- Email: [email protected]
MRI of the knee following prosthetic anterior cruciate ligament reconstruction
Clinical Radiology (1994) 49, 89 99
MRI of the Knee Following Prosthetic Anterior Cruciate Ligament Reconstruction V. N. CASSAR-PULLICINO, I. W. McCALL and A. E. STROVER*
Department of Diagnostic Imaging and *Institute of Orthopaedics, The Robert Jones and Agnes Hunt Orthopaedic Hospital, Oswestry, Shropshire The MRI appearances of the knee in 30 patients with residual symptoms following active biocomposite (ABC) reconstruction of the anterior cruciate ligament were prospectively correlated with the clinical and arthroscopic findings. The variable MR features of the neo-ligament depend on the integrity of the prosthesis, the degree and extent of tissue ingrowth, and the time delay between surgical placement and MRI. The ABC ligament gives a uniform low signal when freshly placed, which is slowly lost by the incorporation of tissue ingrowth and therefore the loss of MR visualization of the ABC ligament is not synonymous with failure. Analysis of the imaging features suggests that a neo-ligament emerging through the tibial tunnel either at or posterior to the mid point of the AP diameter of the tibia is likely to be intact (P = 0.008) compared with one placed anteriorly. The presence of an effusion (64 %) indicates that symptoms are more likely to be due to other internal mechanical derangement rather than failure of the neoligament. Meniscal tears were seen in 67% while osteo-chondral defects were noted in 30% of all patients. Excellent correlation with arthroscopic findings was established confirming intact neoligaments in 16 replacements. MRI in the post-operative period detects rectifiable problems before the development of irreversible mechanical damage and re-rupture of the neo-ligament, but a thorough MRI technique needs to be utilized in examining the entire knee. Cassar-
Pullicino, V.N., McCall, I.W. & Strover, A.E. (1994). Clinical Radiology 49, 89 99. MRI of the Knee Following Prosthetic Anterior Cruciate Ligament Reconstruction
Accepted for Publication 19 July 1993
The anterior cruciate ligament (ACL) is an important part of the stable knee. In the active sportsman, incompetence of the ACL gives symptoms of recurrent and usually painful giving way with repetitive haemarthoses often leading to damage of other intra-articular structures [ 1,2]. The natural history of untreated cruciate ligament tears over a 10 year period indicates that although the majority of patients returned to active sports, few are asymptomatic, 67% showing early radiographic evidence of degenerative changes, while 6% showed frank osteoarthritis [3,6]. Eighty-five per cent required meniscetomy during the 10 year follow-up period [4]. Orthopaedic opinion is increasingly in agreement that the majority of cases of ruptured anterior cruciate ligament in the sportsman require early diagnosis and surgical repair or reconstruction with reinforcement [5]. Reconstruction has also been advocated for the old or chronic cruciate deficient knee [3,4,6]. Methods of ligament reconstruction include autogenous grafts, allografts and synthetic ligaments. Allograft and synthetic ligament reconstructions are particularly attractive by virtue of eliminating donor site problems [7], and the move has been towards arthroscopic surgical introduction, which has dramatically reduced the morbidity associated with this kind of surgery. The synthetic ABC (active bio-composite) ligament is composed of bundles of carbon and polyester fibres combined together in an open partially braided arrangement (Fig. 1) which allows the ingrowth of fibrovascular tissue between the fibre bundles with the development of a 'neoligament' [8]. In the short term the artificial scaffold Correspondence to: Dr V. N. Cassar-Pullicino, Department of Diagnostic Imaging, The Robert Jones and Agnes Hunt Orthopaedic Hospital, Oswestry, Shropshire SY11 7AG.
(a)
(b) Fig. ! (a, b) The active bio-composite ligament. The button and loop at the ends of the two braided components depict the tibial and femoral attachments respectively. The connecting filaments are composed of intertwined carbon and polyester fibres. The construction is designed to promote fibrous ingrowth with neo-ligament formation.
90
CLINICAL RADIOLOGY
(a)
(c)
O)
(d)
Fig. 2 - An intact active bio-composite neo-ligament at 9 months is demonstrated throughout its full length in the relevant projections. (a) The sagittal view shows the ligament exiting the tibia in the correct position just over the 50% ratio. The intercondylar line lies anterior to the ligament with a satisfactory gap between the lower tip of the intercondylar line and the tibia. (b, c) The coronal view shows the ligament in the intercondylar notch curving around the lateral condyle, which is inserted into the lateral cortex of the femur by a carbon pin (d).
91
MR[ IN ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION
(a)
(b)
Fig. 3 - The coronal views show (a) an intact rico-ligament at 12 months, with ingrowth of intermediate signal fibrous tissues which has partly obliterated the low signal of the prosthesis. (b) An intact neo-ligament at 2 years post-operation in which the prosthesis has been completely obscured by fibrous ingrowth. The tibial tract is normal in width and position.
