- Email: [email protected]
Comparison of Postoperative Magnetic Resonance Imaging and Second-Look Arthroscopy for Evaluating Meniscal Allograft Transplantation
Comparison of Postoperative Magnetic Resonance Imaging and Second-Look Arthroscopy for Evaluating Meniscal Allograft Transplantation Jong-Min Kim, M.D., Jong-Min Kim, M.D., Byeong-Sam Jeon, M.D., Chang-Rack Lee, M.D., Sung-Joon Lim, M.D., Kyung-Ah Kim, Ph.D., and Seong-Il Bin, M.D.
Purpose: The aim of this study was to compare the magnetic resonance imaging (MRI) evaluation of transplanted meniscal allograft with second-look arthroscopy and evaluate the sensitivity, specificity, and accuracy of MRI for assessing graft status. Methods: From 1996 to 2012, among 290 knees that underwent meniscal allograft transplantation and received follow-up examination for more than 1 year, those knees that underwent second-look arthroscopy were reviewed. Patients with no postoperative MRI and patients with a time gap between postoperative MRI and second-look arthroscopy of more than 3 months were excluded. Anatomically, the meniscus was divided into 3 segments: anterior one-third, mid body, and posterior one-third. Each part of the meniscus was evaluated using both methods. Grade 3 MRI signal intensity was diagnosed as a meniscal tear radiologically. By use of second-look arthroscopy as the standard, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of postoperative MRI were assessed in each segment of the grafts. Results: Twenty knees were retrospectively enrolled. The specificity, PPV, and accuracy for the anterior one-third were lower than those for the mid body and posterior one-third (specificity of 35.3% v 91.7% and 90%, respectively; PPV of 21.4% v 87.5% and 90.9%, respectively; and accuracy of 45% v 90% and 95%, respectively). However, the sensitivity and NPV were similar among the anterior one-third, mid body, and posterior one-third (sensitivity of 100%, 87.5%, and 100%, respectively; and NPV of 100%, 91.7%, and 100%, respectively). There were no significant differences in the comparison between the diagnostic MRI values of lateral grafts and medial grafts. Of 5 cases that showed grade 3 signal at only the anterior one-third section, 60% had no clinical signs. There were no graft tears in any cases. Conclusions: The anterior one-third of grafts showed low specificity, PPV, and accuracy of postoperative MRI compared with the mid body and posterior one-third. MRI tended to grade the anterior one-third more poorly than second-look arthroscopy. These features should be considered when evaluating transplanted meniscal allografts on postoperative MRI. Level of Evidence: Level III, study of non-consecutive patients evaluating a diagnostic test with a gold standard.
M
eniscal allograft transplantation (MAT) is indicated for unicompartmental knee pain and postactivity effusion after total or near-total meniscectomy in patients with closed physes.1-4 Favorable short- and intermediate-term clinical results have been reported.5-7 Magnetic resonance imaging (MRI) is often obtained to evaluate the allograft tissue in the clinical setting of a
patient with persistent pain after meniscal transplantation. The utility of MRI to evaluate knee allograft tissues has been reported.3,5-11 Second-look arthroscopy is generally regarded as the gold standard for evaluating the postoperative status of meniscal repairs.12-14 However, many patients might be unwilling to undergo additional invasive procedures.
From the Department of Orthopaedic Surgery, College of Medicine, Kosin University, Kosin Gospel Hospital (J-M.K. [first author]), Busan; Department of Orthopedic Surgery, College of Medicine, Asan Medical Center, University of Ulsan (J-M.K. [second author], B-S.J., C-R.L., S-J.L., S-I.B.), Seoul; and Department of Biomedical Engineering, College of Medicine, Chungbuk National University (K-A.K.), Cheongju, Republic of Korea. This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2012-0000479) and by a grant from the Korea Healthcare Technology Research and Development Project, Ministry of Health and Welfare, Republic of Korea (A100054).
The authors report that they have no conflicts of interest in the authorship and publication of this article. Received April 16, 2014; accepted November 26, 2014. Address correspondence to Seong-Il Bin, M.D., Department of Orthopedic Surgery, College of Medicine, University of Ulsan, Asan Medical Center, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 138-736, Republic of Korea. E-mail: [email protected] Ó 2015 by the Arthroscopy Association of North America 0749-8063/14314/$36.00 http://dx.doi.org/10.1016/j.arthro.2014.11.041
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 31, No 5 (May), 2015: pp 859-866
859
860
J-M. KIM ET AL.
Therefore MRI is more often performed postoperatively to evaluate patients rather than second-look arthroscopy.7,15,16 Few prior studies have directly compared MRI and second-look arthroscopy after MAT. The aim of this study was to compare the MRI evaluation of transplanted meniscal allograft with second-look arthroscopy and evaluate the sensitivity, specificity, and accuracy of MRI for assessing graft status. In the native meniscus, increased signal intensity at the anterior horn of the meniscus seen on MRI commonly does not represent a clinically significant lesion.17 We hypothesize that concordance between these 2 diagnostic methods could differ depending on the area of the meniscus.
Methods Patients and Inclusion Criteria Between 1996 and 2012, among 290 knees that underwent medial or lateral MAT at our institution and received follow-up examination for more than 1 year, knees that underwent second-look arthroscopy for the evaluation of a meniscal graft were reviewed. We excluded knees with no postoperative MRI and knees with a time gap between postoperative MRI and second-look arthroscopy of more than 3 months. This study was approved by our institutional review board. Indications for MAT included persistent pain in the affected compartment associated with meniscal deficiency, a well-aligned mechanical axis, and articular chondral wear of International Cartilage Repair Society (ICRS) grade 3 or lower. Focal ICRS grade 4 lesions were not considered contraindications. Contraindications for MAT included patients with generalized arthritis, axial malalignment, uncorrected joint instability, skeletal immaturity, and age greater than 50 years. Radiographic examinations, including the standing anteroposterior view, weight-bearing posteroanterior view with 45 of flexion, lateral view, Merchant view, and standing full-leg anteroposterior view, were performed in every patient to assess the degree of arthritis and mechanical axis of the lower limb. Surgical Technique Proper allograft size was determined using the method described by Pollard et al.18 Lateral MAT was performed using the keyhole technique.2 Medial MAT was performed using the modified bone plug technique.19 All operations were performed by a single surgeon (S-I.B.). The remaining host meniscus was resected arthroscopically, leaving a peripheral rim of about 1 mm. For the bone bridge for the lateral graft, a keyhole tibial slot was made just under the lateral eminence, parallel to the posterior slope of the tibial plateau, and the meniscal allograft was introduced through an anterior mini-arthrotomy. For the bone
plugs for the medial menisci, tibial tunnels were reamed over guidewires positioned in the anatomic anterior and posterior horn attachments. After optimal allograft positioning was confirmed, traditional insideout meniscal repair was performed, using 10 to 12 vertical No. 2-0 nonabsorbable sutures placed 3 to 5 mm apart. Postoperative Rehabilitation Immediately after surgery, we encouraged patients to perform isometric muscle-strengthening exercises. Continuous passive range-of-motion (ROM) exercises were resumed after drainage removal on day 1 after MAT. The goal of ROM was 0 to 60 within the first 3 weeks, up to 120 within 4 to 6 weeks, and full flexion within 8 to 12 weeks. Noneweight-bearing walking with crutches was required during the first 3 weeks. Partial weight-bearing walking (up to 50% of body weight) was required at 4 weeks. Full weight-bearing walking was started at 7 weeks. Rehabilitation was maintained for 3 months to restore full ROM and quadriceps strength. Six months after surgery, patients were allowed to return to all daily activities. MRI Evaluation All postoperative MRI scans were performed to evaluate the cause of persistent knee pain, except for 4 cases undergoing an initial routine check-up and 4 cases undergoing additional procedures. All MRI scans were taken before second-look arthroscopy was performed, except in 1 case. One case with revision anterior cruciate ligament (ACL) reconstruction underwent postoperative MRI 3 days after second-look arthroscopy. All MRI examinations were performed using a conventional 1.5-T magnetic resonance (MR) scanner (Siemens Medical Solutions, Malvern, PA) or 3.0-T MR scanner (Philips Medical Systems, Andover, MA). The 1.5-T MR scanner was used in 45% of cases (9 of 20) before March 2008, and the 3.0-T MR scanner was used in 55% of cases (11 of 20) thereafter. Sagittal T2- and T1-weighted spin-echo images and sagittal/coronal proton densityeweighted fast spin-echo images were obtained using following parameters: (1) T1-weighted imagesdrepetition time (TR) of 530 milliseconds and echo time (TE) of 12 milliseconds; (2) T2-weighted imagesdTR of 3,900 milliseconds and TE of 105 milliseconds; (3) proton densityeweighted imagesdTR of 2,760 milliseconds and TE of 15 milliseconds; and (4) gradient-echo imagesdTR of 700 milliseconds, TE of 18 milliseconds, 20 flip angle with 1 echo train length, 3-mm slice thickness, 0.6- to 0.9-mm intervals, 512 512 matrix, and 16-cm field of view. When we used the 3.0-T MR scanner, sagittal images were obtained at a 1.5-mm slice thickness and coronal images were obtained at a 2.5-mm slice thickness.
