Magnetic resonance imaging for non-invasive clinical evaluation of normal and regenerated cartilage

Abstract

With the development of tissue engineering and regenerative medicine, it is much desired to establish bioimaging techniques to monitor the real-time regeneration efficacy in vivo in a non-invasive way. Herein, we tried magnetic resonance imaging (MRI) to evaluate knee cartilage regeneration after implanting a biomaterial scaffold seeded with chondrocytes, namely, matrix-induced autologous chondrocyte implantation (MACI). After summary of the T2 mapping and the T1-related delayed gadolinium-enhanced MRI imaging of cartilage (dGEMRIC) in vitro and in vivo in the literature, these two MRI techniques were tried clinically. In this study, 18 patients were followed up for 1 year. It was found that there was a significant difference between the regeneration site and the neighboring normal site (control), and the difference gradually diminished with regeneration time up to 1 year according to both the quantitative T1 and T2 MRI methods. We further established the correlation between the quantitative evaluation of MRI and the clinical Lysholm scores for the first time. Hence, the MRI technique was confirmed to be a feasible semi-quantitative yet non-invasive way to evaluate the in vivo regeneration of knee articular cartilage.

Keywords: T2 mapping; cartilage regeneration; delayed gadolinium-enhanced MRI imaging; magnetic resonance imaging; tissue engineering.

Figure 1.

Principles of magnetic resonance imaging of regenerative cartilage. (A) The NMR phenomenon appears when a system of nuclei in a static magnetic field experiences a radiofrequency pulse (RF). (B) Under the action of a 90° RF pulse with Larmor frequency, the magnetization vector is rotated from the z-axis to the xy-plane. After the RF pulse is removed, longitudinal magnetization (M_z) is restored owing to spin–lattice relaxation, and transversal magnetization (M_xy) is decayed owing to spin–spin relaxation. (C) The longitudinal relaxation time T1 represents the recovery of M_z to 63%, and the transversal relaxation time T2 represents the decay of the M_xy to 37%. (D) Correlation between the content of dry components in cartilage and the relaxation time T2. The richest dry component in cartilage is collagen. Data are used and replotted from Lüsse et al.[5] permitted by Elsevier Ltd with copyright 2001. The line comes from linear fitting. (E) Correlation of ΔR1 (△R1 = 1/T1post − 1/T1pre) and glycosaminoglycan (GAG). Excised human cartilage is obtained after total knee and hip replacement surgery. Data are used and replotted from Bursturn et al. [6] permitted by John Wiley and Sons with copyright 1999. The line comes from linear fitting. (F) Schematic presentation of the main dry components of cartilage (collagen and GAG) and change of magnetic resonance signals with tissue regeneration. Along with cartilage regeneration, the maturation of the collagen network leads to, albeit the decrease of the total water content (both free and bound), the increase of bound water and thus the decrease of T2; meanwhile, the increase of GAG in the regenerated cartilage leads to the decrease of the penetration of the negatively charged contrast agent Ga-DTPA²⁻ into the tissue and thus the decrease of ΔR1

Figure 2.

Typical illustration of MACI operation for cartilage regeneration and MRI imaging. (A) Sagittal and transverse proton density-weighted image before surgery. A femoral trochlear cartilage in the right knee joint of a male patient is demonstrated; the arrows indicate the site of cartilage damage. (B) The process of MACI with combination of an ECM-derived scaffold and autologous chondrocytes. The arthroscopic surgery aims at assessing the site of injury and collecting autologous cartilage tissue from the non-weight-bearing area. After the proliferation of cells in vitro, the cells were implanted into the biomimetic cartilage scaffold at a concentration of 1 × 10⁷ cells per milliliter. About 24 h after the cells were loaded into the scaffolds, the tissue-engineered construct was transplanted to the cartilage damage area through stage II surgery to regenerate the cartilage. (C) T1 (left) and T2 (right) map images after 3 months

Figure 3.

T2 mapping at the indicated follow-up times. (A) Sagittal proton density (PD) weighted, T2 mapping with ROI and merged (T2 fused) images of a patient at 3, 6 and 12 months after MACI. Before the tissue engineering treatment, the male patient experienced a femoral trochlear cartilage injury in the right knee joint. (B) Schematic diagram of regeneration site and control (normal) site of cartilage. (C) The line picture shows a clear decrease of T2 relaxation time in the regenerated tissue at 3 and 6 months, and similar T2 values between the regenerated tissue and control normal cartilage at 12 months. The statistics were performed for 25 lesions based on 18 patients in sequence of Table 2. The data are shown as mean ± standard deviation and treated by one-way ANOVA analysis. It is considered to have a significant difference when the P value is less than 0.05. The differences are marked ‘***’ in the cases of P < 0.001

Figure 4.

(A) T1 maps before and after injection of the contrast agent Gd-DTPA²⁻ in the same patient as in Fig. 3 at 3, 6 and 12 months after MACI. (B) The T1pre and T1post relaxation times of regenerated tissue and control cartilage at 3, 6 and 12 months after MACI. (C) The line picture shows a decrease of ΔR1 of regenerated tissue with time. The data are shown as mean ± standard deviation and treated by one-way ANOVA analysis. The differences are marked ‘***’ for P < 0.001 and ‘*’ for 0.01 < P < 0.05

Figure 5.

T2 values and ΔR1 values at different sites in regenerated cartilage and control group in 3, 6 and 12 months. All sites of cartilage had regeneration effects over time, and there was no significant difference between different sites of the cartilage. The differences are marked as ‘*’ for 0.01 < P < 0.05, ‘**’ for 0.001 < P < 0.01 and ‘***’ for P < 0.001. For lateral femoral condyle and medial femoral condyle, n = 4; for patella cartilage, n = 10; for trochlear cartilage, n = 7

Figure 6.

T2 values and corresponding auto-ratio of T2 over the neighbor control. Auto-ratio value means the ratio of the T2 value of the regenerated cartilage to the T2 value of adjacent normal cartilage on the same patient and site. The value of T2/T2_control approaches to unity along with cartilage regeneration. There were 25 lesions based on 18 patients. The P value between any two of the three sets of data is less than 0.001, which is marked as ‘***’

Figure 7.

T1-pertinent results and the reduction of the regenerated sites over the corresponding neighbor healthy control in the same patients and sites. ΔR1 and T1 values before and after injection of the contrast agent in the same patient at 3, 6 and 12 months after MACI (left). These values deviated more from normal in the early stage of cartilage regeneration (3 M) and are close to normal in 12 M. The difference of ΔR1 value is the most obvious in 3 M. Statistical diagrams of auto-ratio values indicate that the ratio of the regenerated tissue to the adjacent normal tissue can be used to evaluate the regeneration effect. There were 25 lesions based on 18 patients. The data are shown as mean ± standard deviation and treated by one-way ANOVA analysis. It is considered to have a significant difference when the P value is less than 0.05. The differences are marked ‘***’ for P < 0.001, ‘**’ for 0.001 < P < 0.01 and ‘*’ for 0.01 < P < 0.05

Figure 8.

The relationship between the auto-ratios to reflect the MRI evaluation of cartilage regeneration and the Lysholm scores to reflect the global clinical performance. The cases are from Supplementary Table S2. With the regeneration of cartilage, the Lysholm score keeps rising and the auto-ratio values gradually decrease and tend to unity. The lines in the bottom figure come from linear fitting of clinical Lysholm scores and auto-ratio data for different patients