DOI: 10.3928/01477447-20110627-10
Abstract
Magnetic resonance imaging (MRI) and arthroscopy have frequently been used to evaluate articular cartilage. Many studies have compared the accuracy of MRI to that of arthroscopy. However, there have been no previous comparison studies between MRI and arthroscopy in the evaluation of repaired cartilage after autologous chondrocyte implantation using the Magnetic Resonance Observation of Cartilage Repair Tissue (MOCART) scoring system. The purpose of this study was to compare the results between MRI and arthroscopy after autologous chondrocyte implantation of an osteochondral lesion of the talus using a modified MOCART scoring system.
Our study investigated 27 consecutive cases in 26 patients who underwent follow-up MRI and second-look arthroscopy 1 year following autologous chondrocyte implantation based on their osteochondral lesion of the talus diagnosis. According to the comparison results of those 5 categories, the agreement between MRI and arthroscopy evaluation results was statistically significant with good reliability in the categories of the degree of defect repair and defect filling, the quality of repaired tissue surface, and synovitis. However, the integration with the border zone and the adhesion category showed poor to moderate reliability. There has been no well-established correlation method between arthroscopy and MRI after autologous chondrocyte implantation of an osteochondral lesion of the talus.
Dr Lee (Kyung Tai) is from the Foot and Ankle Clinic, KT Lee’s Orthopedic Hospital, Dr Choi is from the Department of Radiology, Eulji Hospital, Eulji University School of Medicine, Seoul, Drs Lee (Young Koo) and Koo are from the Department of Orthopedic Surgery, Soonchunhyang University, Bucheon Hospital, Gyeonggi-Do, and Dr Cha is from the Department of Orthopedic Surgery, Kwandong University Hwajung dong, Dukyang-Gu, Koyang-Si, Gyeonggi-Do, Republic of Korea.
Drs Lee (Kyung Tai), Choi, Lee (Young Koo), Cha, and Koo have no relevant financial relationships to disclose.
Correspondence should be addressed to: Young Koo Lee, MD, Department of Orthopedic Surgery, Soonchunhyang University 4 Jung-Dong, Wonmi-Gu, Bucheon-Si, Gyeonggi-Do, 420-767, Republic of Korea (brain0808@hanmail.net).
Magnetic resonance imaging (MRI) and arthroscopy frequently have been used to evaluate articular cartilage. 1,2 Many studies have compared the accuracy of MRI to that of arthroscopy. With regard to the knee joint, O’Connor et al 3 reported the accuracy of MRI compared to arthroscopy for osteochondritis dissecans was 85%. For the ankle joint, Mintz et al 2 reported the accuracy of MRI of osteochondral lesions of the talus was 83% and Lee et al 1 reported an accuracy of 81%. However, none of the prior studies were based on the evaluation results of repaired cartilage after autologous chondrocyte implantation of an osteochondral lesion of the talus.
In addition, there have been no previous comparison studies between MRI and arthroscopy in the evaluation of repaired cartilage after autologous chondrocyte implantation using the Magnetic Resonance Observation of Cartilage Repair Tissue (MOCART) scoring system. The MOCART scoring system is an evaluation method for repaired cartilage with low interobserver variability. 4 It is a useful method for long-term follow-up as it is based on a scoring system. 5 However, previous MOCART-based studies have dealt only with the knee joint; none of the MOCART studies have examined the ankle joint. The purpose of this study was to compare the results between MRI and arthroscopy after autologous chondrocyte implantation of an osteochondral lesion of the talus using a modified MOCART scoring system.
Materials and Methods
This study was performed from September 2005 to March 2008 and included 27 cases in 26 consecutive patients who received a follow-up MRI and second-look arthroscopy 1 year after autologous chondrocyte implantation caused by an osteochondral lesion of the talus. Average patient age was 33.9 years (range, 16–56 years), and the study population comprised 19 men and 7 women. A total of 13 right and 14 left ankles were examined. The study was approved by our Institutional Research Board, and all study participants provided informed consent prior to study enrollment.