861
MENISCAL ALLOGRAFT TRANSPLANTATION
The MRI results regarding graft status after MAT are described using the criteria of Stoller et al.20 Grade 1 signal intensity is defined as single or multiple punctiform hypersignals that do not extend to the joint surface. Grade 2 signal intensity is defined as a linear intrameniscal hypersignal that does not extend to the joint surface. Grade 3 signal intensity is defined as a linear hypersignal that extends to at least 1 of the 2 joint surfaces. Grade 3 MRI signal intensity at the graft was radiologically defined as a meniscal tear. The meniscus was anatomically divided into 3 segments: anterior one-third, mid body, and posterior one-third. These respective segments were assessed using MRI and second-look arthroscopy. Because grade 1 and grade 2 MRI signals cannot be directly visualized on arthroscopy, it was impossible to compare grade 1 and 2 MRI signals with second-look arthroscopy. We evaluated whether grade 3 signal intensity on MRI would present as a meniscal tear on second-look arthroscopy. We also evaluated whether extrusion was associated with a tear. The postoperative MRI findings were compared with second-look arthroscopic findings, used as the gold standard. True-positive, true-negative, false-positive, and false-negative results, as well as the sensitivity, specificity, positive predictive value (PPV), negative predictive value, and accuracy, were assessed for the diagnostic value of postoperative MRI. The postoperative MRI findings were reviewed by a single musculoskeletal radiologist with 10 years of clinical experience and 2 orthopaedic surgeons (J-M.K., J-M.K.) with more than 8 years of knee MRI experience excluding the single operator. When all interpreters reached a consensus through discussion, MRI findings were accepted for this study. Second-Look Arthroscopy Second-look arthroscopy was initially performed in 4 patients as part of routine check-up within 1 year after MAT. However, many patients without pain were reluctant to undergo another invasive procedure. Thereafter second-look arthroscopy was performed when there was persistent knee pain or MRI results suggestive of allograft tears. In addition, second-look arthroscopy was included as part of an additional procedure, such as ACL reconstruction (2 patients), revision ACL reconstruction (1 patient), or osteochondral autograft transfer (1 patient). The intact transplanted allograft was assessed according to the following features: stability at the meniscocapsular junction, normal gross appearance, and no graft tears. When graft tears were confirmed using secondlook arthroscopy, patients underwent treatment with trimming, partial meniscectomy, or subtotal/total meniscectomy according to the extent and type of tear.
Statistics Statistical analyses were performed using SPSS software for Windows, version 16.0 (IBM, Armonk, NY). The data for quantitative variables were given as mean standard deviation. The demographic data of the lateral graft group and medial graft group were assessed with the Mann-Whitney test for continuous variables and Fisher exact test for categorical variables. To compare the lateral graft group and medial graft group regarding the diagnostic value of MRI, including sensitivity, specificity, PPV, negative predictive value, and accuracy, according to graft segments, the Fisher exact test was used. The comparison of diagnostic MRI values between the 1.5-T and 3.0-T MRI groups was also assessed with the Fisher exact test. Differences in the mean values of meniscal extrusion between the graft tear group and the group with no graft tears were assessed with the paired t test. In this study P <.05 was considered statistically significant.
Results Among 25 knees that underwent second-look arthroscopy for the evaluation of meniscal grafts, we excluded 2 knees with no postoperative MRI and 3 knees with a time gap between postoperative MRI and second-look arthroscopy of more than 3 months. Therefore 20 knees were retrospectively enrolled in this study. The study group was composed of 16 male and 4 female patients. Five medial and 15 lateral allograft menisci were assessed. The mean age of the study population was 32.1 years (SD, 8.4 years). The mean time between MAT and second-look arthroscopy was 17.4 months (SD, 11.0 months; range, 7.0 to 49.5 months). The mean time between MAT and postoperative MRI was 15.9 months (SD, 10.8 months; range, 6.0 to 46.4 months). The mean time between postoperative MRI and second-look arthroscopy was 1.6 months (SD, 1.1 months; range, 0.1 to 3.0 months). The demographic data of the lateral graft and medial graft groups are shown in Table 1. Regarding the diagnostic values of MRI, the anterior one-third of meniscal grafts showed lower specificity, PPV, and accuracy (35.3%, 21.4%, and 45%, respectively) compared with the mid body and posterior one-third of meniscal grafts, which both showed high diagnostic values (Table 2). Eleven cases were misdiagnosed with tears at the anterior one-third on MRI (Fig 1). When this group was divided into a lateral graft group and medial graft group, the respective groups also showed low specificity, PPV, and accuracy. There was no significant difference in the diagnostic values of MRI between these 2 groups (Table 2). The diagnostic values of 1.5-T and 3.0-T MRI were also evaluated. These 2 MRI groups showed low specificity, PPV, and accuracy at the anterior one-third. These 2 groups showed no
862
J-M. KIM ET AL.
Table 1. Demographic Data of Patients No. of patients Age at MAT, yr Sex (M/F), n BMI, kg/m2 Time between MAT and postoperative MRI, mo Time between MAT and second-look arthroscopy, mo Time between postoperative MRI and second-look arthroscopy, mo
Overall 20 32.1 8.42 16/4 24.9 3.0 15.9 10.8 17.4 11.0 1.6 1.1
Lateral Graft Group 15 33.7 8.3 13/2 24.5 3.4 14.0 8.9 15.6 8.8 1.6 1.0
Medial Graft Group 5 26.6 6.9 3/2 25.8 1.6 19.0 14.6 23.0 15.5 1.6 1.5
P Value .10 .25 .32 .36 .32 .76
NOTE. Data presented as mean standard deviation unless otherwise indicated. BMI, body mass index; F, female; M, male; MAT, meniscal allograft transplantation; MRI, magnetic resonance imaging.
significant difference in all diagnostic values of MRI according to graft segment (Table 3). Of 12 cases that showed grade 3 MRI signal intensity at the mid body or posterior one-third, 83.3% (10 of 12) had at least 1 symptom or physical sign, such as swelling, joint-line tenderness, locking or blocking, or a positive McMurray test. Of these, 90% (9 of 10, with a lateral graft in 8 cases and a medial graft in 1 case) had a tear on second-look arthroscopy. However, 8.3% of the cases that had no clinical sign (2 of 12) showed a small flap tear at the posterior one-third of the medial graft and shrinkage of the lateral graft. Of 5 cases that showed grade 3 MRI signal intensity only at the anterior one-third section, 20% (1 of 5) had swelling and 20% (1 of 5) had joint-line tenderness, whereas 60% (3 of 5) had no clinical signs. There were no isolated tears of the anterior one-third section on second-look arthroscopy. Ten cases showed graft tears. The graft tears showed a variety of patterns, anatomic locations, and histologic zones. The details of these cases are shown in Table 4. Of these cases, 50% (5 of 10) showed meniscal extrusion (>3 mm). In 10 cases with no graft tear, 50% (5 of 10) showed meniscal extrusion (>3 mm). The mean extrusion of the graft tear group was 3.58 mm (SD, 1.40 mm; range, 1.69 to 6.70 mm), and the mean extrusion of the group with no graft tears was 3.51 mm (SD, 1.20 mm; range, 1.53 to 5.06 mm). There was no significant difference in the
comparison of meniscal extrusion between these 2 groups (P ¼ .90). In the graft tear group, 90% (9 of 10) had ICRS grade 3 cartilage of the femur or tibia. The mean grades of the femur and tibia were 2.7 (SD, 0.5; range, 2 to 3) and 2.8 (SD, 0.6; range, 2 to 3), respectively. In the group with no graft tears, 40% (4 of 10) had ICRS grade 3 cartilage of the femur or tibia. The mean grades of the femur and tibia were 2.1 (SD, 0.6; range, 1 to 3) and 2.1 (SD, 0.7; range, 1 to 3), respectively, in the group with no graft tears. The graft tear group showed significantly higher ICRS cartilage grades (P ¼ .02 for femur and P ¼ .04 for tibia).
