- Email: [email protected]
T2 relaxation times of the anterolateral femoral cartilage in patients after ACL-reconstruction with and without a deep lateral femoral notch sign
European Journal of Radiology 106 (2018) 85–91
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Research article
T2 relaxation times of the anterolateral femoral cartilage in patients after ACL-reconstruction with and without a deep lateral femoral notch sign
T
⁎
Cyrus Behzadia, , Goetz H. Welschb, Jan-Philipp Petersend, Bjoern P. Schoennagela, Peter Bannasa, Michael G. Kaula, Gerhard Schoenc, Josephine Berger-Grochd, Gerhard Adama, Marc Regiera a
Department of Diagnostic and Interventional Radiology and Nuclear Medicine, University Medical Center Hamburg-Eppendorf, Hamburg, 20246, Germany Department of Athletics and Sports Medicine, University Medical Center Hamburg-Eppendorf, Hamburg, 20246, Germany c Department of Medical Biometry and Epidemiology, University Medical Center Hamburg-Eppendorf, Hamburg, 20246, Germany d Department of Trauma, Hand and Reconstructive Surgery, University Medical Center Hamburg-Eppendorf, Hamburg, 20246, Germany b
A R T I C LE I N FO
A B S T R A C T
Keywords: Quantitative MRI T2 relaxation time Articular cartilage Deep lateral femoral notch sign Anterior cruciate ligament rupture
Purpose: To quantitatively assess T2 relaxation times of the anterolateral femoral cartilage following anterior cruciate ligament (ACL)-reconstruction with and without a positive deep lateral femoral notch sign (DLNS) at post-traumatic MRI. Materials and Methods: In 52 patients post-traumatic MRI as well as 12 months after ACL-rupture (ACLR) and surgical treatment were analysed. In 28 patients a positive DLNS was present at post-traumatic MRI. For quantitative analysis, T2 relaxation time measurements (7 TE: 10–70 ms) were performed at time of reevaluation. Three polygonal ROIs encompassing the full cartilage layer were placed in the anterolateral femoral cartilage. Clinical assessment included Lysholm-Tegner-Activity-Score, Rasmussen's clinical score and modified Cincinnati-Rating-System-Questionnaire. Description and differences were calculated as means and confidence intervals of means, controlled for the cluster effect of person, if appropriate. Results: In patients with a positive DLNS after ACLR, relaxation times in the notch region were significantly prolonged compared to patients without a positive DLNS (Δ 7.4 ms, CI: 5.6–9.2; p-value < 0.001) as well as to the adjacent anterior (Δ 5.7 ms, CI: 4.7–6.7; p-value < 0.001) and central femoral cartilage (Δ 6.6 ms, CI: 5.7–7.6; p-value < 0.001). Solely insignificant differences were noticed in the performed clinical scores comparing the two groups (p > 0.05). Conclusion: Significantly prolonged T2 relaxation times of the anterolateral femoral cartilage were found in patients with a positive DLNS following ACL-reconstruction compared to patients without a DLNS. Based on these results, it has to be assumed that a positive DLNS is associated with higher cartilage degradation.
1. Introduction Acute rupture of the anterior cruciate ligament (ACL) is one of the most feared injuries in sports overall. Considering the high incidence of Anterior-Cruciate-LigamentRuptures (ACLRs) of more than 200.000 injuries in the US alone [1,2], the immense relevance becomes evident. The impact on every athlete’s career is underlined by reviewing the therapeutic consequences after ACLRs. Usually, post-traumatic surgery and at least six months of rehabilitation are mandatory in order to return to regular athletic activity
[3]. However, it is reported that only 48% of elite Australian football players could return-to-play in less than one year [4]. Furthermore, studies could demonstrate that in the majority of patients after an ACLR signs of osteoarthritis (OA) are detectable within 10–15 years [5,6]. In post-traumatic diagnostic evaluation of ACLR, concomitant injuries are presumably to be diagnosed. In approximately half of all patients suffering from ACLR, an articular cartilage injury is noted [7]. In an extensive trauma mechanism, due to a dorsolateral movement of the femur and a fixed lower leg ("pivot shift manoeuvre"), a
⁎ Corresponding author at: Center for Radiology and Endoscopy, Department of Diagnostic and Interventional Radiology and Nuclearmedicine, University Medical Center Hamburg-Eppendorf, Martinistrasse 52, 20246, Hamburg, Germany. E-mail address: [email protected] (C. Behzadi).
https://doi.org/10.1016/j.ejrad.2018.07.007 Received 9 March 2018; Received in revised form 24 May 2018; Accepted 8 July 2018 0720-048X/ © 2018 Elsevier B.V. All rights reserved.
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
system (Ingenia, Philips, Best, the Netherlands). A 16-channel knee coil was used for signal reception. The investigated knee was placed in the centre of the coil and sandbags positioned on patient’s leg.
transchondral impression of the anterolateral femoral condyle versus the dorsolateral tibial plateau can result. This constellation can be detected at post-traumatic radiography (lateral view), known as the deep lateral notch (sulcus) sign (DLNS) [8]. It has been reported that in high-pivoting sports in about 33% of patients after acute ACLR, a positive DLNS deeper than 2 mm can be observed [9]. At MRI a DLNS is accompanied by subchondral bone marrow edema in the distal femur in about 50% of patients [9]. In the recent decade various techniques have drawn attention to quantitative MRI measurements of articular cartilage. Techniques such as T1rho or T2/T2* have demonstrated that early stages of OA might be diagnosed prior to irreversible stages [10–13]. T2 mapping is based on the detection of increased water observed in cartilage degradation due to biochemical changes in the extracellular matrix (e.g. collagen fraction). High sensitivity rates for the indirect assessment of early stages of OA for T2 mapping have been published in various studies [10,28]. In our investigation of post-operative femorotibial joints in patients after ACLR, we decided to include T2 mapping as the most validated technique for quantitative measurements. T2 mapping has been used in many clinical trials for it offers several advantages (e.g. no i.v. contrast media required, robust against susceptibility artifacts) while being highly sensitive to biochemical changes in articular cartilage composition [10,14,15]. Progressive cartilage degradation has been generally confirmed in recent publications after ACLR up to two years after surgery compared to healthy individuals [16–18]. However, to the best of our knowledge, no dedicated quantitative analysis of the DLNS has been performed yet. Therefore, the purpose of the presented study was to analyse and compare T2 relaxation times in patients suffering from ACLR with and without a DLNS. We intended to evaluate if the DLNS is accompanied with altered relaxation times indicating a more severe traumatic event and higher cartilage degradation. Therefore, all patients were re-evaluated including quantitative T2 measurements as well as several clinical kneescores 12 months post-trauma.
2.1.2. Morphological sequences Sagittally oriented 3D fat-saturated proton-density weighted (Pdw) Turbo Spin Echo (TSE) sequence was performed and reformations subsequently generated. The parameters were kept identical at both time points: Time to repetition (TR) 1300 ms; echo-time (TE) 3.7 ms; flip angle: 90°, slice thickness: 0.6 mm (in total: 330 slices), field of view (FoV): 185 × 185 mm, matrix: 320 × 320, scan time: 8:24 min. In addition, a coronal T1w TSE sequence was acquired (TR: 818 ms, TE: 10.8 ms, flip angle: 90°, slice thickness: 2.2 mm (in total: 60 slices), matrix: 672 × 672, FoV: 180 × 180 mm, scan time 1:50 min. For quantitative measurements, a 3D multi-echo T2w Turbo Spin Echo (TSE) sequence was used (TR: 192.8 ms, 7 echo times (10–70 ms), flip angle: 90°, slice thickness: 3 mm, matrix: 320 × 320, FoV of 160 × 160 mm, Sense factor: 4.4, scan time: 6:15 min). T2 mapping was performed by fitting a monoexponential function A*exp(–TE/T2) to the multiecho data using an in-house quantification plugin (qMapIt) extending ImageJ (National Institutes of Health, Bethesda, MD). The first echo was neglected because it does not contain any signal of a stimulated echo in contrast to the later ones. 2.1.3. Morphological image analysis All patients underwent biplane radiography (anterior-posterior and lateral view) of the injured knee in the emergency department at the day of trauma. These were evaluated in accordance to the KellgrenLawrence grading for signs of OA [19]. All MRIs between 2014 and 2015 were scanned for ACLRs by two radiologists with special interest in sports imaging (five and 12 years of experience). If both radiologists diagnosed ACLR, MRIs were re-evaluated for additional injuries. If the readers had divergent opinion concerning the status of the ACL or concomitant injuries, they reviewed the anonymized data sets and formed a consensual diagnosis [2,20,21]. Morphological analysis was performed by using a commercially available post processing workstation (Extended Brilliance Workspace, Version 2.0, Philips Healthcare, Best, the Netherlands). Concomitant injuries were noted as listed below: a) impression of the anterolateral femoral condyle, b) bone marrow edema, c) rupture of the collateral ligaments, d) cartilage lesions according to Noyes classification higher than grade 2 [22], e) rupture of the retinaculum patellae, f) injury of the Hoffa fat pad, g) fracture, h) meniscal injuries (lesions of the anterior or posterior horn as well as bucket handle tears, i) rupture of the patellar tendon. In patients with a DLNS sign, maximum length of impression in each patient was measured and the mean length for all impressions calculated (in mm). The depth of the impression (in mm) was measured perpendicular to the surface of the articular cartilage of the femoral condyle in sagittal orientation and categorized as follows: I: 0 - < 1 mm; II: 1 - < 2 mm, III: 2 - < 3 mm; IV: 3 - < 4 mm; V: ≥ 4 mm. Clinical evaluation at re-examination included three different Knee scoring systems (modified Lysholm-Tegner-Activity-Score, Rasmussen's clinical score and modified Cincinnati Rating System Questionnaire) [23–25]. Apart from the time-point at re-evaluation, all patients had to complete the scores retrospectively concerning their immediate posttraumatic status.
