men, mean age: 39±12 years, BMI: 23 ± 1.5) and 10 age, BMI and sex matched controls (5 women, 5 men, mean age: 37±13 years, BMI: 23 ± 1.5). The affected and contralateral knees of the ACL-reconstructed subjects and the right knee of the control subjects were scanned in a 3T MRI scanner using 2 quantitative double-echo in steady-state (qDESS with low and high diffusion sensitivity) sequences (10-minute scan-time per knee) at 3 time-points: 3-weeks post-ACL reconstruction (baseline), 3-months post-ACL reconstruction and 9-months post-ACL-reconstruction.

T2 and ADC values were calculated from the qDESS sequences as previously described6,7. The femoral cartilage was manually segmented in the sagittal plane (Fig.1a). Femoral cartilage T2 and ADC projections were created by fitting a cylinder to the segmentations and radially projecting the data into angular bins4 (Fig.1b). T2 maps were separated into superficial and deep layers (Fig.1c) based on the midpoint of thickness. Femoral cartilage T2 projections were registered to the baseline time point for each knee (Elastix)8, and difference maps were created by subtracting the baseline projection from each of the 3-month and 9-month projections (for both superficial and deep T2 maps) (Fig.1d). Clusters in difference maps were quantified as a contiguous set of pixels with an area greater than 12.4mm2 consisting of values greater than twice the standard deviation of control subjects’ difference maps (Fig.1e)4. ADC values were averaged over the femoral cartilage at each time-point (Fig.2). Our outcomes were reported as the change in T2 percent cluster area (ΔT2%CA) and average ADC. We used a general linear model with Bonferroni’s correction to test for differences in ΔT2%CA and ADC values between ACL-reconstructed, contralateral, and control knees (α<0.05).


The ΔT2%CA in superficial and deep regions were significantly higher for ACL-reconstructed knees compared to both the contralateral and control knees in femoral cartilage (p<0.001) (Fig.3&4). The contralateral knee and control knee were not significantly different in ΔT2%CA (p=0.507). There were no significant differences between the 3-month and 9-month ΔT2%CA (p=0.207). However, there was a notable increase in deep cartilage ΔT2%CA from 3 to 9-months for the injured knee (Fig.3), as opposed to superficial cartilage ΔT2%CA which remained relatively constant (Fig.4). ADC values were significantly higher for contralateral knees compared to the control knees (p=0.012). Although not significant, the ADC values for injured knees from 3-months to 9-months saw a noticeable increase (Fig.5).


The higher ΔT2%CA for ACL-reconstructed compared to the contralateral and control knees is indicative of early cartilage changes occurring following ACL-reconstruction surgery. The ΔT2%CA is representative of local, elevated T2 and could represent specific areas within cartilage where degeneration may be first occurring. The trend of higher superficial ΔT2%CA at 3-months may indicate that the superficial cartilage is experiencing early changes, whereas by 9-months ΔT2%CA is very similar for deep and superficial cartilage. Long-term follow-up is needed to understand what happens to cartilage that experienced early ΔT2%CA elevations. ADC increases may be an effect of the contralateral knee becoming the primary load-bearing knee while patients recover, potentially causing cartilage changes. However, more study is necessary to understand ADC changes.


ΔT2%CA, acquired using a fast, 3D scan, can detect changes in the ACL-reconstructed within months post-surgery, which speaks to the promising potential of this method for detecting early osteoarthritis degeneration.


NIH R01 AR0063643, NIH R01 EB002524, NIH K24 AR062068, GE Healthcare.


[1] D. Simon, R. Mascarenhas, B.M. Saltzman, M. Rollins, B. R. Bach Jr., and P.MacDonald, “The Relationship between Anterior Cruciate Ligament Injury and Osteoarthritis of the Knee,” Advances in Orthopedics, vol. 2015, Article ID 928301, 11 pages, 2015.

[2] D.E. Meuffels, M.M. Favejee, M.M. Vissers, et al, “Ten year follow-up study comparing conservative versus operative treatment of anterior cruciate ligament ruptures. A matched-pair analysis of high level athletes,” British Journal of Sports Medicine 2009;43:347-351.

[3] X. Li, D. Kuo, A. Theologis, J. Carballido-Gamio, C. Stehling, T. M. Link, B. Ma, and S. Majumdar. “Cartilage in Anterior Cruciate Ligament–Reconstructed Knees: MR Imaging T1ρ and T2—Initial Experience with 1-year Follow-up,” Radiology 2011 258:2, 505-514

[4] U.D. Monu, C.D. Jordan, B.L. Samuelson, B.A. Hargreaves, G.E. Gold, E.J. McWalter, “Cluster analysis of quantitative MRI T2 and T1ρ relaxation times of cartilage identifies differences between healthy and ACL-injured individuals at 3T”, Osteoarthritis and Cartilage, Volume 25, Issue 4, 2017, Pages 513-520, ISSN 1063-4584.

[5] J.G. Raya, G. Melkus, S. Adam-Neumair, et al. Change of Diffusion Tensor Imaging Parameters in Articular Cartilage With Progressive Proteoglycan Extraction. Investigative Radiology 2011;46:401.

[6] B. Sveinsson, A.S. Chaudhari, G.E. Gold, B.A. Hargreaves, “A simple analytic method for estimating T2 in the knee from DESS”, Magnetic Resonance Imaging, Volume 38, 2017, Pages 63-70, ISSN 0730-725X.

[7] B. Sveinsson, A.S. Chaudhari, G.E. Gold, B.A. Hargreaves, “A simple analytic method for estimating T2 in the knee from DESS”, Magnetic Resonance Imaging, Volume 38, 2017, Pages 63-70, ISSN 0730-725X.

[8] S. Klein, M. Staring, K. Murphy, M.A. Viergever and J.P.W. Pluim, "elastix: a toolbox for intensity based medical image registration," IEEE Transactions on Medical Imaging, vol. 29, no. 1, pp. 196 - 205, January 2010.


Figure 1: (a) Cartilage is segmented and (b) a cylinder is fit to femoral cartilage to allow for unwrapping and projecting the femoral cartilage map. (c) Cartilage is divided into superficial and deep layers at each time point. (d) Difference maps are made by subtracting the baseline (3-week post-surgery scan) from the 3- and 9-month post-surgery timepoint. (e) Thresholds are applied to reveal clusters of elevated T2 relaxation times, and %CA is calculated as the percentage of the femoral cartilage that is covered by clusters.

Figure 2: Projected femoral cartilage ADC maps for an ACL-reconstructed knee show regional variation in ADC values. The ADC values remain relatively consistent over the time points, with moderate increases noticed.

Figure 3: Deep femoral cartilage ΔT2%CA are significantly higher for the injured knee, and notably it appears there is a trend towards an increased ΔT2%CA in the deep femoral cartilage from 3 to 9-months for the injured knee.

Figure 4: Superficial femoral cartilage ΔT2%CA are significantly higher for the injured knee, and remain relatively consistent from 3 to 9-months for the injured knee.

Figure 5: Contralateral femoral cartilage ADC values were significantly higher than control femoral cartilage. Of note, the average ADC values of the control knees remain very consistent over time, whereas the injured (ACL-reconstructed) knees’ ADC values trend towards an increase at the 9-month time point for the injured knee.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)