Discussion The most important finding of this study was that the anterior one-third of meniscal grafts showed lower specificity, PPV, and accuracy on MRI in comparison with the mid body and posterior one-third. We also found that MRI tended to overestimate tears at the anterior one-third when the criteria of Stoller et al.20 were applied. The combined application of clinical signs including swelling, locking, joint-line tenderness, and the McMurray test might be useful in the diagnosis of a graft tear. MAT shows valuable clinical outcomes, including pain relief and functional improvement. However, because these clinical results are based on subjective evaluation, it is difficult to compare results between
Table 2. Diagnostic Results of Postoperative MRI According to Graft Side Meniscal Segment Anterior one-third
P value Mid body
P value Posterior one-third
P value
Graft Side (n) All (20) Lateral graft (15) Medial graft (5)
TP, n 3 3 0
TN, n 6 5 1
FP, n 11 7 4
FN, n 0 0 0
All (20) Lateral graft (15) Medial graft (5)
7 7 0
11 7 4
1 1 0
1 0 1
All (20) Lateral graft (15) Medial graft (5)
10 9 1
9 5 4
1 1 0
0 0 0
Sensitivity 100% 100% d d 87.5% 100% 0% .13 100% 100% 100% d
Specificity 35.3% 41.7% 20% .60 91.7% 87.5% 100% > .99 90% 83.3% 100% > .99
PPV 21.4% 30% 0% .51 87.5% 87.5% d d 90.9% 90% 100% > .99
NPV 100% 100% 100% d 91.7% 100% 80% .42 100% 100% 100% d
Accuracy 45% 53% 20% .32 90% 93.3% 80% .45 95% 93.3% 100% > .99
FN, false negative; FP, false positive; MRI, magnetic resonance imaging; NPV, negative predictive value; PPV, positive predictive value; TN, true negative; TP, true positive.
863
MENISCAL ALLOGRAFT TRANSPLANTATION
Fig 1. A 24-year-old male patient who underwent left knee medial meniscal allograft transplantation and anterior cruciate ligament reconstruction. (A) At the follow-up examination at 49.5 months, sagittal proton densityeweighted magnetic resonance imaging (MRI) showed grade 3 signal intensity at the anterior one-third (circle). (B) Sagittal T2-weighted MRI showed grade 3 signal intensity at the anterior one-third (circle). (C) Good stability and no tear in the anterior one-third were verified using second-look arthroscopy (circle). (MFC, medial femoral condyle; MM, medial meniscal graft; MTP, medial tibial plateau.)
individual studies. Therefore the importance of objective evaluation is on the rise. Accordingly, objective evaluations such as MRI and second-look arthroscopy have been applied to determine clinical results after MAT.4,6,7,10,11,15,16,19,21 Although second-look arthroscopy is accepted as the gold standard for evaluating postoperative results after knee surgery, it is an invasive procedure.12,13,21,22 Accordingly, MRI has been proposed as an alternative for evaluating graft status after MAT.6,7,11,13,15,16 The accuracy of MRI for assessing graft status after MAT remains uncertain. In this study we used the criteria of Stoller et al.20 to evaluate postoperative meniscal allografts. Although these criteria have been established as useful criteria for evaluating the degeneration of the native meniscus on MRI,23,24 they may not be relevant when describing graft status after MAT. In fact, we often observed that the postoperative MRI findings did not match the second-look arthroscopic
findings at the anterior third of transplanted meniscal allografts. We hypothesized that the consistency of these 2 diagnostic methods could be different according to the segments of the meniscus. To investigate this hypothesis, a direct comparison of postoperative MRI findings and second-look arthroscopic findings was carried out in this study. Several authors have reported that high signal intensity can be observed in grafts after MAT.15,25-27 Verdonk et al.3 reported that of 17 patients who underwent MRI at a mean of 12 years after MAT, 10 (59%) showed grade 3 signal intensity. Using MRI within the first postoperative year, Lee et al.10 evaluated 43 patients who underwent MAT. Higher signals were observed for the transplanted allograft in comparison with the normal meniscus. In particular, the anterior horn showed higher signal intensity than the posterior horn throughout the first year after MAT. Lee et al. also noted that persistently increased
Table 3. Diagnostic Results of Postoperative MRI According to MRI Type Meniscal Segment Anterior one-third
P value Mid body
P value Posterior one-third
P value
MRI Type (n) All (20) 1.5-T MRI (9) 3.0-T MRI (11)
TP, n 3 3 0
TN, n 6 3 3
FP, n 11 3 8
FN, n 0 0 0
All (20) 1.5-T MRI (9) 3.0-T MRI (11)
7 3 4
11 6 5
1 0 1
1 0 1
All (20) 1.5-T MRI (15) 3.0-T MRI (5)
10 5 1
9 4 4
1 0 0
0 0 0
Sensitivity 100% 100% d d 87.%5 100% 80% > .99 100% 100% 100% d
Specificity 35.3% 50% 27.3% .60 91.7% 100% 83.3% > .99 90% 100% 83.3% > .99
PPV 21.4% 50% 0% .06 87.5% 100% 80% > .99 90.9% 100% 83.3% > .99
NPV 100% 100% 100% d 91.7% 100% 83.8% > .99 100% 100% 100% d
Accuracy 45% 66.6% 27.3% .18 90% 93.3% 81.8% .48 95% 100% 90.9% > .99
FN, false negative; FP, false positive; MRI, magnetic resonance imaging; NPV, negative predictive value; PPV, positive predictive value; TN, true negative; TP, true positive.
864 White zone Red-white zone Red zone Red-white zone Red zone Red zone Red zone Small flap Flap Longitudinal Complex Longitudinal Longitudinal B-H 2.96 2.91 3.84 4.60 2.32 3.84 6.70 3 2 2 3 3 3 2 2 3 3 3 2 3 2 2 1 2 0 0 1 2 21.9 23.0 25.8 23.3 25.7 27.8 24.4 Lateral Medial Lateral Lateral Lateral Lateral Lateral 26 35 26 37 43 40 32 M M F M M M M 4 5 6 7 8 9 10
ant, anterior; B-H, bucket handle; BMI, body mass index; F, female; ICRS, International Cartilage Repair Society; M, male; MAT, meniscal allograft transplantation; post, posterior. *Positive values indicate varus, whereas negative values indicate valgus.
Red-white zone Red-white zone Flap Horizontal 4.09 1.69 3 3 3 3 2 1 23.1 22.3 Lateral Lateral M M 2 3
41 21
Tibia 3 Femur 3 BMI, kg/m2 31.1 Graft Side Lateral Age at MAT, yr 43 Sex M Case No. 1
Table 4. Details of 10 Cases With Meniscal Graft Tears
Alignment* 3
ICRS Grade
Amount of Extrusion, mm 2.89
Anatomic Location Ant one-third þ mid body þ post one-third Post one-third Ant one-third þ mid body þ post one-third Post one-third Mid body þ post one-third Mid body þ post one-third Mid body þ post one-third Mid body þ post one-third Mid body þ post one-third Ant one-third þ mid body þ post one-third
Tear Pattern B-H
Histologic Location Red zone
J-M. KIM ET AL.