2. Materials & methods Approval from the local institutional review board was received prior to initiation of the study. Informed patient consent was obtained from all patients before image acquisition. 2.1. Study population and Inclusion/Exclusion criteria All patients in our database, who were diagnosed with ACLR (between 2014–2015) prior to the study were contacted and asked to voluntarily participate in our study. In total, 59 out of 124 patients took part in our investigation (47%). Four patients refused an MRI examination at re-evaluation and were consequently excluded. Three patients were excluded because their knees did not fit inside the 16channel knee coil. Therefore, the quantitative imaging protocol was not applicable. Therefore, 52 patients (18 female, 34 male; mean age at reevaluation: 32.5 years) were enrolled. Exclusion criteria were: 1.) Surgery prior to ACLR on the injured knee, 2.) Acute injury or history of major injury after ACLR (e.g. fracture) or 3.) Systemic diseases like e.g. rheumatoid arthritis. All scans at time of re-evaluation were conducted between August 2015 and April 2016.
2.2. Quantitative analysis
2.1.1. Image acquisition All patients underwent morphological and quantitative MRI imaging protocol at re-evaluation. Morphological sequences were acquired prior to T2 relaxation measurements. The order was kept constant in all patients. At both time points, all MRIs were performed using a 3 T MRI
Three cartilage regions of the lateral femoral condyle were analysed in all participants (the anterolateral as well as the bordering anterior and central cartilage, Fig. 1). In patients with a positive DLNS in posttraumatic MRI, maximum length of the impression (in mm) in the anterolateral cartilage was measured. Subsequently, the ROI in the 86
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
likelihood ratio tests. Measurement such as Lysholm-Tegner-Activity-Score, Rasmussen's clinical score and modified Cincinnati Rating System Questionnaire, which were measured on a person level, were described by means and confidence intervals of means. The standard level of significance used to identify statistically significant effects was set to 0.05. All statistical analyses were performed with the statistical software package R, version 3.4.1 [37]. 3. Results 3.1. Morphological post-traumatic injuries No positive signs of initial stages of OA were diagnosed reading biplane radiography (grade 0: n = 52). At posttraumatic MRI, a positive DLNS was detected in 28 patients (54%; impression < 1 mm: n = 9; 1 - < 2 mm: n = 11; 2 − < 3 mm: n = 3; 3 − < 4 mm: n = 3; ≥4 mm: n = 2). The mean length of the impression was calculated with 16 mm (+− 3 mm). 50 patients had a bone marrow edema in the tibial plateau or distal femur (96%). In the lateral tibial plateau three fractures were detected (6%). At follow-up, bone marrow edema had fully declined in all cases. Nine cartilage lesions (grade 3: n = 4; grade 4: n = 5) were detected in eight patients (15%). 29 meniscal lesions were diagnosed in 25 patients (48%): 16 ruptures of the posterior horn of the medial meniscus, 11 ruptures of the posterior horn of the lateral meniscus, one rupture of the anterior horn of the lateral meniscus and one bucket handle tear. One patellar tendon rupture as well as one rupture of the medial collateral ligament (2%) and two ruptures of the Hoffa fat pad (4%) were found. No fracture was detected reading post-traumatic MRI.
Fig. 1. Sagittal Pdw sequence including T2w-colour overlay in a patient with a positive DLNS. In each participant, three regions of the anterolateral femoral cartilage (A = Anterior; N = Notch Region, C = Central) were defined and subsequently analysed. As exemplarily demonstrated in this patient, elevated T2 relaxation times of the Notch Region (42.6 ± 9.1 ms) in comparison to the Anterior (35.1 ± 9.9 ms) and Central Regions (31.3 ± 5.7 ms) of the femoral cartilage could be detected.
3.2. Quantitative analysis of the anterolateral femur
anterolateral cartilage was adjusted to the initial length of the impression for each patient. For the adjacent anterior and central femoral cartilage, the identical length was chosen for ROI placements separately. In patients without a DLNS, ROI placements were performed after calculating the mean length of the impression in patients with a DLNS. Therefore, ROIs for the three investigated cartilage segments had the identical length as the impression of the anterolateral cartilage in patients with a DLNS. Due to the presence of rather thin cartilage at the anterolateral femoral condyle, we decided in all patients to place ROIs in three consecutive slices in each segment. Each ROI encovered the full depth of the cartilage layer. Therefore, 468 ROIs were placed and analysed using a dedicated ImageJ based software tool (qMapIt) by one radiologist (XX) in order to prevent interobserver measurement error as performed in previous studies and under supervision by a senior radiologist (12 years of experience in musculoskeletal imaging [13,26]. In order to ensure our quantitative analysis, ROI placement were repeated after a four-week interval in 10 randomly selected patients (5 of each cohort). Intraclass correlation coefficient (ICC) was calculated in order to compare the measurements (Rating: < 0.4: poor; 0.4–0.75: moderate; ≥0.75: good) [27,28]. Subchondral bone and joint effusion were excluded from ROI placements. The reader was blinded to patient information while performing the placements.
The ICC for intraobserver reliability in 10 randomly selected individuals was ≥0.8 in each cartilage segment. In patients with a DLNS, differences between the notch region and the anterior femoral cartilage (Δ 5.7 ms; CI: 4.7–6.7 ms; p-value: < 0.001) were highly significant. The same was found comparing the notch region to the central femoral cartilage (Δ 6.6 ms; CI: 5.7–7.6 ms; p-value: < 0.001). Mean relaxation time for the notch region was calculated with 42.0 ms (CI: 38.9–45.1 ms, Table 1). The adjacent anterior femoral cartilage revealed a mean relaxation time of 36.3 ms (CI: 33.8–38.9 ms) and the central femoral cartilage of 35.4 ms (CI: 32.8–38.0 ms). In patients without a DLNS, mean relaxation times for the anterior femoral cartilage (Δ 1.5 ms; CI: 0.5–2.5 ms; p-value: < 0.001) as well as for central femoral cartilage (Δ 0.1 ms; CI: −0.9 to 1.2 ms; p-value: 0.829) were prolonged if compared to the notch region. The comparison of the anterior to the central femoral cartilage revealed significantly higher results for the anterior femoral cartilage (Δ 1.4 ms; CI: 0.5–2.2 ms; p-value: 0.002). In patients without a DLNS, mean relaxation times for the notch region were calculated with 34.6 ms (CI: 33.3–35.9 ms). Anterior femoral cartilage revealed a mean relaxation time of 36.1 ms (CI: 35.0–37.2 ms) and the central femoral cartilage of 34.7 ms (CI: 33.6–35.9 ms). 3.3. Comparison between the two groups (positive vs. negative DLNS)
2.3. Statistical analysis
The calculated mean difference for the notch region revealed significantly elevated relaxation times in patients with a positive DLNS (Δ 7.4 ms; CI: 5.6–9.2 ms; p-value: < 0.001) compared to patients without a DLNS. In patients presenting a positive DLNS, mean relaxation time for anterior femoral cartilage were 0.2 ms longer than in patients without a DLNS (CI: −1.2 to 1.7 ms, p-value: 0.763). Differences between the two groups were insignificant for central
Due to violation of the independence assumption of single measurements, we estimated differences and confidence intervals of differences for relaxation time with random effects models. Fixed effects were "region" (anterior/notch/central) and "DLNS" (yes/no), controlled for the random effect "person". If appropriate, we conducted statistical models with interaction terms. Appropriateness was tested by 87
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
Table 1 Detailed illustration of T2 relaxation times (ms) in the anterolateral femoral cartilage in patients after ACLR with and without a positive deep lateral notch sign (DLNS).
femoral cartilage revealing slightly prolonged relaxation times in patients with a DLNS (Δ 0.7; CI: -−0.8 to 2.1 ms; p: 0.394).
model of the notch region was 37.6 ms (CI: 35.9–39.3 ms, Table 2). Differences between the five sub-categories in patients with a DLNS were insignificant (p = 0.693). For each category, a mean increase of 0.14 ms (CI: −0.54 to 0.82 ms) was detected.
3.4. Detailed evaluation of the notch region in patients with a DLNS
3.5. Evaluation of the performed clinical scores
Mean relaxation time for the intercept of the performed regression
Detailed evaluation is presented in Table 3 demonstrating insignificant differences for all scoring systems between patients with and without a DLNS at both time points (p > 0.05). Furthermore, we performed a statistical analysis in patients with a DLNS regarding any additional injuries. In patients with a DLNS, mean relaxation times between patients with and without concomitant injuries were statistically insignificant (p = 0.21). Concerning existing cartilage injuries higher than grade II at followup imaging of patients with a DLNS, differences of the mean relaxation times for this sub-cohort were as well statistically insignificant (p = 0.95).