All patients were diagnosed based on MRI and physical examinations, and the diagnosis was confirmed during surgery using arthroscopy. Approximately 1 year following their initial surgery, patients underwent MRI and second-look arthroscopy. A modified MOCART scoring system was used to classify MRI and arthroscopy evaluation of the osteochondral lesion of the talus. Five categories of the MOCART scoring system that could be applied to both MRI and arthroscopy were used with some modification to compare the results of the 1-year follow-up MRI and second-look arthroscopy (Table 1).
 | Modified MOCART Scoring System for the Evaluation of Autologous Chondrocyte Implantation |
Surgical Technique
During the first surgical stage, a cartilage biopsy was performed after a local anesthesia block was administered to the foot and ankle. Autologous chondrocyte implantation, the second-stage surgery, was performed under spinal anesthesia, with patients placed in a frog-leg position. An incision was made along the midline of the medial malleolous, which exposed approximately 10 cm of the posterior tibial tendon and anterior ankle joint. Oblique medial malleolar osteotomy was performed under fluoroscopy.
The osteochondral lesion of the talus then was exposed by retracting the distal malleolar fragment. The lesion was removed, and a 2.5-mm drill was used to prevent delamination of the gel matrix. The area was injected with Chondron (Sewoncellontech Corp, Seoul, Korea), which hardened within 5 minutes given its properties to convert the fibrin-mixed chondrocyte liquid mixture into the gel matrix within the lesion. Through a pre-drilled hole, the osteotomy site was fixed using a 4.5-mm diameter cannulated screw and a K-wire without a rotational deformation. The ankle was immobilized at 90° using a plaster splint, and nonweight-bearing measures were initiated.
The third-stage surgery was performed under spinal anesthesia. Patients were placed in a kneeling position to retract the ankle using an 8-lb weight to check the joint condition through the anteromedial and anterolateral portals. Patients then were placed in a supine position with their legs on the table, and the previous incision scar was incised to remove the screw.
Imaging Technique
Magnetic resonance imaging was performed 1 year postoperatively using a 1.5-Tesla unit (Twinspeed; General Electric Health Care, Milwaukee, Wisconsin) and an extremity coil. Sagittal inversion recovery images were obtained with a 16-cm field of view, repetition time of 5000 milliseconds, echo time of 16 ms, inversion time of 150 milliseconds, echo train length 8, and slice thickness of 3.5 mm with no interslice gap. Sagittal, axial, and coronal intermediate-weighted fast-spin echo images were obtained with a repetition time of 4000 milliseconds, effective echo time of 25 to 26 milliseconds, echo train length 8, slice thickness of 3 mm with no inter-slice gap, and field view of 11 to 15 cm. A matrix of 512×256 was obtained with a number of excitations of 1 to 2. Coronal fat-suppressed 3-dimensional spoiled gradient-recalled images were added with a 12-cm field of view, repetition time of 40 milliseconds, echo time of 6 milliseconds, flip angle of 40°, slice thickness of 1.5 mm with no interslice gap, and matrix of 256×192. An experienced musculoskeletal radiologist assessed all MRIs and was unaware of the second-look arthroscopic findings.
Statistical Analysis
Statistical analysis was performed using the modified MOCART scoring system under the hypothesis that there is no difference between MRI and arthroscopy and under the alternative hypothesis that there is a difference between MRI and arthroscopy, to determine any difference between MRI and arthroscopy. To assess agreement and reliability, the intraclass correlation coefficients (ICC) were obtained from random effects 1-way analysis of variance. Intraclass correlation coefficient values close to zero indicated no interrater reliability, and ICC values close to 1 indicated perfect reliability. All analyses were performed using SPSS version 12.0 (SPSS Institute, Chicago, Illinois), and all significance tests were 2-tailed. For all tests, a P value of <.05 was considered statistically significant.