intrameniscal intensity might indicate normal healing processes rather than degenerative changes or tears. We think that this phenomenon may often be observed at the anterior one-third because most surgical trauma occurs at the anterior portion during MAT. MRI provides only limited evaluations of postoperative status after meniscal repair because scarring at the repaired site can present as a grade 3 signal.13,22,28 Miao et al.14 reported that sagittal T2-weighted imaging shows the highest diagnostic accuracy (60.5%) and sagittal T1-weighted imaging shows the lowest diagnostic accuracy (28.9%) in comparison with second-look arthroscopy (38 repaired menisci were evaluated using postoperative MRI at 15 months [SD, 5.1 months] after meniscal repair). In a different study, Miao et al.13 also noted that second-look arthroscopy is the most reliable way to determine meniscal healing and that the combined application of different sequences on MRI could improve its diagnostic value. Accordingly, these studies showed the limitations of MRI, even when evaluating native meniscal healing after meniscal repair. Because there are few previous studies that have compared postoperative MRI and second-look arthroscopy after MAT21,25 and these studies did not directly compare both methods, it is difficult to compare our results with those of other studies. We found that the diagnostic values of postoperative MRI after MAT can be affected by the segments of the graft. The anterior one-third showed low specificity, PPV, and accuracy compared with the mid body and posterior body. These results indicate that MRI can reliably assess the mid body and posterior one-third of the transplanted graft but not the anterior one-third. Grade 3 signal intensity at the anterior one-third could be misdiagnosed as a tear when using the criteria of Stoller et al.20 According to our study results, grade 3 signal intensity at the anterior one-third of a graft does not necessarily indicate a graft tear (according to the criteria of Stoller et al.) but grade 3 signal intensity at the mid body and posterior onethird could be associated with graft tears. Therefore we believe that the criteria of Stoller et al. are unsuitable for interpreting allograft meniscus. New evaluation systems that can more reliably interpret graft status on MRI are needed. For the clinical assessment of meniscal repair, Barrett et al.29 performed physical examination to assess the following clinical signs: swelling, joint-line tenderness, locking, and McMurray test. Absence of the 4 signs indicated a healed meniscus. Miao et al.13 also used these clinical signs for the evaluation of meniscal repair. They noted that the healing rate was 70.8%, significantly lower than the healing rate found by
865
MENISCAL ALLOGRAFT TRANSPLANTATION
second-look arthroscopy. When a repaired meniscus with 2 or more signs was defined as unhealed, the diagnostic value could be improved. Similarly, our study found that there were no graft tears in 5 cases that showed grade 3 signal intensity at only the anterior one-third without clinical signs or with 1 clinical sign. We believed that the combination of postoperative MRI and clinical assessment could reduce the rate of misdiagnosis of graft tears. The clinical relevance of our study is that taking into account both the characteristics of MRI findings at the anterior one-third and the clinical assessment findings, unnecessary planning of second-look arthroscopy may be avoided in patients with abnormally high signal intensity in the anterior one-third of the meniscal allograft. Limitations This study has several limitations. First, this is a retrospective study without a control group, and the number of cases was small. The 20 patients enrolled in this study may not represent the entire group of 290 patients. This study did not include the MR findings of the remaining large group of patients who were not selected for surgery. Second, second-look arthroscopy was not performed according to consistent surgical indications, which might affect the homogeneity of our study group. These factors could induce selection bias, which might yield an inability to comment on the diagnostic values of MRI for the remaining group. Third, 1.5-T and 3.0-T MRI scanners were used to evaluate the status of the meniscal grafts after MAT. MRI findings associated with second-look arthroscopy can be affected by the type of MRI scanner, which might induce bias regarding the diagnostic results. However, our study showed that there was no difference in the comparison of diagnostic MRI values between these 2 MRI scanners. Lastly, the relations between the pattern and degree of grade 3 signal intensity and second-look arthroscopic findings were not discussed sufficiently. We will consider these limitations in subsequent research on new criteria regarding graft status on postoperative MRI after MAT.
Conclusions The anterior one-third of a transplanted meniscal allograft showed lower specificity, PPV, and accuracy of postoperative MRI compared with the mid body and posterior one-third. When the criteria of Stoller et al.20 were used to evaluate grafts after MAT, MRI tended to grade the anterior one-third more poorly than secondlook arthroscopy. These features should be considered when evaluating transplanted meniscal allografts on postoperative MRI.
References 1. LaPrade RF, Wills NJ, Spiridonov SI, Perkinson S. A prospective outcomes study of meniscal allograft transplantation. Am J Sports Med 2010;38:1804-1812. 2. Wilcox TR, Goble EM. Indications for meniscal allograft reconstruction. Am J Knee Surg 1996;9:35-36. 3. Verdonk PC, Verstraete KL, Almqvist KF, et al. Meniscal allograft transplantation: Long-term clinical results with radiological and magnetic resonance imaging correlations. Knee Surg Sports Traumatol Arthrosc 2006;14:694-706. 4. Ha JK, Shim JC, Kim DW, Lee YS, Ra HJ, Kim JG. Relationship between meniscal extrusion and various clinical findings after meniscus allograft transplantation. Am J Sports Med 2010;38:2448-2455. 5. Noyes FR, Barber-Westin SD. Prospective evaluation of allograft meniscus transplantation: A minimum 2-year follow-up. Am J Sports Med 2006;34:2038-2039 (author reply 2039). 6. Kim JM, Lee BS, Kim KH, Kim KA, Bin SI. Results of meniscus allograft transplantation using bone fixation: 110 cases with objective evaluation. Am J Sports Med 2012;40:1027-1034. 7. Lee DH, Kim SB, Kim TH, Cha EJ, Bin SI. Midterm outcomes after meniscal allograft transplantation: Comparison of cases with extrusion versus without extrusion. Am J Sports Med 2010;38:247-254. 8. Stollsteimer GT, Shelton WR, Dukes A, Bomboy AL. Meniscal allograft transplantation: A 1- to 5-year followup of 22 patients. Arthroscopy 2000;16:343-347. 9. Graf KW Jr, Sekiya JK, Wojtys EM. Long-term results after combined medial meniscal allograft transplantation and anterior cruciate ligament reconstruction: Minimum 8.5-year follow-up study. Arthroscopy 2004;20:129-140. 10. Lee DH, Lee BS, Chung JW, et al. Changes in magnetic resonance imaging signal intensity of transplanted meniscus allografts are not associated with clinical outcomes. Arthroscopy 2011;27:1211-1218. 11. Lee BS, Kim JM, Sohn DW, Bin SI. Review of meniscal allograft transplantation focusing on long-term results and evaluation methods. Knee Surg Relat Res 2013;25:1-6. 12. Morgan CD, Wojtys EM, Casscells CD, Casscells SW. Arthroscopic meniscal repair evaluated by second-look arthroscopy. Am J Sports Med 1991;19:632-637 (discussion 637-638). 13. Miao Y, Yu JK, Ao YF, Zheng ZZ, Gong X, Leung KK. Diagnostic values of 3 methods for evaluating meniscal healing status after meniscal repair: Comparison among second-look arthroscopy, clinical assessment, and magnetic resonance imaging. Am J Sports Med 2011;39: 735-742. 14. Miao Y, Yu JK, Zheng ZZ, et al. MRI signal changes in completely healed meniscus confirmed by second-look arthroscopy after meniscal repair with bioabsorbable arrows. Knee Surg Sports Traumatol Arthrosc 2009;17: 622-630. 15. Lee DH, Kim TH, Lee SH, Kim CW, Kim JM, Bin SI. Evaluation of meniscus allograft transplantation with serial magnetic resonance imaging during the first postoperative year: Focus on graft extrusion. Arthroscopy 2008;24:1115-1121.
866
J-M. KIM ET AL.
16. Lee BS, Chung JW, Kim JM, Cho WJ, Kim KA, Bin SI. Morphologic changes in fresh-frozen meniscus allografts over 1 year: A prospective magnetic resonance imaging study on the width and thickness of transplants. Am J Sports Med 2012;40:1384-1391. 17. Shepard MF, Hunter DM, Davies MR, Shapiro MS, Seeger LL. The clinical significance of anterior horn meniscal tears diagnosed on magnetic resonance images. Am J Sports Med 2002;30:189-192. 18. Pollard ME, Kang Q, Berg EE. Radiographic sizing for meniscal transplantation. Arthroscopy 1995;11:684-687. 19. Gonzalez-Lucena G, Gelber PE, Pelfort X, Tey M, Monllau JC. Meniscal allograft transplantation without bone blocks: A 5- to 8-year follow-up of 33 patients. Arthroscopy 2010;26:1633-1640. 20. Stoller DW, Cannon WD, Anderson LJ. Magnetic resonance imaging in orthopaedics and sports medicine, Ed 2. Philadelphia: Lippincott Williams & Wilkins, 1997. 21. van Arkel ER, Goei R, de Ploeg I, de Boer HH. Meniscal allografts: Evaluation with magnetic resonance imaging and correlation with arthroscopy. Arthroscopy 2000;16: 517-521. 22. Steenbrugge F, Verstraete K, Verdonk R. Magnetic resonance imaging of the surgically repaired meniscus: A
23.
24.
25.
26.
27.
28.
29.
13-year follow-up study of 13 knees. Acta Orthop Scand 2004;75:323-327. Crues JV III, Mink J, Levy TL, Lotysch M, Stoller DW. Meniscal tears of the knee: Accuracy of MR imaging. Radiology 1987;164:445-448. Stoller DW, Martin C, Crues JV III, Kaplan L, Mink JH. Meniscal tears: Pathologic correlation with MR imaging. Radiology 1987;163:731-735. Potter HG, Rodeo SA, Wickiewicz TL, Warren RF. MR imaging of meniscal allografts: Correlation with clinical and arthroscopic outcomes. Radiology 1996;198:509-514. Noyes FR, Barber-Westin SD, Rankin M. Meniscal transplantation in symptomatic patients less than fifty years old. J Bone Joint Surg Am 2004;86:1392-1404. Verstraete KL, Verdonk R, Lootens T, Verstraete P, De Rooy J, Kunnen M. Current status and imaging of allograft meniscal transplantation. Eur J Radiol 1997;26: 16-22. Mustonen AO, Tielinen L, Lindahl J, Hirvensalo E, Kiuru M, Koskinen SK. MRI of menisci repaired with bioabsorbable arrows. Skeletal Radiol 2006;35:515-521. Barrett GR, Field MH, Treacy SH, Ruff CG. Clinical results of meniscus repair in patients 40 years and older. Arthroscopy 1998;14:824-829.