Table 2 For the 28 patients with a positive DLNS, a special analysis revealed a constantly rising relaxation time (0.14 ms with each Millimetre; CI: -0.54 - 0.82) in comparison to the depth of the impression, although the correlation was insignificant (p = 0.693).
4. Discussion In the presented study, T2 relaxation times of patients suffering from ACLR were analysed. The main focus was to determine if a DLNS had any impact on T2 relaxation times and to compare their results to patients without a DLNS as well as to diverse clinical assessment scores. In our preliminary study analysing patients 12 months after ACLreconstruction, the presented results indicated significantly prolonged relaxation times in the anterolateral femoral cartilage in those patients with a DLNS (p < 0.001) compared to individuals without a DLNS. Interestingly, in patients with a DLNS, relaxation times of the notch region were significantly prolonged compared to the adjacent femoral 88
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
volunteers in all compartments [2]. Furthermore, they found a significant correlation between the superficial layers in T1rho and T2 relaxation times, indicating a similar value of these two diverse quantitative techniques. One might argue that their results were based on a small study collective consisting of seven men and five women only. However, in the limitation section the authors stated that the results could be considered as preliminary, encouraging further studies. In concordance to Li et al. we chose a time span of 12 months as well and re-evaluated 52 patients to compare our results. The detection of significantly prolonged relaxation times within the notch region in our patient collective further confirms their results of generally prolonged relaxation times after ACLR. Moreover, our findings strongly indicate higher biochemical cartilage alterations in patients in case of deep lateral notch sign. We believe that these quantitative measurements are capable of detecting patients at higher risk for manifestations of OA at an earlier time-point. This assumption is supported by further findings of Williams et al. [29]. They reported that initial changes in T2 relaxation times six months after ACLR were correlated to biochemical as well as morphological changes at a twoyear re-evaluation. If this impairment of biochemical cartilage composition accelerates onset of OA and subsequently leads to a reduced quality of life, has to be further evaluated in a longitudinal study design. A recently published study by Su et al. concluded that initially elevated relaxation times after ACLR might lead to a worse clinical outcome [17]. The authors analysed T1rho and T2 relaxation times after ACLR prior to surgery comparing the results to the contralateral, healthy knee of each individual. In addition, they completed at baseline as well at 6 and 12 months post surgery a clinical evaluation (KneeInjury and Osteoarthritis score (KOOS) and Marx activity level questionnaire). For instance, they observed at baseline significantly correlated side-to-side differences for T1rho and T2 relaxation times for the posterolateral tibia (p < 0.001). Interestingly, they only noticed a correlation between T2 and KOOS at baseline whereas no such association was observed at 6 and 12 months follow-up. The authors concluded that in contrast to T2 relaxation time measurement, T1rho might be a more accurate tool to longitudinally evaluate compositional changes in articular cartilage. However, their conclusions were exclusively based on the assessment of subjective questionnaires and not confirmed with quantitative measurements post-surgery. It would be of interest to further evaluate their findings in future quantitative analysis. As a further limitation, the authors mentioned that nowadays T1rho measurements are exclusively used in research settings and not as established as T2 relaxation measurements. We agree with the authors and therefore focused on T2 measurements enabling future multi-centre re-evaluations. Divergent results have been published in recent literature concerning the question if concomitant injuries in ACLR affect cartilage relaxation times. Gheno et al. investigated patients post-traumatic and six months after ACL-reconstruction and correlated T2 relaxation times to clinical scores [15]. They found prolonged relaxation times in the weight bearing and non-weight bearing femorotibial cartilage when compared to a control group. Due to their findings, one might conclude that articular cartilage is generally altered in patients after ACL-reconstruction. Furthermore, they postulated that the meniscal status does not interfere with T2 relaxation measurements after ACLR which is in contrast to previous studies [30,31]. The results of our analysis seem to support the findings by Gheno et al.. In our patient collective non-significant differences were detected when comparing patients with isolated ACLR to those presenting with concomitant injuries (p > 0.05). However, it needs to be taken into account that our investigation focused on the anterolateral femoral cartilage alone at a singular
Table 3 For the three clinical scores (LGS = modified Lysholm-Tegner-Activity-Score; RS = Rasmussen's clinical score, C = modified Cincinnati Rating System Questionnaire) no significant differences in the analysis of the immediate posttraumatic situation and at the time of re-evaluation were diagnosed (p > 0.05).
cartilage (p < 0.001) affirming the hypothesis that the DLNS represents higher cartilage degradation. However, the observed differences in our quantitative cartilage analysis could not be found likewise in the patients’ clinical symptoms. This lack of proportionality might be attributed to subtle cartilage injuries prior to clinical manifestations or the dominance of an ACLR as the major injury masking other symptoms. To the best of our knowledge, the presented study is the first to perform a specific quantitative analysis of the DLNS in patients after ACLRs. Recent publications focused on consequences of ACLR on articular cartilage of the knee have alluded to similar results revealing prolonged relaxation times after surgical treatment [16,18,29]. Therefore, we constrained our investigation on the anterolateral femoral cartilage including the notch region and the surrounding cartilage in patients with a positive DLNS and compared these results to patients without a positive DLNS. Li et al. investigated T1rho and T2 relaxation times immediately after an ACLR and at a 1-year follow up investigating 12 patients. They could successfully demonstrate that cartilage relaxation times in patients after ACLR were significantly prolonged compared to healthy 89
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
evaluation 12 months after ACLR. These cartilage areas might not be as affected as the central weight bearing cartilage areas after e.g. meniscectomy in ACLR. As a primary limitation, one might argue that we have only compared T2 relaxation times at a 1-year follow-up after surgery and did not perform a comparison to a healthy control collective. However, various investigators have recently described generally elevated relaxation times in post-traumatic MRI as well as in follow-up examinations up to 2-years after ACLR when compared to healthy individuals [16,30,32]. Even in non-operatively treated patients after various knee injuries generally elevated T2 relaxation times could be demonstrated compared to healthy individuals indicating cartilage abnormalities [33]. Therefore, our aim was to distinctively focus on the intra- and interindividual analysis of T2 relaxation times of the anterolateral and adjacent femoral cartilage in patients after ACLR with and without a DLNS. Furthermore, one might argue that the analysis of the contralateral knee could have led to an internal control ruling out physiological interindividual alterations. However, quantitative measurements of the healthy knee in patients following ACLR might have been confounded due to adapted post-traumatic kinematics [34,35]. These clinical results have also been lately confirmed by studies demonstrating generally elevated relaxation times in patients after ACLR compared to the uninjured contralateral knee [11,36]. Apart from the above-mentioned limitations, one might criticize that we decided against performing a horizontal subdivision of the femoral articular cartilage. However, articular cartilage of the anterolateral femoral cartilage is rather thin compared to e.g. the patellar cartilage. Therefore we consensually decided to investigate the full cartilage layer. Moreover, we intended to generally analyse if the DLNS has a measurable influence on cartilage relaxation times and compare the findings to the adjacent femoral cartilage. Therefore we focused on the general influence of a DLNS rather than creating inaccurate measurements based on a limited amount of pixels. Finally, it might be criticized that only one Radiologist performed ROI placements as the main investigator. However, ROI placements were repeated in 10 randomly selected individuals (5 of each cohort). We consider the distinct intrarater reproducibility of ≥0.8 as an indicator of a very high conformity of our measurements. Apart from the above-mentioned limitations, the presented results of our investigation lead to the assumption that a positive DLNS in case of ACLR alludes to a more severe injury representing progressive cartilage degradation due to an extensive trauma mechanism. Further investigation has to be encouraged focussing on the clinical course of patients presenting with a DLNS after an ACLR including quantitative measurements as well as the assessment of clinical scores. Prospective longitudinal research might demonstrate if prolonged relaxation times in post-traumatic MRI can reliably detect severe cartilage degradation and therefore identify those patients, who might benefit from invasive cartilage repair in order to prevent early OA.