Results
Among the 5 categories of the MO-CART scoring system that are applicable to both second-look arthroscopy and MRI, the degree of defect repair and filling of the defect category (Table 2) showed congruent results in 16 of 27 cases. Of the 14 cases in which arthroscopy showed complete manifestations, MRI showed complete manifestations in 12 cases (Figure 1) and >50% of the adjacent cartilage filling manifestations in 2 cases. However, in 5 cases in which arthroscopy showed <50% of the adjacent cartilage filling, MRI showed >50%, indicating complete disagreement. Therefore, the degree of defect repair and filling of the defect category showed a relatively high correlation, resulting in an ICC value of 0.7222 ( P<.0001).
 | Degree of Defect Repair and Filling Category |
 | Coronal (A) and sagittal (B) intermediate-weighted fast-spin echo MRIs of a 16-year-old girl 1 year after autologous chondrocyte implantation of the medial talar dome show complete filling of the defect of the osteochondral lesion (white arrows) and complete integration. The repaired tissue surface is relatively smooth and synovitis (black arrows) is seen. Arthroscopic photograph shows complete defect filling with a smooth surface over the talus. In this case, MRI demonstrates a good correlation with arthroscopy (C). |
In the integration into the border zone category (Table 3), arthroscopy showed complete manifestations in 18 cases; MRI results were in agreement in 17 cases, with 1 case displaying a demarcating border. Nevertheless, of the 5 cases in which arthroscopy showed visible integration defects, MRI displayed incongruent results in 3 cases (Figure 2). Therefore, in this category, there was a tendency to disagree, with an ICC value of 0.48 ( P=.0034).
 | Integration Into the Border Zone |
 | Sagittal (A) and coronal (B) intermediate-weighted fast-spin echo and coronal fat-suppressed 3D-spoiled gradient-recalled MRIs (C) of a 39-year-old woman show complete filling of the defect (arrows) of the osteochondral lesion of talus and complete integration. The repaired tissue shows damaged surface with fibrillation. Arthroscopic photograph shows complete filling of the defect with fibrillation of the surface but incomplete integration (arrow), which reveals disagreement with the MRI findings (D). |
For the surface of the repaired tissue category (Table 4), the results were in agreement in 24 of 27 cases (Figure 3), with an ICC value of 0.8523 ( P<.0001). In the adhesion category (Table 5), MRI failed to detect any adhesions, whereas arthroscopy showed adhesion in 2 cases. Therefore, there was complete disagreement in the adhesion category, with an ICC of 0.00 ( P=.50). Finally, with regard to the synovitis category (Table 6), MRI and arthroscopy agreed in 24 of 27 cases, showing a fairly high degree of agreement and an ICC value of 0.7797 ( P<.0001).
 | Surface of the Repaired Tissue |
 | Sagittal (A) and coronal (B) intermediate-weighted fast-spin echo MRIs of a 28-year-old man show complete filling of the defect of the osteochondral lesion of the medial talar dome. The repaired tissue shows damaged surface (arrows) with ulcer and fibrillation. Arthroscopic photograph shows complete defect filling and the surface of repaired tissue is damaged. In this case, MRI demonstrates a good correlation with arthroscopy (C). |
 | Adhesion |
 | Synovitis |
Among the 5 categories, arthroscopy and MRI showed a high degree of agreement, producing ICC values with good reliability in the categories of degree of defect repair and filling, surface of the repair tissue, and synovitis. However, moderate to poor reliability was noted with respect to the categories of integration into the border zone and adhesion, suggesting that there might be some limitations to using these categories. In addition, the adhesion category showed complete disagreement, with an ICC value of 0.00 (Table 7).
 | Statistical Analysis Results |
Discussion
Articular cartilage represents the most easily injured portion of a joint for which there is still no completely established treatment. There are various treatment methods for articular cartilage injury, the first being a primary repair technique in patients with acute symptoms 6 and attempted lavage and debridement in patients with chronic symptoms. 7,8 Marrow-inducing repair techniques are widely used, although they have low biomechanical properties as cartilage is formed by fibrous cartilage. 9Restorative techniques recently have attracted attention as the cartilage is formed by hyaline cartilage restored to the mechanical properties of an uninjured normal cartilage. Our study technique also was used in patients with articular cartilage injury who underwent autologous chondrocyte implantation using restorative techniques.