Purpose: The aim of this study was to compare the magnetic resonance imaging (MRI) evaluation of transplanted meniscal allograft with second-look arthroscopy and evaluate the sensitivity, specificity, and accuracy of MRI for assessing graft status. Methods: From 1996 to 2012, among 290 knees that underwent meniscal allograft transplantation and received follow-up examination for more than 1 year, those knees that underwent second-look arthroscopy were reviewed. Patients with no postoperative MRI and patients with a time gap between postoperative MRI and second-look arthroscopy of more than 3 months were excluded. Anatomically, the meniscus was divided into 3 segments: anterior one-third, mid body, and posterior one-third. Each part of the meniscus was evaluated using both methods. Grade 3 MRI signal intensity was diagnosed as a meniscal tear radiologically. By use of second-look arthroscopy as the standard, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of postoperative MRI were assessed in each segment of the grafts. Results: Twenty knees were retrospectively enrolled. The specificity, PPV, and accuracy for the anterior one-third were lower than those for the mid body and posterior one-third (specificity of 35.3% v 91.7% and 90%, respectively; PPV of 21.4% v 87.5% and 90.9%, respectively; and accuracy of 45% v 90% and 95%, respectively). However, the sensitivity and NPV were similar among the anterior one-third, mid body, and posterior one-third (sensitivity of 100%, 87.5%, and 100%, respectively; and NPV of 100%, 91.7%, and 100%, respectively). There were no significant differences in the comparison between the diagnostic MRI values of lateral grafts and medial grafts. Of 5 cases that showed grade 3 signal at only the anterior one-third section, 60% had no clinical signs. There were no graft tears in any cases. Conclusions: The anterior one-third of grafts showed low specificity, PPV, and accuracy of postoperative MRI compared with the mid body and posterior one-third. MRI tended to grade the anterior one-third more poorly than second-look arthroscopy. These features should be considered when evaluating transplanted meniscal allografts on postoperative MRI. Level of Evidence: Level III, study of non-consecutive patients evaluating a diagnostic test with a gold standard.
M
eniscal allograft transplantation (MAT) is indicated for unicompartmental knee pain and postactivity effusion after total or near-total meniscectomy in patients with closed physes.1-4 Favorable short- and intermediate-term clinical results have been reported.5-7 Magnetic resonance imaging (MRI) is often obtained to evaluate the allograft tissue in the clinical setting of a
patient with persistent pain after meniscal transplantation. The utility of MRI to evaluate knee allograft tissues has been reported.3,5-11 Second-look arthroscopy is generally regarded as the gold standard for evaluating the postoperative status of meniscal repairs.12-14 However, many patients might be unwilling to undergo additional invasive procedures.
From the Department of Orthopaedic Surgery, College of Medicine, Kosin University, Kosin Gospel Hospital (J-M.K. [first author]), Busan; Department of Orthopedic Surgery, College of Medicine, Asan Medical Center, University of Ulsan (J-M.K. [second author], B-S.J., C-R.L., S-J.L., S-I.B.), Seoul; and Department of Biomedical Engineering, College of Medicine, Chungbuk National University (K-A.K.), Cheongju, Republic of Korea. This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2012-0000479) and by a grant from the Korea Healthcare Technology Research and Development Project, Ministry of Health and Welfare, Republic of Korea (A100054).
The authors report that they have no conflicts of interest in the authorship and publication of this article. Received April 16, 2014; accepted November 26, 2014. Address correspondence to Seong-Il Bin, M.D., Department of Orthopedic Surgery, College of Medicine, University of Ulsan, Asan Medical Center, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 138-736, Republic of Korea. E-mail: [email protected] Ó 2015 by the Arthroscopy Association of North America 0749-8063/14314/$36.00 http://dx.doi.org/10.1016/j.arthro.2014.11.041
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 31, No 5 (May), 2015: pp 859-866
859
860
J-M. KIM ET AL.
Therefore MRI is more often performed postoperatively to evaluate patients rather than second-look arthroscopy.7,15,16 Few prior studies have directly compared MRI and second-look arthroscopy after MAT. The aim of this study was to compare the MRI evaluation of transplanted meniscal allograft with second-look arthroscopy and evaluate the sensitivity, specificity, and accuracy of MRI for assessing graft status. In the native meniscus, increased signal intensity at the anterior horn of the meniscus seen on MRI commonly does not represent a clinically significant lesion.17 We hypothesize that concordance between these 2 diagnostic methods could differ depending on the area of the meniscus.
Methods Patients and Inclusion Criteria Between 1996 and 2012, among 290 knees that underwent medial or lateral MAT at our institution and received follow-up examination for more than 1 year, knees that underwent second-look arthroscopy for the evaluation of a meniscal graft were reviewed. We excluded knees with no postoperative MRI and knees with a time gap between postoperative MRI and second-look arthroscopy of more than 3 months. This study was approved by our institutional review board. Indications for MAT included persistent pain in the affected compartment associated with meniscal deficiency, a well-aligned mechanical axis, and articular chondral wear of International Cartilage Repair Society (ICRS) grade 3 or lower. Focal ICRS grade 4 lesions were not considered contraindications. Contraindications for MAT included patients with generalized arthritis, axial malalignment, uncorrected joint instability, skeletal immaturity, and age greater than 50 years. Radiographic examinations, including the standing anteroposterior view, weight-bearing posteroanterior view with 45 of flexion, lateral view, Merchant view, and standing full-leg anteroposterior view, were performed in every patient to assess the degree of arthritis and mechanical axis of the lower limb. Surgical Technique Proper allograft size was determined using the method described by Pollard et al.18 Lateral MAT was performed using the keyhole technique.2 Medial MAT was performed using the modified bone plug technique.19 All operations were performed by a single surgeon (S-I.B.). The remaining host meniscus was resected arthroscopically, leaving a peripheral rim of about 1 mm. For the bone bridge for the lateral graft, a keyhole tibial slot was made just under the lateral eminence, parallel to the posterior slope of the tibial plateau, and the meniscal allograft was introduced through an anterior mini-arthrotomy. For the bone
plugs for the medial menisci, tibial tunnels were reamed over guidewires positioned in the anatomic anterior and posterior horn attachments. After optimal allograft positioning was confirmed, traditional insideout meniscal repair was performed, using 10 to 12 vertical No. 2-0 nonabsorbable sutures placed 3 to 5 mm apart. Postoperative Rehabilitation Immediately after surgery, we encouraged patients to perform isometric muscle-strengthening exercises. Continuous passive range-of-motion (ROM) exercises were resumed after drainage removal on day 1 after MAT. The goal of ROM was 0 to 60 within the first 3 weeks, up to 120 within 4 to 6 weeks, and full flexion within 8 to 12 weeks. Noneweight-bearing walking with crutches was required during the first 3 weeks. Partial weight-bearing walking (up to 50% of body weight) was required at 4 weeks. Full weight-bearing walking was started at 7 weeks. Rehabilitation was maintained for 3 months to restore full ROM and quadriceps strength. Six months after surgery, patients were allowed to return to all daily activities. MRI Evaluation All postoperative MRI scans were performed to evaluate the cause of persistent knee pain, except for 4 cases undergoing an initial routine check-up and 4 cases undergoing additional procedures. All MRI scans were taken before second-look arthroscopy was performed, except in 1 case. One case with revision anterior cruciate ligament (ACL) reconstruction underwent postoperative MRI 3 days after second-look arthroscopy. All MRI examinations were performed using a conventional 1.5-T magnetic resonance (MR) scanner (Siemens Medical Solutions, Malvern, PA) or 3.0-T MR scanner (Philips Medical Systems, Andover, MA). The 1.5-T MR scanner was used in 45% of cases (9 of 20) before March 2008, and the 3.0-T MR scanner was used in 55% of cases (11 of 20) thereafter. Sagittal T2- and T1-weighted spin-echo images and sagittal/coronal proton densityeweighted fast spin-echo images were obtained using following parameters: (1) T1-weighted imagesdrepetition time (TR) of 530 milliseconds and echo time (TE) of 12 milliseconds; (2) T2-weighted imagesdTR of 3,900 milliseconds and TE of 105 milliseconds; (3) proton densityeweighted imagesdTR of 2,760 milliseconds and TE of 15 milliseconds; and (4) gradient-echo imagesdTR of 700 milliseconds, TE of 18 milliseconds, 20 flip angle with 1 echo train length, 3-mm slice thickness, 0.6- to 0.9-mm intervals, 512 512 matrix, and 16-cm field of view. When we used the 3.0-T MR scanner, sagittal images were obtained at a 1.5-mm slice thickness and coronal images were obtained at a 2.5-mm slice thickness.