(2016) 162–168. [4] M.G. Liptak, K.R. Angel, Return to play and player performance after anterior cruciate ligament injury in elite Australian rules football players, Orthop. J. Sports Med. 5 (June (6)) (2017) 2325967117711885. [5] H. Roos, T. Adalberth, L. Dahlberg, L.S. Lohmander, Osteoarthritis of the knee after injury to the anterior cruciate ligament or meniscus: the influence of time and age, Osteoarthr. Cartil. 3 (December (4)) (1995) 261–267. [6] L.S. Lohmander, A. Stenberg, M. Englund, H. Roos, High prevalence of knee osteoarthritis, pain, and functional limitations in female soccer players twelve years after anterior cruciate ligament injury, Arthritis Rheum. 50 (October (10)) (2004) 3145–3152. [7] S. Bollen, Epidemiology of knee injuries: diagnosis and triage, Br. J. Sports Med. 34 (June (3)) (2000) 227–228. [8] A.R. Jones, D.B.L. Finlay, D.J. Learmonth, A deep lateral femoral notch as a sign of acutely torn anterior cruciate ligament, Injury 24 (October (9)) (1993) 601–602. [9] E. Herbst, C. Hoser, K. Tecklenburg, M. Filipovic, C. Dallapozza, M. Herbort, et al., The lateral femoral notch sign following ACL injury: frequency, morphology and relation to meniscal injury and sports activity, Knee Surg. Sports Traumatol. Arthrosc. 23 (August (8)) (2015) 2250–2258. [10] T. Baum, G.B. Joseph, D.C. Karampinos, P.M. Jungmann, T.M. Link, J.S. Bauer, Cartilage and meniscal T2 relaxation time as non-invasive biomarker for knee osteoarthritis and cartilage repair procedures, Osteoarthr. Cartil. 21 (October (10)) (2013) 1474–1484. [11] A.A. Theologis, B. Haughom, F. Liang, Y. Zhang, S. Majumdar, T.M. Link, et al., Comparison of T1rho relaxation times between ACL-reconstructed knees and contralateral uninjured knees, Knee Surg. Sports Traumatol. Arthrosc. 22 (January (2)) (2013) 298–307. [12] C. Behzadi, G.H. Welsch, A. Laqmani, F.O. Henes, M.G. Kaul, G. Schoen, et al., Comparison of T2* relaxation times of articular cartilage of the knee in elite professional football players and age-and BMI-matched amateur athletes, Eur. J. Radiol. 1 (January (86)) (2017) 105–111. [13] C. Behzadi, K.J. Maas, G. Welsch, M. Kaul, G. Schoen, A. Laqmani, et al., Quantitative T2* relaxation time analysis of articular cartilage of the tibiotalar joint in professional football players and healthy volunteers at 3T MRI, J. Magn. Reson. Imaging 47 (February (2)) (2018) 372–379. [14] M.D. Crema, F.W. Roemer, M.D. Marra, D. Burstein, G.E. Gold, F. Eckstein, et al., Articular cartilage in the knee: current MR imaging techniques and applications in clinical practice and research 1, RadioGraphics 31 (January (1)) (2011) 37–61. [15] R. Gheno, Y.C. Yoon, J.H. Wang, K. Kim, S.-Y. Baek, Changes in the T2relaxation value of the tibiofemoral articular cartilage about 6 months after anterior cruciate ligament reconstruction using the double-bundle technique, Br. J. Radiol. 89 (April (1060)) (2016) 20151002–20151010. [16] F. Su, J.F. Hilton, L. Nardo, S. Wu, F. Liang, T.M. Link, et al., Cartilage morphology and T1rho; And T2 quantification in ACL-reconstructed knees: a 2-year follow-up, Osteoarthr. Cartil. 21 (August (8)) (2013) 1058–1067. [17] F. Su, V. Pedoia, H.L. Teng, M. Kretzschmar, B.C. Lau, C.E. McCulloch, et al., The association between MR T1p and T2 of cartilage and patient-reported outcomes after ACL injury and reconstruction, Osteoarthr. Cartil. 24 (July (7)) (2016) 1180–1189. [18] H. Li, A. Hosseini, J.-S. Li, T.J. Gill, G. Li, Quantitative magnetic resonance imaging (MRI) morphological analysis of knee cartilage in healthy and anterior cruciate ligament-injured knees, Knee Surg. Sports Traumatol. Arthrosc. 20 (August (8)) (2012) 1496–1502. [19] J.H. Kellgren, J.S. Lawrence, Radiografic criteria for classification of osteoarthritis, Ann. Rheum. Dis. 16 (1957) 494–502. [20] R.B. Souza, B.T. Feeley, Z.A. Zarins, T.M. Link, X. Li, S. Majumdar, T1rho MRI relaxation in knee OA subjects with varying sizes of cartilage lesions, Knee 20 (March (2)) (2013) 113–119. [21] K.K. Hovis, H. Alizai, S.-C. Tham, R.B. Souza, M.C. Nevitt, C.E. McCulloch, et al., Non-traumatic anterior cruciate ligament abnormalities and their relationship to Osteoarthritis using morphological grading and cartilage T2 relaxation times: data from the Osteoarthritis Initiative (OAI), Skeletal Radiol. 41 (February (11) (2012) 1435–1443. [22] F.R. Noyes, C.L. Stabler, A system for grading articular cartilage lesions at arthroscopy, Am. J. Sports Med. 17 (July (4)) (1989) 505–513. [23] P.S. Rasmussen, Tibial condylar fractures. Impairment of knee joint stability as an indication for surgical treatment, J. Bone Joint Surg. Am. 55 (October (7)) (1973) 1331–1350. [24] Y. Tegner, J. Lysholm, Rating systems in the evaluation of knee ligament injuries, Clin. Orthop. Relat. Res. (198) (1985) 42–49. [25] S.D. Barber-Westin, F.R. Noyes, J.W. McCloskey, Rigorous statistical reliability, validity, and responsiveness testing of the cincinnati knee rating system in 350 subjects with uninjured, injured, or anterior cruciate ligament-reconstructed knees, Am. J. Sports Med. 27 (4) (1999) 402–416. [26] S. Shiraj, H.K. Kim, C. Anton, P.S. Horn, T. Laor, Spatial variation of T2 relaxation times of patellar cartilage and physeal patency: an in vivo study in children and young adults, Am. J. Roentgenol. 202 (March (3)) (2014) W292–7. [27] J.H. Ahn, S.H. Lee, S.H. Choi, T.K. Lim, Magnetic resonance imaging evaluation of anterior cruciate ligament reconstruction using quadrupled hamstring tendon autografts, Am. J. Sports Med. 38 (August (9)) (2010) 1768–1777. [28] H. Tao, Y. Hu, Y. Qiao, K. Ma, X. Yan, Y. Hua, et al., T2-Mapping evaluation of early cartilage alteration of talus for chronic lateral ankle instability with isolated anterior talofibular ligament tear or combined with calcaneofibular ligament tear, J. Magn. Reson. Imaging 47 (January (1)) (2018) 69–77. [29] A. Williams, C.S. Winalski, C.R. Chu, Early articular cartilage MRI T2 changes after anterior cruciate ligament reconstruction correlate with later changes in T2 and
Conflict of interest There are no known conflicts of interest. References [1] R.I. Bolbos, T.M. Link, C.B. Ma, S. Majumdar, X. Li, T1rho relaxation time of the meniscus and its relationship with T1rho of adjacent cartilage in knees with acute ACL injuries at 3 T, Osteoarthr. Cartil. 17 (January (1)) (2009) 12–18. [2] X. Li, D. Kuo, A. Theologis, J. Carballido-Gamio, C. Stehling, T.M. Link, et al., Cartilage in anterior cruciate ligament–reconstructed knees: MR imaging T1ρ and T2 - initial experience with 1-year follow-up, Radiology 258 (February (2)) (2011) 505–514. [3] P. Kannus, M. Jarvinen, R. Johnson, P. Renström, M. Pope, B. Beynnon, et al., Function of the quadriceps and hamstrings muscles in knees with chronic partial deficiency of the anterior cruciate ligament, Am. J. Sports Med. 20 (April (2))
90
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
osteoarthritis initiative, Skeletal Radiol. 16 (November) (2017) 1–14. [34] J.A. Vilensky, B.L. O’Connor, K.D. Brandt, E.A. Dunn, P.I. Rogers, C.A. DeLong, Serial kinematic analysis of the unstable knee after transection of the anterior cruciate ligament: temporal and angular changes in a canine model of osteoarthritis, J. Orthop. Res. 12 (March (2)) (1994) 229–237. [35] J.K. Seon, H.R. Gadikota, M. Kozanek, L.S. Oh, T.J. Gill, G. Li, The effect of anterior cruciate ligament reconstruction on kinematics of the knee with combined anterior cruciate ligament injury and subtotal medial meniscectomy: an in vitro robotic investigation, YJARS 1 (February (2)) (2009) 123–130. [36] C.W. Kim, A. Hosseini, L. Lin, Y. Wang, M. Torriani, T.J. Gill, A.J. Grodzinsky, et al., Quantitative analysis of T2 relaxation times of the patellofemoral jointcartilage 3 years after anterior cruciate ligament reconstruction, J. Orthop. Transl. 2 (July (12)) (2017) 85–92. [37] R Core Team, R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, Austria, 2017.