Regarding patients who underwent various types of surgery, we compared the MRI and arthroscopy results. 3,9 Although Mintz et al 2 reported that the accuracy of MRI for osteochondral lesions of the talus was 83%, this report was not used in the postautologous chondrocyte implantation outcome. To address the increased need for such an outcome, Marlovits et al 10 reported the MOCART scoring system in 2004. The MOCART scoring system gradually developed for high-resolution MRI with the intent of defining pertinent variables for an objective description of repaired tissue, given the current technological limitations. Since Marlovits’ study, multiple articles have been published reporting the use of the MOCART scoring system, suggesting it is a useful method for long-term follow-up using point scales. 5
Recent articles also have highlighted the accuracy of MOCART for evaluating repaired knee cartilage after surgery for osteochondritis dissecans producing assessments with low interobserver variability. 4However, previous articles primarily addressed knee evaluations, prompting the need for evaluation studies on osteochondral lesions of the talus. In our study, the MOCART scoring system was modified (Table 1) by excluding the status of the subchondral lamina injured by drilling from the original MO-CART scoring system. In the arthroscopic fields, not only the structure of repaired tissue but also their signal characteristics, as shown in Table 1, could not be evaluated.
In our study, the categories for the degree of the defect repair and filling, surface of the repair tissue, and synovitis all showed fairly high ICC results. However, the categories of integration and adhesion showed low ICC. Although Mintz et al 2 reported that the accuracy of MRI for osteochondral lesions of the talus was 83%, the report was not based on the results after autologous chondrocyte implantation. Therefore, an evaluation using the MO-CART scoring system was performed to verify accuracy.
According to the results of our study, the integration and adhesion category showed a low ICC, which was an unexpected result. A possible reason for this finding may be because the ankle cartilage is one-third thinner than knee joint cartilage 11 and the articular surfaces of the ankle are naturally closely packed and congruous, unlike the knee where the joint surfaces are not congruent and there is no clear separation of the articular surfaces of the tibiotalar joint. As a result, ankle cartilage is hardly ever evaluated using MRI, in contrast to knee cartilage. It may be necessary to consider new isotropic MRI techniques at a high field (3T) MRI with a dedicated multichannel coil that potentially improves evaluation of repair tissue in the anatomically challenging ankle joint. 12
Furthermore, the adhesion category in our study showed complete disagreement, possibly due to one of the postautologous chondrocyte implantation complications where routine MRI may be limited in distinguishing adhesion from a closely packed joint with thin cartilage. In addition, compared to the results reported by Mintz et al, 2 the MOCART scoring system is much more subdivided than in their study. Consequently, there will be some limitations in correlating arthroscopy and MRI after autologous chondrocyte implantation using those subtypes, which would be better used as an alternative method for following long-term clinical outcomes.
This study has limitations in that we studied relatively few cases and did not include all subtypes of the MOCART scoring system. Our results suggest further studies with more cases are needed. In addition, if modifications of the procedure or advancements in techniques allow for the study of the rest of the MOCART scoring system subtypes not included in this study because of the surgical procedure performed, more accurate data may be obtained in the future.
Conclusion
According to this comparison study between arthroscopy and MRI using the modified MOCART scoring system after autologous chondrocyte implantation in osteochondral lesion of the talus cases, the categories of degree of defect repair and filling, surface of the repair tissue, and synovitis all showed high ICC results, whereas the categories of integration into the border zone and adhesion showed low ICC values. There has been no well-established correlation method between arthroscopy and MRI after autologous chondrocyte implantation for osteochondral lesions of the talus. Although MRI allows for noninvasive evaluation of the repair tissue, limitations remain in using the MOCART scoring system to evaluate repaired cartilage after autologous chondrocyte implantation for osteochondral lesions of the talus.
No comments:
Post a Comment