861
MENISCAL ALLOGRAFT TRANSPLANTATION
The MRI results regarding graft status after MAT are described using the criteria of Stoller et al.20 Grade 1 signal intensity is defined as single or multiple punctiform hypersignals that do not extend to the joint surface. Grade 2 signal intensity is defined as a linear intrameniscal hypersignal that does not extend to the joint surface. Grade 3 signal intensity is defined as a linear hypersignal that extends to at least 1 of the 2 joint surfaces. Grade 3 MRI signal intensity at the graft was radiologically defined as a meniscal tear. The meniscus was anatomically divided into 3 segments: anterior one-third, mid body, and posterior one-third. These respective segments were assessed using MRI and second-look arthroscopy. Because grade 1 and grade 2 MRI signals cannot be directly visualized on arthroscopy, it was impossible to compare grade 1 and 2 MRI signals with second-look arthroscopy. We evaluated whether grade 3 signal intensity on MRI would present as a meniscal tear on second-look arthroscopy. We also evaluated whether extrusion was associated with a tear. The postoperative MRI findings were compared with second-look arthroscopic findings, used as the gold standard. True-positive, true-negative, false-positive, and false-negative results, as well as the sensitivity, specificity, positive predictive value (PPV), negative predictive value, and accuracy, were assessed for the diagnostic value of postoperative MRI. The postoperative MRI findings were reviewed by a single musculoskeletal radiologist with 10 years of clinical experience and 2 orthopaedic surgeons (J-M.K., J-M.K.) with more than 8 years of knee MRI experience excluding the single operator. When all interpreters reached a consensus through discussion, MRI findings were accepted for this study. Second-Look Arthroscopy Second-look arthroscopy was initially performed in 4 patients as part of routine check-up within 1 year after MAT. However, many patients without pain were reluctant to undergo another invasive procedure. Thereafter second-look arthroscopy was performed when there was persistent knee pain or MRI results suggestive of allograft tears. In addition, second-look arthroscopy was included as part of an additional procedure, such as ACL reconstruction (2 patients), revision ACL reconstruction (1 patient), or osteochondral autograft transfer (1 patient). The intact transplanted allograft was assessed according to the following features: stability at the meniscocapsular junction, normal gross appearance, and no graft tears. When graft tears were confirmed using secondlook arthroscopy, patients underwent treatment with trimming, partial meniscectomy, or subtotal/total meniscectomy according to the extent and type of tear.
Statistics Statistical analyses were performed using SPSS software for Windows, version 16.0 (IBM, Armonk, NY). The data for quantitative variables were given as mean standard deviation. The demographic data of the lateral graft group and medial graft group were assessed with the Mann-Whitney test for continuous variables and Fisher exact test for categorical variables. To compare the lateral graft group and medial graft group regarding the diagnostic value of MRI, including sensitivity, specificity, PPV, negative predictive value, and accuracy, according to graft segments, the Fisher exact test was used. The comparison of diagnostic MRI values between the 1.5-T and 3.0-T MRI groups was also assessed with the Fisher exact test. Differences in the mean values of meniscal extrusion between the graft tear group and the group with no graft tears were assessed with the paired t test. In this study P <.05 was considered statistically significant.
Results Among 25 knees that underwent second-look arthroscopy for the evaluation of meniscal grafts, we excluded 2 knees with no postoperative MRI and 3 knees with a time gap between postoperative MRI and second-look arthroscopy of more than 3 months. Therefore 20 knees were retrospectively enrolled in this study. The study group was composed of 16 male and 4 female patients. Five medial and 15 lateral allograft menisci were assessed. The mean age of the study population was 32.1 years (SD, 8.4 years). The mean time between MAT and second-look arthroscopy was 17.4 months (SD, 11.0 months; range, 7.0 to 49.5 months). The mean time between MAT and postoperative MRI was 15.9 months (SD, 10.8 months; range, 6.0 to 46.4 months). The mean time between postoperative MRI and second-look arthroscopy was 1.6 months (SD, 1.1 months; range, 0.1 to 3.0 months). The demographic data of the lateral graft and medial graft groups are shown in Table 1. Regarding the diagnostic values of MRI, the anterior one-third of meniscal grafts showed lower specificity, PPV, and accuracy (35.3%, 21.4%, and 45%, respectively) compared with the mid body and posterior one-third of meniscal grafts, which both showed high diagnostic values (Table 2). Eleven cases were misdiagnosed with tears at the anterior one-third on MRI (Fig 1). When this group was divided into a lateral graft group and medial graft group, the respective groups also showed low specificity, PPV, and accuracy. There was no significant difference in the diagnostic values of MRI between these 2 groups (Table 2). The diagnostic values of 1.5-T and 3.0-T MRI were also evaluated. These 2 MRI groups showed low specificity, PPV, and accuracy at the anterior one-third. These 2 groups showed no
862
J-M. KIM ET AL.
Table 1. Demographic Data of Patients No. of patients Age at MAT, yr Sex (M/F), n BMI, kg/m2 Time between MAT and postoperative MRI, mo Time between MAT and second-look arthroscopy, mo Time between postoperative MRI and second-look arthroscopy, mo
Overall 20 32.1 8.42 16/4 24.9 3.0 15.9 10.8 17.4 11.0 1.6 1.1
Lateral Graft Group 15 33.7 8.3 13/2 24.5 3.4 14.0 8.9 15.6 8.8 1.6 1.0
Medial Graft Group 5 26.6 6.9 3/2 25.8 1.6 19.0 14.6 23.0 15.5 1.6 1.5
P Value .10 .25 .32 .36 .32 .76
NOTE. Data presented as mean standard deviation unless otherwise indicated. BMI, body mass index; F, female; M, male; MAT, meniscal allograft transplantation; MRI, magnetic resonance imaging.
significant difference in all diagnostic values of MRI according to graft segment (Table 3). Of 12 cases that showed grade 3 MRI signal intensity at the mid body or posterior one-third, 83.3% (10 of 12) had at least 1 symptom or physical sign, such as swelling, joint-line tenderness, locking or blocking, or a positive McMurray test. Of these, 90% (9 of 10, with a lateral graft in 8 cases and a medial graft in 1 case) had a tear on second-look arthroscopy. However, 8.3% of the cases that had no clinical sign (2 of 12) showed a small flap tear at the posterior one-third of the medial graft and shrinkage of the lateral graft. Of 5 cases that showed grade 3 MRI signal intensity only at the anterior one-third section, 20% (1 of 5) had swelling and 20% (1 of 5) had joint-line tenderness, whereas 60% (3 of 5) had no clinical signs. There were no isolated tears of the anterior one-third section on second-look arthroscopy. Ten cases showed graft tears. The graft tears showed a variety of patterns, anatomic locations, and histologic zones. The details of these cases are shown in Table 4. Of these cases, 50% (5 of 10) showed meniscal extrusion (>3 mm). In 10 cases with no graft tear, 50% (5 of 10) showed meniscal extrusion (>3 mm). The mean extrusion of the graft tear group was 3.58 mm (SD, 1.40 mm; range, 1.69 to 6.70 mm), and the mean extrusion of the group with no graft tears was 3.51 mm (SD, 1.20 mm; range, 1.53 to 5.06 mm). There was no significant difference in the
comparison of meniscal extrusion between these 2 groups (P ¼ .90). In the graft tear group, 90% (9 of 10) had ICRS grade 3 cartilage of the femur or tibia. The mean grades of the femur and tibia were 2.7 (SD, 0.5; range, 2 to 3) and 2.8 (SD, 0.6; range, 2 to 3), respectively. In the group with no graft tears, 40% (4 of 10) had ICRS grade 3 cartilage of the femur or tibia. The mean grades of the femur and tibia were 2.1 (SD, 0.6; range, 1 to 3) and 2.1 (SD, 0.7; range, 1 to 3), respectively, in the group with no graft tears. The graft tear group showed significantly higher ICRS cartilage grades (P ¼ .02 for femur and P ¼ .04 for tibia).