cartilage thickness, J. Orthop. Res. (July) (2016) 1–8. [30] R.B. Souza, S.J. Wu, L.J. Morse, K. Subburaj, C.R. Allen, B.T. Feeley, Cartilage MRI relaxation times after arthroscopic partial medial meniscectomy reveal localized degeneration, Knee Surg. Sports Traumatol. Arthrosc. 23 (May (1)) (2014) 188–197. [31] C.R. Chu, A.A. Williams, R.V. West, Y. Qian, F.H. Fu, B.H. Do, et al., Quantitative magnetic resonance imaging UTE-T2* mapping of cartilage and meniscus healing after anatomic anterior cruciate ligament reconstruction, Am. J. Sports Med. 42 (August (8)) (2014) 1847–1856. [32] H. Li, S. Chen, H. Tao, S. Chen, Quantitative MRI T2 relaxation time evaluation of knee cartilage: comparison of meniscus-intact and -injured knees after anterior cruciate ligament reconstruction, Am. J. Sports Med. 43 (April (4)) (2015) 865–872. [33] F.C. Hofmann, J. Neumann, U. Heilmeier, G.B. Joseph, M.C. Nevitt, C.E. McCulloch, et al., Conservatively treated knee injury is associated with knee cartilage matrix degeneration measured with MRI-based T2 relaxation times: data from the
91
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Research article
T2 relaxation times of the anterolateral femoral cartilage in patients after ACL-reconstruction with and without a deep lateral femoral notch sign
T
⁎
Cyrus Behzadia, , Goetz H. Welschb, Jan-Philipp Petersend, Bjoern P. Schoennagela, Peter Bannasa, Michael G. Kaula, Gerhard Schoenc, Josephine Berger-Grochd, Gerhard Adama, Marc Regiera a
Department of Diagnostic and Interventional Radiology and Nuclear Medicine, University Medical Center Hamburg-Eppendorf, Hamburg, 20246, Germany Department of Athletics and Sports Medicine, University Medical Center Hamburg-Eppendorf, Hamburg, 20246, Germany c Department of Medical Biometry and Epidemiology, University Medical Center Hamburg-Eppendorf, Hamburg, 20246, Germany d Department of Trauma, Hand and Reconstructive Surgery, University Medical Center Hamburg-Eppendorf, Hamburg, 20246, Germany b
A R T I C LE I N FO
A B S T R A C T
Keywords: Quantitative MRI T2 relaxation time Articular cartilage Deep lateral femoral notch sign Anterior cruciate ligament rupture
Purpose: To quantitatively assess T2 relaxation times of the anterolateral femoral cartilage following anterior cruciate ligament (ACL)-reconstruction with and without a positive deep lateral femoral notch sign (DLNS) at post-traumatic MRI. Materials and Methods: In 52 patients post-traumatic MRI as well as 12 months after ACL-rupture (ACLR) and surgical treatment were analysed. In 28 patients a positive DLNS was present at post-traumatic MRI. For quantitative analysis, T2 relaxation time measurements (7 TE: 10–70 ms) were performed at time of reevaluation. Three polygonal ROIs encompassing the full cartilage layer were placed in the anterolateral femoral cartilage. Clinical assessment included Lysholm-Tegner-Activity-Score, Rasmussen's clinical score and modified Cincinnati-Rating-System-Questionnaire. Description and differences were calculated as means and confidence intervals of means, controlled for the cluster effect of person, if appropriate. Results: In patients with a positive DLNS after ACLR, relaxation times in the notch region were significantly prolonged compared to patients without a positive DLNS (Δ 7.4 ms, CI: 5.6–9.2; p-value < 0.001) as well as to the adjacent anterior (Δ 5.7 ms, CI: 4.7–6.7; p-value < 0.001) and central femoral cartilage (Δ 6.6 ms, CI: 5.7–7.6; p-value < 0.001). Solely insignificant differences were noticed in the performed clinical scores comparing the two groups (p > 0.05). Conclusion: Significantly prolonged T2 relaxation times of the anterolateral femoral cartilage were found in patients with a positive DLNS following ACL-reconstruction compared to patients without a DLNS. Based on these results, it has to be assumed that a positive DLNS is associated with higher cartilage degradation.
1. Introduction Acute rupture of the anterior cruciate ligament (ACL) is one of the most feared injuries in sports overall. Considering the high incidence of Anterior-Cruciate-LigamentRuptures (ACLRs) of more than 200.000 injuries in the US alone [1,2], the immense relevance becomes evident. The impact on every athlete’s career is underlined by reviewing the therapeutic consequences after ACLRs. Usually, post-traumatic surgery and at least six months of rehabilitation are mandatory in order to return to regular athletic activity
[3]. However, it is reported that only 48% of elite Australian football players could return-to-play in less than one year [4]. Furthermore, studies could demonstrate that in the majority of patients after an ACLR signs of osteoarthritis (OA) are detectable within 10–15 years [5,6]. In post-traumatic diagnostic evaluation of ACLR, concomitant injuries are presumably to be diagnosed. In approximately half of all patients suffering from ACLR, an articular cartilage injury is noted [7]. In an extensive trauma mechanism, due to a dorsolateral movement of the femur and a fixed lower leg ("pivot shift manoeuvre"), a
⁎ Corresponding author at: Center for Radiology and Endoscopy, Department of Diagnostic and Interventional Radiology and Nuclearmedicine, University Medical Center Hamburg-Eppendorf, Martinistrasse 52, 20246, Hamburg, Germany. E-mail address: [email protected] (C. Behzadi).
https://doi.org/10.1016/j.ejrad.2018.07.007 Received 9 March 2018; Received in revised form 24 May 2018; Accepted 8 July 2018 0720-048X/ © 2018 Elsevier B.V. All rights reserved.
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
system (Ingenia, Philips, Best, the Netherlands). A 16-channel knee coil was used for signal reception. The investigated knee was placed in the centre of the coil and sandbags positioned on patient’s leg.
transchondral impression of the anterolateral femoral condyle versus the dorsolateral tibial plateau can result. This constellation can be detected at post-traumatic radiography (lateral view), known as the deep lateral notch (sulcus) sign (DLNS) [8]. It has been reported that in high-pivoting sports in about 33% of patients after acute ACLR, a positive DLNS deeper than 2 mm can be observed [9]. At MRI a DLNS is accompanied by subchondral bone marrow edema in the distal femur in about 50% of patients [9]. In the recent decade various techniques have drawn attention to quantitative MRI measurements of articular cartilage. Techniques such as T1rho or T2/T2* have demonstrated that early stages of OA might be diagnosed prior to irreversible stages [10–13]. T2 mapping is based on the detection of increased water observed in cartilage degradation due to biochemical changes in the extracellular matrix (e.g. collagen fraction). High sensitivity rates for the indirect assessment of early stages of OA for T2 mapping have been published in various studies [10,28]. In our investigation of post-operative femorotibial joints in patients after ACLR, we decided to include T2 mapping as the most validated technique for quantitative measurements. T2 mapping has been used in many clinical trials for it offers several advantages (e.g. no i.v. contrast media required, robust against susceptibility artifacts) while being highly sensitive to biochemical changes in articular cartilage composition [10,14,15]. Progressive cartilage degradation has been generally confirmed in recent publications after ACLR up to two years after surgery compared to healthy individuals [16–18]. However, to the best of our knowledge, no dedicated quantitative analysis of the DLNS has been performed yet. Therefore, the purpose of the presented study was to analyse and compare T2 relaxation times in patients suffering from ACLR with and without a DLNS. We intended to evaluate if the DLNS is accompanied with altered relaxation times indicating a more severe traumatic event and higher cartilage degradation. Therefore, all patients were re-evaluated including quantitative T2 measurements as well as several clinical kneescores 12 months post-trauma.
2.1.2. Morphological sequences Sagittally oriented 3D fat-saturated proton-density weighted (Pdw) Turbo Spin Echo (TSE) sequence was performed and reformations subsequently generated. The parameters were kept identical at both time points: Time to repetition (TR) 1300 ms; echo-time (TE) 3.7 ms; flip angle: 90°, slice thickness: 0.6 mm (in total: 330 slices), field of view (FoV): 185 × 185 mm, matrix: 320 × 320, scan time: 8:24 min. In addition, a coronal T1w TSE sequence was acquired (TR: 818 ms, TE: 10.8 ms, flip angle: 90°, slice thickness: 2.2 mm (in total: 60 slices), matrix: 672 × 672, FoV: 180 × 180 mm, scan time 1:50 min. For quantitative measurements, a 3D multi-echo T2w Turbo Spin Echo (TSE) sequence was used (TR: 192.8 ms, 7 echo times (10–70 ms), flip angle: 90°, slice thickness: 3 mm, matrix: 320 × 320, FoV of 160 × 160 mm, Sense factor: 4.4, scan time: 6:15 min). T2 mapping was performed by fitting a monoexponential function A*exp(–TE/T2) to the multiecho data using an in-house quantification plugin (qMapIt) extending ImageJ (National Institutes of Health, Bethesda, MD). The first echo was neglected because it does not contain any signal of a stimulated echo in contrast to the later ones. 2.1.3. Morphological image analysis All patients underwent biplane radiography (anterior-posterior and lateral view) of the injured knee in the emergency department at the day of trauma. These were evaluated in accordance to the KellgrenLawrence grading for signs of OA [19]. All MRIs between 2014 and 2015 were scanned for ACLRs by two radiologists with special interest in sports imaging (five and 12 years of experience). If both radiologists diagnosed ACLR, MRIs were re-evaluated for additional injuries. If the readers had divergent opinion concerning the status of the ACL or concomitant injuries, they reviewed the anonymized data sets and formed a consensual diagnosis [2,20,21]. Morphological analysis was performed by using a commercially available post processing workstation (Extended Brilliance Workspace, Version 2.0, Philips Healthcare, Best, the Netherlands). Concomitant injuries were noted as listed below: a) impression of the anterolateral femoral condyle, b) bone marrow edema, c) rupture of the collateral ligaments, d) cartilage lesions according to Noyes classification higher than grade 2 [22], e) rupture of the retinaculum patellae, f) injury of the Hoffa fat pad, g) fracture, h) meniscal injuries (lesions of the anterior or posterior horn as well as bucket handle tears, i) rupture of the patellar tendon. In patients with a DLNS sign, maximum length of impression in each patient was measured and the mean length for all impressions calculated (in mm). The depth of the impression (in mm) was measured perpendicular to the surface of the articular cartilage of the femoral condyle in sagittal orientation and categorized as follows: I: 0 - < 1 mm; II: 1 - < 2 mm, III: 2 - < 3 mm; IV: 3 - < 4 mm; V: ≥ 4 mm. Clinical evaluation at re-examination included three different Knee scoring systems (modified Lysholm-Tegner-Activity-Score, Rasmussen's clinical score and modified Cincinnati Rating System Questionnaire) [23–25]. Apart from the time-point at re-evaluation, all patients had to complete the scores retrospectively concerning their immediate posttraumatic status.