Discussion The most important finding of this study was that the anterior one-third of meniscal grafts showed lower specificity, PPV, and accuracy on MRI in comparison with the mid body and posterior one-third. We also found that MRI tended to overestimate tears at the anterior one-third when the criteria of Stoller et al.20 were applied. The combined application of clinical signs including swelling, locking, joint-line tenderness, and the McMurray test might be useful in the diagnosis of a graft tear. MAT shows valuable clinical outcomes, including pain relief and functional improvement. However, because these clinical results are based on subjective evaluation, it is difficult to compare results between
Table 2. Diagnostic Results of Postoperative MRI According to Graft Side Meniscal Segment Anterior one-third
P value Mid body
P value Posterior one-third
P value
Graft Side (n) All (20) Lateral graft (15) Medial graft (5)
TP, n 3 3 0
TN, n 6 5 1
FP, n 11 7 4
FN, n 0 0 0
All (20) Lateral graft (15) Medial graft (5)
7 7 0
11 7 4
1 1 0
1 0 1
All (20) Lateral graft (15) Medial graft (5)
10 9 1
9 5 4
1 1 0
0 0 0
Sensitivity 100% 100% d d 87.5% 100% 0% .13 100% 100% 100% d
Specificity 35.3% 41.7% 20% .60 91.7% 87.5% 100% > .99 90% 83.3% 100% > .99
PPV 21.4% 30% 0% .51 87.5% 87.5% d d 90.9% 90% 100% > .99
NPV 100% 100% 100% d 91.7% 100% 80% .42 100% 100% 100% d
Accuracy 45% 53% 20% .32 90% 93.3% 80% .45 95% 93.3% 100% > .99
FN, false negative; FP, false positive; MRI, magnetic resonance imaging; NPV, negative predictive value; PPV, positive predictive value; TN, true negative; TP, true positive.
863
MENISCAL ALLOGRAFT TRANSPLANTATION
Fig 1. A 24-year-old male patient who underwent left knee medial meniscal allograft transplantation and anterior cruciate ligament reconstruction. (A) At the follow-up examination at 49.5 months, sagittal proton densityeweighted magnetic resonance imaging (MRI) showed grade 3 signal intensity at the anterior one-third (circle). (B) Sagittal T2-weighted MRI showed grade 3 signal intensity at the anterior one-third (circle). (C) Good stability and no tear in the anterior one-third were verified using second-look arthroscopy (circle). (MFC, medial femoral condyle; MM, medial meniscal graft; MTP, medial tibial plateau.)
individual studies. Therefore the importance of objective evaluation is on the rise. Accordingly, objective evaluations such as MRI and second-look arthroscopy have been applied to determine clinical results after MAT.4,6,7,10,11,15,16,19,21 Although second-look arthroscopy is accepted as the gold standard for evaluating postoperative results after knee surgery, it is an invasive procedure.12,13,21,22 Accordingly, MRI has been proposed as an alternative for evaluating graft status after MAT.6,7,11,13,15,16 The accuracy of MRI for assessing graft status after MAT remains uncertain. In this study we used the criteria of Stoller et al.20 to evaluate postoperative meniscal allografts. Although these criteria have been established as useful criteria for evaluating the degeneration of the native meniscus on MRI,23,24 they may not be relevant when describing graft status after MAT. In fact, we often observed that the postoperative MRI findings did not match the second-look arthroscopic
findings at the anterior third of transplanted meniscal allografts. We hypothesized that the consistency of these 2 diagnostic methods could be different according to the segments of the meniscus. To investigate this hypothesis, a direct comparison of postoperative MRI findings and second-look arthroscopic findings was carried out in this study. Several authors have reported that high signal intensity can be observed in grafts after MAT.15,25-27 Verdonk et al.3 reported that of 17 patients who underwent MRI at a mean of 12 years after MAT, 10 (59%) showed grade 3 signal intensity. Using MRI within the first postoperative year, Lee et al.10 evaluated 43 patients who underwent MAT. Higher signals were observed for the transplanted allograft in comparison with the normal meniscus. In particular, the anterior horn showed higher signal intensity than the posterior horn throughout the first year after MAT. Lee et al. also noted that persistently increased
Table 3. Diagnostic Results of Postoperative MRI According to MRI Type Meniscal Segment Anterior one-third
P value Mid body
P value Posterior one-third
P value
MRI Type (n) All (20) 1.5-T MRI (9) 3.0-T MRI (11)
TP, n 3 3 0
TN, n 6 3 3
FP, n 11 3 8
FN, n 0 0 0
All (20) 1.5-T MRI (9) 3.0-T MRI (11)
7 3 4
11 6 5
1 0 1
1 0 1
All (20) 1.5-T MRI (15) 3.0-T MRI (5)
10 5 1
9 4 4
1 0 0
0 0 0
Sensitivity 100% 100% d d 87.%5 100% 80% > .99 100% 100% 100% d
Specificity 35.3% 50% 27.3% .60 91.7% 100% 83.3% > .99 90% 100% 83.3% > .99
PPV 21.4% 50% 0% .06 87.5% 100% 80% > .99 90.9% 100% 83.3% > .99
NPV 100% 100% 100% d 91.7% 100% 83.8% > .99 100% 100% 100% d
Accuracy 45% 66.6% 27.3% .18 90% 93.3% 81.8% .48 95% 100% 90.9% > .99
FN, false negative; FP, false positive; MRI, magnetic resonance imaging; NPV, negative predictive value; PPV, positive predictive value; TN, true negative; TP, true positive.
864 White zone Red-white zone Red zone Red-white zone Red zone Red zone Red zone Small flap Flap Longitudinal Complex Longitudinal Longitudinal B-H 2.96 2.91 3.84 4.60 2.32 3.84 6.70 3 2 2 3 3 3 2 2 3 3 3 2 3 2 2 1 2 0 0 1 2 21.9 23.0 25.8 23.3 25.7 27.8 24.4 Lateral Medial Lateral Lateral Lateral Lateral Lateral 26 35 26 37 43 40 32 M M F M M M M 4 5 6 7 8 9 10
ant, anterior; B-H, bucket handle; BMI, body mass index; F, female; ICRS, International Cartilage Repair Society; M, male; MAT, meniscal allograft transplantation; post, posterior. *Positive values indicate varus, whereas negative values indicate valgus.
Red-white zone Red-white zone Flap Horizontal 4.09 1.69 3 3 3 3 2 1 23.1 22.3 Lateral Lateral M M 2 3
41 21
Tibia 3 Femur 3 BMI, kg/m2 31.1 Graft Side Lateral Age at MAT, yr 43 Sex M Case No. 1
Table 4. Details of 10 Cases With Meniscal Graft Tears
Alignment* 3
ICRS Grade
Amount of Extrusion, mm 2.89
Anatomic Location Ant one-third þ mid body þ post one-third Post one-third Ant one-third þ mid body þ post one-third Post one-third Mid body þ post one-third Mid body þ post one-third Mid body þ post one-third Mid body þ post one-third Mid body þ post one-third Ant one-third þ mid body þ post one-third
Tear Pattern B-H
Histologic Location Red zone
J-M. KIM ET AL.