2. Materials & methods Approval from the local institutional review board was received prior to initiation of the study. Informed patient consent was obtained from all patients before image acquisition. 2.1. Study population and Inclusion/Exclusion criteria All patients in our database, who were diagnosed with ACLR (between 2014–2015) prior to the study were contacted and asked to voluntarily participate in our study. In total, 59 out of 124 patients took part in our investigation (47%). Four patients refused an MRI examination at re-evaluation and were consequently excluded. Three patients were excluded because their knees did not fit inside the 16channel knee coil. Therefore, the quantitative imaging protocol was not applicable. Therefore, 52 patients (18 female, 34 male; mean age at reevaluation: 32.5 years) were enrolled. Exclusion criteria were: 1.) Surgery prior to ACLR on the injured knee, 2.) Acute injury or history of major injury after ACLR (e.g. fracture) or 3.) Systemic diseases like e.g. rheumatoid arthritis. All scans at time of re-evaluation were conducted between August 2015 and April 2016.
2.2. Quantitative analysis
2.1.1. Image acquisition All patients underwent morphological and quantitative MRI imaging protocol at re-evaluation. Morphological sequences were acquired prior to T2 relaxation measurements. The order was kept constant in all patients. At both time points, all MRIs were performed using a 3 T MRI
Three cartilage regions of the lateral femoral condyle were analysed in all participants (the anterolateral as well as the bordering anterior and central cartilage, Fig. 1). In patients with a positive DLNS in posttraumatic MRI, maximum length of the impression (in mm) in the anterolateral cartilage was measured. Subsequently, the ROI in the 86
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
likelihood ratio tests. Measurement such as Lysholm-Tegner-Activity-Score, Rasmussen's clinical score and modified Cincinnati Rating System Questionnaire, which were measured on a person level, were described by means and confidence intervals of means. The standard level of significance used to identify statistically significant effects was set to 0.05. All statistical analyses were performed with the statistical software package R, version 3.4.1 [37]. 3. Results 3.1. Morphological post-traumatic injuries No positive signs of initial stages of OA were diagnosed reading biplane radiography (grade 0: n = 52). At posttraumatic MRI, a positive DLNS was detected in 28 patients (54%; impression < 1 mm: n = 9; 1 - < 2 mm: n = 11; 2 − < 3 mm: n = 3; 3 − < 4 mm: n = 3; ≥4 mm: n = 2). The mean length of the impression was calculated with 16 mm (+− 3 mm). 50 patients had a bone marrow edema in the tibial plateau or distal femur (96%). In the lateral tibial plateau three fractures were detected (6%). At follow-up, bone marrow edema had fully declined in all cases. Nine cartilage lesions (grade 3: n = 4; grade 4: n = 5) were detected in eight patients (15%). 29 meniscal lesions were diagnosed in 25 patients (48%): 16 ruptures of the posterior horn of the medial meniscus, 11 ruptures of the posterior horn of the lateral meniscus, one rupture of the anterior horn of the lateral meniscus and one bucket handle tear. One patellar tendon rupture as well as one rupture of the medial collateral ligament (2%) and two ruptures of the Hoffa fat pad (4%) were found. No fracture was detected reading post-traumatic MRI.
Fig. 1. Sagittal Pdw sequence including T2w-colour overlay in a patient with a positive DLNS. In each participant, three regions of the anterolateral femoral cartilage (A = Anterior; N = Notch Region, C = Central) were defined and subsequently analysed. As exemplarily demonstrated in this patient, elevated T2 relaxation times of the Notch Region (42.6 ± 9.1 ms) in comparison to the Anterior (35.1 ± 9.9 ms) and Central Regions (31.3 ± 5.7 ms) of the femoral cartilage could be detected.
3.2. Quantitative analysis of the anterolateral femur
anterolateral cartilage was adjusted to the initial length of the impression for each patient. For the adjacent anterior and central femoral cartilage, the identical length was chosen for ROI placements separately. In patients without a DLNS, ROI placements were performed after calculating the mean length of the impression in patients with a DLNS. Therefore, ROIs for the three investigated cartilage segments had the identical length as the impression of the anterolateral cartilage in patients with a DLNS. Due to the presence of rather thin cartilage at the anterolateral femoral condyle, we decided in all patients to place ROIs in three consecutive slices in each segment. Each ROI encovered the full depth of the cartilage layer. Therefore, 468 ROIs were placed and analysed using a dedicated ImageJ based software tool (qMapIt) by one radiologist (XX) in order to prevent interobserver measurement error as performed in previous studies and under supervision by a senior radiologist (12 years of experience in musculoskeletal imaging [13,26]. In order to ensure our quantitative analysis, ROI placement were repeated after a four-week interval in 10 randomly selected patients (5 of each cohort). Intraclass correlation coefficient (ICC) was calculated in order to compare the measurements (Rating: < 0.4: poor; 0.4–0.75: moderate; ≥0.75: good) [27,28]. Subchondral bone and joint effusion were excluded from ROI placements. The reader was blinded to patient information while performing the placements.
The ICC for intraobserver reliability in 10 randomly selected individuals was ≥0.8 in each cartilage segment. In patients with a DLNS, differences between the notch region and the anterior femoral cartilage (Δ 5.7 ms; CI: 4.7–6.7 ms; p-value: < 0.001) were highly significant. The same was found comparing the notch region to the central femoral cartilage (Δ 6.6 ms; CI: 5.7–7.6 ms; p-value: < 0.001). Mean relaxation time for the notch region was calculated with 42.0 ms (CI: 38.9–45.1 ms, Table 1). The adjacent anterior femoral cartilage revealed a mean relaxation time of 36.3 ms (CI: 33.8–38.9 ms) and the central femoral cartilage of 35.4 ms (CI: 32.8–38.0 ms). In patients without a DLNS, mean relaxation times for the anterior femoral cartilage (Δ 1.5 ms; CI: 0.5–2.5 ms; p-value: < 0.001) as well as for central femoral cartilage (Δ 0.1 ms; CI: −0.9 to 1.2 ms; p-value: 0.829) were prolonged if compared to the notch region. The comparison of the anterior to the central femoral cartilage revealed significantly higher results for the anterior femoral cartilage (Δ 1.4 ms; CI: 0.5–2.2 ms; p-value: 0.002). In patients without a DLNS, mean relaxation times for the notch region were calculated with 34.6 ms (CI: 33.3–35.9 ms). Anterior femoral cartilage revealed a mean relaxation time of 36.1 ms (CI: 35.0–37.2 ms) and the central femoral cartilage of 34.7 ms (CI: 33.6–35.9 ms). 3.3. Comparison between the two groups (positive vs. negative DLNS)
2.3. Statistical analysis
The calculated mean difference for the notch region revealed significantly elevated relaxation times in patients with a positive DLNS (Δ 7.4 ms; CI: 5.6–9.2 ms; p-value: < 0.001) compared to patients without a DLNS. In patients presenting a positive DLNS, mean relaxation time for anterior femoral cartilage were 0.2 ms longer than in patients without a DLNS (CI: −1.2 to 1.7 ms, p-value: 0.763). Differences between the two groups were insignificant for central
Due to violation of the independence assumption of single measurements, we estimated differences and confidence intervals of differences for relaxation time with random effects models. Fixed effects were "region" (anterior/notch/central) and "DLNS" (yes/no), controlled for the random effect "person". If appropriate, we conducted statistical models with interaction terms. Appropriateness was tested by 87
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
Table 1 Detailed illustration of T2 relaxation times (ms) in the anterolateral femoral cartilage in patients after ACLR with and without a positive deep lateral notch sign (DLNS).
femoral cartilage revealing slightly prolonged relaxation times in patients with a DLNS (Δ 0.7; CI: -−0.8 to 2.1 ms; p: 0.394).
model of the notch region was 37.6 ms (CI: 35.9–39.3 ms, Table 2). Differences between the five sub-categories in patients with a DLNS were insignificant (p = 0.693). For each category, a mean increase of 0.14 ms (CI: −0.54 to 0.82 ms) was detected.
3.4. Detailed evaluation of the notch region in patients with a DLNS
3.5. Evaluation of the performed clinical scores
Mean relaxation time for the intercept of the performed regression
Detailed evaluation is presented in Table 3 demonstrating insignificant differences for all scoring systems between patients with and without a DLNS at both time points (p > 0.05). Furthermore, we performed a statistical analysis in patients with a DLNS regarding any additional injuries. In patients with a DLNS, mean relaxation times between patients with and without concomitant injuries were statistically insignificant (p = 0.21). Concerning existing cartilage injuries higher than grade II at followup imaging of patients with a DLNS, differences of the mean relaxation times for this sub-cohort were as well statistically insignificant (p = 0.95).