intrameniscal intensity might indicate normal healing processes rather than degenerative changes or tears. We think that this phenomenon may often be observed at the anterior one-third because most surgical trauma occurs at the anterior portion during MAT. MRI provides only limited evaluations of postoperative status after meniscal repair because scarring at the repaired site can present as a grade 3 signal.13,22,28 Miao et al.14 reported that sagittal T2-weighted imaging shows the highest diagnostic accuracy (60.5%) and sagittal T1-weighted imaging shows the lowest diagnostic accuracy (28.9%) in comparison with second-look arthroscopy (38 repaired menisci were evaluated using postoperative MRI at 15 months [SD, 5.1 months] after meniscal repair). In a different study, Miao et al.13 also noted that second-look arthroscopy is the most reliable way to determine meniscal healing and that the combined application of different sequences on MRI could improve its diagnostic value. Accordingly, these studies showed the limitations of MRI, even when evaluating native meniscal healing after meniscal repair. Because there are few previous studies that have compared postoperative MRI and second-look arthroscopy after MAT21,25 and these studies did not directly compare both methods, it is difficult to compare our results with those of other studies. We found that the diagnostic values of postoperative MRI after MAT can be affected by the segments of the graft. The anterior one-third showed low specificity, PPV, and accuracy compared with the mid body and posterior body. These results indicate that MRI can reliably assess the mid body and posterior one-third of the transplanted graft but not the anterior one-third. Grade 3 signal intensity at the anterior one-third could be misdiagnosed as a tear when using the criteria of Stoller et al.20 According to our study results, grade 3 signal intensity at the anterior one-third of a graft does not necessarily indicate a graft tear (according to the criteria of Stoller et al.) but grade 3 signal intensity at the mid body and posterior onethird could be associated with graft tears. Therefore we believe that the criteria of Stoller et al. are unsuitable for interpreting allograft meniscus. New evaluation systems that can more reliably interpret graft status on MRI are needed. For the clinical assessment of meniscal repair, Barrett et al.29 performed physical examination to assess the following clinical signs: swelling, joint-line tenderness, locking, and McMurray test. Absence of the 4 signs indicated a healed meniscus. Miao et al.13 also used these clinical signs for the evaluation of meniscal repair. They noted that the healing rate was 70.8%, significantly lower than the healing rate found by
865
MENISCAL ALLOGRAFT TRANSPLANTATION
second-look arthroscopy. When a repaired meniscus with 2 or more signs was defined as unhealed, the diagnostic value could be improved. Similarly, our study found that there were no graft tears in 5 cases that showed grade 3 signal intensity at only the anterior one-third without clinical signs or with 1 clinical sign. We believed that the combination of postoperative MRI and clinical assessment could reduce the rate of misdiagnosis of graft tears. The clinical relevance of our study is that taking into account both the characteristics of MRI findings at the anterior one-third and the clinical assessment findings, unnecessary planning of second-look arthroscopy may be avoided in patients with abnormally high signal intensity in the anterior one-third of the meniscal allograft. Limitations This study has several limitations. First, this is a retrospective study without a control group, and the number of cases was small. The 20 patients enrolled in this study may not represent the entire group of 290 patients. This study did not include the MR findings of the remaining large group of patients who were not selected for surgery. Second, second-look arthroscopy was not performed according to consistent surgical indications, which might affect the homogeneity of our study group. These factors could induce selection bias, which might yield an inability to comment on the diagnostic values of MRI for the remaining group. Third, 1.5-T and 3.0-T MRI scanners were used to evaluate the status of the meniscal grafts after MAT. MRI findings associated with second-look arthroscopy can be affected by the type of MRI scanner, which might induce bias regarding the diagnostic results. However, our study showed that there was no difference in the comparison of diagnostic MRI values between these 2 MRI scanners. Lastly, the relations between the pattern and degree of grade 3 signal intensity and second-look arthroscopic findings were not discussed sufficiently. We will consider these limitations in subsequent research on new criteria regarding graft status on postoperative MRI after MAT.
Conclusions The anterior one-third of a transplanted meniscal allograft showed lower specificity, PPV, and accuracy of postoperative MRI compared with the mid body and posterior one-third. When the criteria of Stoller et al.20 were used to evaluate grafts after MAT, MRI tended to grade the anterior one-third more poorly than secondlook arthroscopy. These features should be considered when evaluating transplanted meniscal allografts on postoperative MRI.
References 1. LaPrade RF, Wills NJ, Spiridonov SI, Perkinson S. A prospective outcomes study of meniscal allograft transplantation. Am J Sports Med 2010;38:1804-1812. 2. Wilcox TR, Goble EM. Indications for meniscal allograft reconstruction. Am J Knee Surg 1996;9:35-36. 3. Verdonk PC, Verstraete KL, Almqvist KF, et al. Meniscal allograft transplantation: Long-term clinical results with radiological and magnetic resonance imaging correlations. Knee Surg Sports Traumatol Arthrosc 2006;14:694-706. 4. Ha JK, Shim JC, Kim DW, Lee YS, Ra HJ, Kim JG. Relationship between meniscal extrusion and various clinical findings after meniscus allograft transplantation. Am J Sports Med 2010;38:2448-2455. 5. Noyes FR, Barber-Westin SD. Prospective evaluation of allograft meniscus transplantation: A minimum 2-year follow-up. Am J Sports Med 2006;34:2038-2039 (author reply 2039). 6. Kim JM, Lee BS, Kim KH, Kim KA, Bin SI. Results of meniscus allograft transplantation using bone fixation: 110 cases with objective evaluation. Am J Sports Med 2012;40:1027-1034. 7. Lee DH, Kim SB, Kim TH, Cha EJ, Bin SI. Midterm outcomes after meniscal allograft transplantation: Comparison of cases with extrusion versus without extrusion. Am J Sports Med 2010;38:247-254. 8. Stollsteimer GT, Shelton WR, Dukes A, Bomboy AL. Meniscal allograft transplantation: A 1- to 5-year followup of 22 patients. Arthroscopy 2000;16:343-347. 9. Graf KW Jr, Sekiya JK, Wojtys EM. Long-term results after combined medial meniscal allograft transplantation and anterior cruciate ligament reconstruction: Minimum 8.5-year follow-up study. Arthroscopy 2004;20:129-140. 10. Lee DH, Lee BS, Chung JW, et al. Changes in magnetic resonance imaging signal intensity of transplanted meniscus allografts are not associated with clinical outcomes. Arthroscopy 2011;27:1211-1218. 11. Lee BS, Kim JM, Sohn DW, Bin SI. Review of meniscal allograft transplantation focusing on long-term results and evaluation methods. Knee Surg Relat Res 2013;25:1-6. 12. Morgan CD, Wojtys EM, Casscells CD, Casscells SW. Arthroscopic meniscal repair evaluated by second-look arthroscopy. Am J Sports Med 1991;19:632-637 (discussion 637-638). 13. Miao Y, Yu JK, Ao YF, Zheng ZZ, Gong X, Leung KK. Diagnostic values of 3 methods for evaluating meniscal healing status after meniscal repair: Comparison among second-look arthroscopy, clinical assessment, and magnetic resonance imaging. Am J Sports Med 2011;39: 735-742. 14. Miao Y, Yu JK, Zheng ZZ, et al. MRI signal changes in completely healed meniscus confirmed by second-look arthroscopy after meniscal repair with bioabsorbable arrows. Knee Surg Sports Traumatol Arthrosc 2009;17: 622-630. 15. Lee DH, Kim TH, Lee SH, Kim CW, Kim JM, Bin SI. Evaluation of meniscus allograft transplantation with serial magnetic resonance imaging during the first postoperative year: Focus on graft extrusion. Arthroscopy 2008;24:1115-1121.
866
J-M. KIM ET AL.
16. Lee BS, Chung JW, Kim JM, Cho WJ, Kim KA, Bin SI. Morphologic changes in fresh-frozen meniscus allografts over 1 year: A prospective magnetic resonance imaging study on the width and thickness of transplants. Am J Sports Med 2012;40:1384-1391. 17. Shepard MF, Hunter DM, Davies MR, Shapiro MS, Seeger LL. The clinical significance of anterior horn meniscal tears diagnosed on magnetic resonance images. Am J Sports Med 2002;30:189-192. 18. Pollard ME, Kang Q, Berg EE. Radiographic sizing for meniscal transplantation. Arthroscopy 1995;11:684-687. 19. Gonzalez-Lucena G, Gelber PE, Pelfort X, Tey M, Monllau JC. Meniscal allograft transplantation without bone blocks: A 5- to 8-year follow-up of 33 patients. Arthroscopy 2010;26:1633-1640. 20. Stoller DW, Cannon WD, Anderson LJ. Magnetic resonance imaging in orthopaedics and sports medicine, Ed 2. Philadelphia: Lippincott Williams & Wilkins, 1997. 21. van Arkel ER, Goei R, de Ploeg I, de Boer HH. Meniscal allografts: Evaluation with magnetic resonance imaging and correlation with arthroscopy. Arthroscopy 2000;16: 517-521. 22. Steenbrugge F, Verstraete K, Verdonk R. Magnetic resonance imaging of the surgically repaired meniscus: A
23.
24.
25.
26.
27.
28.
29.
13-year follow-up study of 13 knees. Acta Orthop Scand 2004;75:323-327. Crues JV III, Mink J, Levy TL, Lotysch M, Stoller DW. Meniscal tears of the knee: Accuracy of MR imaging. Radiology 1987;164:445-448. Stoller DW, Martin C, Crues JV III, Kaplan L, Mink JH. Meniscal tears: Pathologic correlation with MR imaging. Radiology 1987;163:731-735. Potter HG, Rodeo SA, Wickiewicz TL, Warren RF. MR imaging of meniscal allografts: Correlation with clinical and arthroscopic outcomes. Radiology 1996;198:509-514. Noyes FR, Barber-Westin SD, Rankin M. Meniscal transplantation in symptomatic patients less than fifty years old. J Bone Joint Surg Am 2004;86:1392-1404. Verstraete KL, Verdonk R, Lootens T, Verstraete P, De Rooy J, Kunnen M. Current status and imaging of allograft meniscal transplantation. Eur J Radiol 1997;26: 16-22. Mustonen AO, Tielinen L, Lindahl J, Hirvensalo E, Kiuru M, Koskinen SK. MRI of menisci repaired with bioabsorbable arrows. Skeletal Radiol 2006;35:515-521. Barrett GR, Field MH, Treacy SH, Ruff CG. Clinical results of meniscus repair in patients 40 years and older. Arthroscopy 1998;14:824-829.