Table 2 For the 28 patients with a positive DLNS, a special analysis revealed a constantly rising relaxation time (0.14 ms with each Millimetre; CI: -0.54 - 0.82) in comparison to the depth of the impression, although the correlation was insignificant (p = 0.693).
4. Discussion In the presented study, T2 relaxation times of patients suffering from ACLR were analysed. The main focus was to determine if a DLNS had any impact on T2 relaxation times and to compare their results to patients without a DLNS as well as to diverse clinical assessment scores. In our preliminary study analysing patients 12 months after ACLreconstruction, the presented results indicated significantly prolonged relaxation times in the anterolateral femoral cartilage in those patients with a DLNS (p < 0.001) compared to individuals without a DLNS. Interestingly, in patients with a DLNS, relaxation times of the notch region were significantly prolonged compared to the adjacent femoral 88
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
volunteers in all compartments [2]. Furthermore, they found a significant correlation between the superficial layers in T1rho and T2 relaxation times, indicating a similar value of these two diverse quantitative techniques. One might argue that their results were based on a small study collective consisting of seven men and five women only. However, in the limitation section the authors stated that the results could be considered as preliminary, encouraging further studies. In concordance to Li et al. we chose a time span of 12 months as well and re-evaluated 52 patients to compare our results. The detection of significantly prolonged relaxation times within the notch region in our patient collective further confirms their results of generally prolonged relaxation times after ACLR. Moreover, our findings strongly indicate higher biochemical cartilage alterations in patients in case of deep lateral notch sign. We believe that these quantitative measurements are capable of detecting patients at higher risk for manifestations of OA at an earlier time-point. This assumption is supported by further findings of Williams et al. [29]. They reported that initial changes in T2 relaxation times six months after ACLR were correlated to biochemical as well as morphological changes at a twoyear re-evaluation. If this impairment of biochemical cartilage composition accelerates onset of OA and subsequently leads to a reduced quality of life, has to be further evaluated in a longitudinal study design. A recently published study by Su et al. concluded that initially elevated relaxation times after ACLR might lead to a worse clinical outcome [17]. The authors analysed T1rho and T2 relaxation times after ACLR prior to surgery comparing the results to the contralateral, healthy knee of each individual. In addition, they completed at baseline as well at 6 and 12 months post surgery a clinical evaluation (KneeInjury and Osteoarthritis score (KOOS) and Marx activity level questionnaire). For instance, they observed at baseline significantly correlated side-to-side differences for T1rho and T2 relaxation times for the posterolateral tibia (p < 0.001). Interestingly, they only noticed a correlation between T2 and KOOS at baseline whereas no such association was observed at 6 and 12 months follow-up. The authors concluded that in contrast to T2 relaxation time measurement, T1rho might be a more accurate tool to longitudinally evaluate compositional changes in articular cartilage. However, their conclusions were exclusively based on the assessment of subjective questionnaires and not confirmed with quantitative measurements post-surgery. It would be of interest to further evaluate their findings in future quantitative analysis. As a further limitation, the authors mentioned that nowadays T1rho measurements are exclusively used in research settings and not as established as T2 relaxation measurements. We agree with the authors and therefore focused on T2 measurements enabling future multi-centre re-evaluations. Divergent results have been published in recent literature concerning the question if concomitant injuries in ACLR affect cartilage relaxation times. Gheno et al. investigated patients post-traumatic and six months after ACL-reconstruction and correlated T2 relaxation times to clinical scores [15]. They found prolonged relaxation times in the weight bearing and non-weight bearing femorotibial cartilage when compared to a control group. Due to their findings, one might conclude that articular cartilage is generally altered in patients after ACL-reconstruction. Furthermore, they postulated that the meniscal status does not interfere with T2 relaxation measurements after ACLR which is in contrast to previous studies [30,31]. The results of our analysis seem to support the findings by Gheno et al.. In our patient collective non-significant differences were detected when comparing patients with isolated ACLR to those presenting with concomitant injuries (p > 0.05). However, it needs to be taken into account that our investigation focused on the anterolateral femoral cartilage alone at a singular
Table 3 For the three clinical scores (LGS = modified Lysholm-Tegner-Activity-Score; RS = Rasmussen's clinical score, C = modified Cincinnati Rating System Questionnaire) no significant differences in the analysis of the immediate posttraumatic situation and at the time of re-evaluation were diagnosed (p > 0.05).
cartilage (p < 0.001) affirming the hypothesis that the DLNS represents higher cartilage degradation. However, the observed differences in our quantitative cartilage analysis could not be found likewise in the patients’ clinical symptoms. This lack of proportionality might be attributed to subtle cartilage injuries prior to clinical manifestations or the dominance of an ACLR as the major injury masking other symptoms. To the best of our knowledge, the presented study is the first to perform a specific quantitative analysis of the DLNS in patients after ACLRs. Recent publications focused on consequences of ACLR on articular cartilage of the knee have alluded to similar results revealing prolonged relaxation times after surgical treatment [16,18,29]. Therefore, we constrained our investigation on the anterolateral femoral cartilage including the notch region and the surrounding cartilage in patients with a positive DLNS and compared these results to patients without a positive DLNS. Li et al. investigated T1rho and T2 relaxation times immediately after an ACLR and at a 1-year follow up investigating 12 patients. They could successfully demonstrate that cartilage relaxation times in patients after ACLR were significantly prolonged compared to healthy 89
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
evaluation 12 months after ACLR. These cartilage areas might not be as affected as the central weight bearing cartilage areas after e.g. meniscectomy in ACLR. As a primary limitation, one might argue that we have only compared T2 relaxation times at a 1-year follow-up after surgery and did not perform a comparison to a healthy control collective. However, various investigators have recently described generally elevated relaxation times in post-traumatic MRI as well as in follow-up examinations up to 2-years after ACLR when compared to healthy individuals [16,30,32]. Even in non-operatively treated patients after various knee injuries generally elevated T2 relaxation times could be demonstrated compared to healthy individuals indicating cartilage abnormalities [33]. Therefore, our aim was to distinctively focus on the intra- and interindividual analysis of T2 relaxation times of the anterolateral and adjacent femoral cartilage in patients after ACLR with and without a DLNS. Furthermore, one might argue that the analysis of the contralateral knee could have led to an internal control ruling out physiological interindividual alterations. However, quantitative measurements of the healthy knee in patients following ACLR might have been confounded due to adapted post-traumatic kinematics [34,35]. These clinical results have also been lately confirmed by studies demonstrating generally elevated relaxation times in patients after ACLR compared to the uninjured contralateral knee [11,36]. Apart from the above-mentioned limitations, one might criticize that we decided against performing a horizontal subdivision of the femoral articular cartilage. However, articular cartilage of the anterolateral femoral cartilage is rather thin compared to e.g. the patellar cartilage. Therefore we consensually decided to investigate the full cartilage layer. Moreover, we intended to generally analyse if the DLNS has a measurable influence on cartilage relaxation times and compare the findings to the adjacent femoral cartilage. Therefore we focused on the general influence of a DLNS rather than creating inaccurate measurements based on a limited amount of pixels. Finally, it might be criticized that only one Radiologist performed ROI placements as the main investigator. However, ROI placements were repeated in 10 randomly selected individuals (5 of each cohort). We consider the distinct intrarater reproducibility of ≥0.8 as an indicator of a very high conformity of our measurements. Apart from the above-mentioned limitations, the presented results of our investigation lead to the assumption that a positive DLNS in case of ACLR alludes to a more severe injury representing progressive cartilage degradation due to an extensive trauma mechanism. Further investigation has to be encouraged focussing on the clinical course of patients presenting with a DLNS after an ACLR including quantitative measurements as well as the assessment of clinical scores. Prospective longitudinal research might demonstrate if prolonged relaxation times in post-traumatic MRI can reliably detect severe cartilage degradation and therefore identify those patients, who might benefit from invasive cartilage repair in order to prevent early OA.
(2016) 162–168. [4] M.G. Liptak, K.R. Angel, Return to play and player performance after anterior cruciate ligament injury in elite Australian rules football players, Orthop. J. Sports Med. 5 (June (6)) (2017) 2325967117711885. [5] H. Roos, T. Adalberth, L. Dahlberg, L.S. Lohmander, Osteoarthritis of the knee after injury to the anterior cruciate ligament or meniscus: the influence of time and age, Osteoarthr. Cartil. 3 (December (4)) (1995) 261–267. [6] L.S. Lohmander, A. Stenberg, M. Englund, H. Roos, High prevalence of knee osteoarthritis, pain, and functional limitations in female soccer players twelve years after anterior cruciate ligament injury, Arthritis Rheum. 50 (October (10)) (2004) 3145–3152. [7] S. Bollen, Epidemiology of knee injuries: diagnosis and triage, Br. J. Sports Med. 34 (June (3)) (2000) 227–228. [8] A.R. Jones, D.B.L. Finlay, D.J. Learmonth, A deep lateral femoral notch as a sign of acutely torn anterior cruciate ligament, Injury 24 (October (9)) (1993) 601–602. [9] E. Herbst, C. Hoser, K. Tecklenburg, M. Filipovic, C. Dallapozza, M. Herbort, et al., The lateral femoral notch sign following ACL injury: frequency, morphology and relation to meniscal injury and sports activity, Knee Surg. Sports Traumatol. Arthrosc. 23 (August (8)) (2015) 2250–2258. [10] T. Baum, G.B. Joseph, D.C. Karampinos, P.M. Jungmann, T.M. Link, J.S. Bauer, Cartilage and meniscal T2 relaxation time as non-invasive biomarker for knee osteoarthritis and cartilage repair procedures, Osteoarthr. Cartil. 21 (October (10)) (2013) 1474–1484. [11] A.A. Theologis, B. Haughom, F. Liang, Y. Zhang, S. Majumdar, T.M. Link, et al., Comparison of T1rho relaxation times between ACL-reconstructed knees and contralateral uninjured knees, Knee Surg. Sports Traumatol. Arthrosc. 22 (January (2)) (2013) 298–307. [12] C. Behzadi, G.H. Welsch, A. Laqmani, F.O. Henes, M.G. Kaul, G. Schoen, et al., Comparison of T2* relaxation times of articular cartilage of the knee in elite professional football players and age-and BMI-matched amateur athletes, Eur. J. Radiol. 1 (January (86)) (2017) 105–111. [13] C. Behzadi, K.J. Maas, G. Welsch, M. Kaul, G. Schoen, A. Laqmani, et al., Quantitative T2* relaxation time analysis of articular cartilage of the tibiotalar joint in professional football players and healthy volunteers at 3T MRI, J. Magn. Reson. Imaging 47 (February (2)) (2018) 372–379. [14] M.D. Crema, F.W. Roemer, M.D. Marra, D. Burstein, G.E. Gold, F. Eckstein, et al., Articular cartilage in the knee: current MR imaging techniques and applications in clinical practice and research 1, RadioGraphics 31 (January (1)) (2011) 37–61. [15] R. Gheno, Y.C. Yoon, J.H. Wang, K. Kim, S.-Y. Baek, Changes in the T2relaxation value of the tibiofemoral articular cartilage about 6 months after anterior cruciate ligament reconstruction using the double-bundle technique, Br. J. Radiol. 89 (April (1060)) (2016) 20151002–20151010. [16] F. Su, J.F. Hilton, L. Nardo, S. Wu, F. Liang, T.M. Link, et al., Cartilage morphology and T1rho; And T2 quantification in ACL-reconstructed knees: a 2-year follow-up, Osteoarthr. Cartil. 21 (August (8)) (2013) 1058–1067. [17] F. Su, V. Pedoia, H.L. Teng, M. Kretzschmar, B.C. Lau, C.E. McCulloch, et al., The association between MR T1p and T2 of cartilage and patient-reported outcomes after ACL injury and reconstruction, Osteoarthr. Cartil. 24 (July (7)) (2016) 1180–1189. [18] H. Li, A. Hosseini, J.-S. Li, T.J. Gill, G. Li, Quantitative magnetic resonance imaging (MRI) morphological analysis of knee cartilage in healthy and anterior cruciate ligament-injured knees, Knee Surg. Sports Traumatol. Arthrosc. 20 (August (8)) (2012) 1496–1502. [19] J.H. Kellgren, J.S. Lawrence, Radiografic criteria for classification of osteoarthritis, Ann. Rheum. Dis. 16 (1957) 494–502. [20] R.B. Souza, B.T. Feeley, Z.A. Zarins, T.M. Link, X. Li, S. Majumdar, T1rho MRI relaxation in knee OA subjects with varying sizes of cartilage lesions, Knee 20 (March (2)) (2013) 113–119. [21] K.K. Hovis, H. Alizai, S.-C. Tham, R.B. Souza, M.C. Nevitt, C.E. McCulloch, et al., Non-traumatic anterior cruciate ligament abnormalities and their relationship to Osteoarthritis using morphological grading and cartilage T2 relaxation times: data from the Osteoarthritis Initiative (OAI), Skeletal Radiol. 41 (February (11) (2012) 1435–1443. [22] F.R. Noyes, C.L. Stabler, A system for grading articular cartilage lesions at arthroscopy, Am. J. Sports Med. 17 (July (4)) (1989) 505–513. [23] P.S. Rasmussen, Tibial condylar fractures. Impairment of knee joint stability as an indication for surgical treatment, J. Bone Joint Surg. Am. 55 (October (7)) (1973) 1331–1350. [24] Y. Tegner, J. Lysholm, Rating systems in the evaluation of knee ligament injuries, Clin. Orthop. Relat. Res. (198) (1985) 42–49. [25] S.D. Barber-Westin, F.R. Noyes, J.W. McCloskey, Rigorous statistical reliability, validity, and responsiveness testing of the cincinnati knee rating system in 350 subjects with uninjured, injured, or anterior cruciate ligament-reconstructed knees, Am. J. Sports Med. 27 (4) (1999) 402–416. [26] S. Shiraj, H.K. Kim, C. Anton, P.S. Horn, T. Laor, Spatial variation of T2 relaxation times of patellar cartilage and physeal patency: an in vivo study in children and young adults, Am. J. Roentgenol. 202 (March (3)) (2014) W292–7. [27] J.H. Ahn, S.H. Lee, S.H. Choi, T.K. Lim, Magnetic resonance imaging evaluation of anterior cruciate ligament reconstruction using quadrupled hamstring tendon autografts, Am. J. Sports Med. 38 (August (9)) (2010) 1768–1777. [28] H. Tao, Y. Hu, Y. Qiao, K. Ma, X. Yan, Y. Hua, et al., T2-Mapping evaluation of early cartilage alteration of talus for chronic lateral ankle instability with isolated anterior talofibular ligament tear or combined with calcaneofibular ligament tear, J. Magn. Reson. Imaging 47 (January (1)) (2018) 69–77. [29] A. Williams, C.S. Winalski, C.R. Chu, Early articular cartilage MRI T2 changes after anterior cruciate ligament reconstruction correlate with later changes in T2 and
Conflict of interest There are no known conflicts of interest. References [1] R.I. Bolbos, T.M. Link, C.B. Ma, S. Majumdar, X. Li, T1rho relaxation time of the meniscus and its relationship with T1rho of adjacent cartilage in knees with acute ACL injuries at 3 T, Osteoarthr. Cartil. 17 (January (1)) (2009) 12–18. [2] X. Li, D. Kuo, A. Theologis, J. Carballido-Gamio, C. Stehling, T.M. Link, et al., Cartilage in anterior cruciate ligament–reconstructed knees: MR imaging T1ρ and T2 - initial experience with 1-year follow-up, Radiology 258 (February (2)) (2011) 505–514. [3] P. Kannus, M. Jarvinen, R. Johnson, P. Renström, M. Pope, B. Beynnon, et al., Function of the quadriceps and hamstrings muscles in knees with chronic partial deficiency of the anterior cruciate ligament, Am. J. Sports Med. 20 (April (2))
90
European Journal of Radiology 106 (2018) 85–91
C. Behzadi et al.
osteoarthritis initiative, Skeletal Radiol. 16 (November) (2017) 1–14. [34] J.A. Vilensky, B.L. O’Connor, K.D. Brandt, E.A. Dunn, P.I. Rogers, C.A. DeLong, Serial kinematic analysis of the unstable knee after transection of the anterior cruciate ligament: temporal and angular changes in a canine model of osteoarthritis, J. Orthop. Res. 12 (March (2)) (1994) 229–237. [35] J.K. Seon, H.R. Gadikota, M. Kozanek, L.S. Oh, T.J. Gill, G. Li, The effect of anterior cruciate ligament reconstruction on kinematics of the knee with combined anterior cruciate ligament injury and subtotal medial meniscectomy: an in vitro robotic investigation, YJARS 1 (February (2)) (2009) 123–130. [36] C.W. Kim, A. Hosseini, L. Lin, Y. Wang, M. Torriani, T.J. Gill, A.J. Grodzinsky, et al., Quantitative analysis of T2 relaxation times of the patellofemoral jointcartilage 3 years after anterior cruciate ligament reconstruction, J. Orthop. Transl. 2 (July (12)) (2017) 85–92. [37] R Core Team, R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, Austria, 2017.
cartilage thickness, J. Orthop. Res. (July) (2016) 1–8. [30] R.B. Souza, S.J. Wu, L.J. Morse, K. Subburaj, C.R. Allen, B.T. Feeley, Cartilage MRI relaxation times after arthroscopic partial medial meniscectomy reveal localized degeneration, Knee Surg. Sports Traumatol. Arthrosc. 23 (May (1)) (2014) 188–197. [31] C.R. Chu, A.A. Williams, R.V. West, Y. Qian, F.H. Fu, B.H. Do, et al., Quantitative magnetic resonance imaging UTE-T2* mapping of cartilage and meniscus healing after anatomic anterior cruciate ligament reconstruction, Am. J. Sports Med. 42 (August (8)) (2014) 1847–1856. [32] H. Li, S. Chen, H. Tao, S. Chen, Quantitative MRI T2 relaxation time evaluation of knee cartilage: comparison of meniscus-intact and -injured knees after anterior cruciate ligament reconstruction, Am. J. Sports Med. 43 (April (4)) (2015) 865–872. [33] F.C. Hofmann, J. Neumann, U. Heilmeier, G.B. Joseph, M.C. Nevitt, C.E. McCulloch, et al., Conservatively treated knee injury is associated with knee cartilage matrix degeneration measured with MRI-based T2 relaxation times: data from the
91