should be strong enough at the time of implantation to allow early functional rehabilitation of the knee, eventually maturing to become a permanent and durable reconstruction of the anterior cruciate ligament. Due to the fibrogenic properties of carbon, prosthetic degradation becomes an asset, rather than a deficiency [9]. Magnetic resonance scanning is now established as an accurate and non-invasive method of evaluating the integrity of the anterior cruciate ligament [10]. The entire length of the ligament can be visualized on a sagittal view, which is slightly oblique in the coronal plane. Magnetic resonance has also been used to evaluate autologous grafts with considerable success [i2,13] and should therefore provide a valuable method to evaluate postoperative problems in scaffold types of synthetic ligaments. This study was undertaken to assess the post-operative status of ACL reconstructions using the ABC ligament in patients who continued to have symptoms of pain, swelling and instability, and to correlate the M R appearances with clinical and arthroscopic findings. The postoperative integrity of the synthetic ligament and internal structures of the knee were studied by MRI to investigate ligament failure. MATERIALS AND METHODS
The prospective study involved 30 patients who had
POSTERIOR =~ 75 -o ~._ 65
+ +
"6 ~ 55 50%
~ 45 c = -~ '~ 35 ,= "6 g ~ 25 _o ANTERIOB
q + +
I Failed
P = 0.008
I Intact
L i g a m e n t status
Fig. 4 'Box and whisker' plot display of one-tailed 't'-test showing the effect of tibial tunnel location on eventual outcome of ligament status. The control box covers the middle 50~ of data values, between the upper and lower quartiles. The 'whiskers' extend to 1.5 times the interquartile range and the control line is the median. Extreme values are plotted as separate plots.
92
CLINICAL RADIOLOGY
(a)
(b) Fig. 5 - (a) The sagittal view of a normal knee with an intact natural anterior cruciate ligament demonstrating the relationship of the intercondylar line to the AP diameter o f the tibia (arrow). (b) An intact ligament shown well posterior to the lower tip of the intercondylar line. (c) A failed ligament shown emerging from the tibia anterior to the intercondylar line with a narrow gap between the lower tip and the tibia.
(c)
returned to the 'problem knee clinic' at the Robert Jones and Agnes Hunt Orthopaedic Hospital complaining of symptoms of pain, swelling or instability following reconstruction of the anterior cruciate ligament using the active biological composite (ABC) ligament. It was ascertained at the outset that the carbon and polyester fabric of the ABC ligament has a low signal on MRI by undertaking an M R study of the ligament with a phantom. MRI of the whole knee was performed utilizing a 0.5 T G E Max scanner, with a dedicated knee surface coil. The knee was positioned in full extension as this was thought to provide a better assessment of the whole length of the neo-ligament through both the tibial tunnel and within the joint. The entire knee was scanned in the sagittal plane, using Tl-weighted spin-echo (TR 800/TE 30), and T2-weighted Gradient-echo (TR 300/TE 20/20 ~ sequences, followed by coronal S.E. Tl-weighted scans and axial S.E. Tl-weighted scans of the patello-femoral articulation to include the neo-ligament, the intercondylar notch and its contents. The optimum coronal view was used to measure accurately the inclination of the neoligament to the sagittal plane. A Tl-weighted S.E. sequence (TR 800 TE 30) with thin (3 mm) continuous
MRI IN ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION
(a)
93
(b)
Fig. 6 - (a, b) The braided portion of the neo-ligament lies in the tibial canal. The tibial exit is anterior to the mid point of the AP tibial diameter and the process of impingement has resulted in enlargement of tibial orifice prior to ligament failure. The stabilizing button has been drawn into the tibia due to the excess tension of the ligament within this joint.
sections was directed in the oblique sagittal plane along the axis of the neo-ligament. F r o m this sequence, the optimum oblique sagittal image was used as a further scout view to measure the inclination of the neo-ligament to the coronal plane. Further oblique 3 m m contiguous coronal images were obtained through the axis of the neoligament. The total examination time was 50 min. Longitudinal sectional images were thus obtained of the neo-ligament from its point of entry into the tibial drill hole through the intercondylar notch of the knee to its point of exit via the 'over-the-top' route (Fig. 2). The first six patients were also imaged after intravenous Gadolinium enhancement (0.1 mmol/kg body weight) using T1 S.E. sequences. Preoperative and post-operative clinical assessment, anterior cruciate ligament reconstruction and arthroscopic assessment were undertaken by the same orthopaedic surgeon (AES). The M R images were independently assessed by two experienced radiologists (IWM and VCP). The radiologists were blind to the clinical findings, operative details and post-operative status of the patients. At final follow-up the knees were classified as either stable or unstable. A stable knee had to demonstrate the following findings: extension to at least 0 ~ both Lachman and drawer tests demonstrating no more than grade 1 laxity with a solid end point and an absent pivot shift. Tibial translation was measured using the West-
minster cruciometer [11] and stability in these cases was defined as an increased tibial translation of not more than 2.5 m m of difference when compared with the intact knee. All patients had arthroscopic 'second look' assessments 3 months following the initial operation, confirming an intact prosthetic ligament. Ninety per cent of patients had at least one other operative arthroscopy at various intervals during the next two years. Three patients (10%) refused a further follow-up arthroscopy, and were evaluated clinically. The interval between either arthroscopy or clinical evaluation and the M R I examination varied from 3 to 6 months. Four patients had more than one M R examination. A one-tailed 't'-test was performed to compare the effect of the location of the tibial tunnel expressed as a percentage of the tibial sagittal distance, with ligament status in failed and intact reconstructions. RESULTS The study group consisted of 30 patients with a mean age of 29 years (range 20-42 years). There were 26 males and four females. Surgery was performed for chronic knee instability in 25 patients and for an acute A C L rupture in five. The neo-ligament was seen as a thick, distinct structure with a uniformly low signal in 10 patients (Fig. 2). All
94
CLINICAL RADIOLOGY
(a)
(b)
Fig, 7 - (a) Enlargement of the tibial orifice is again shown on Tl-weighted image. (b) The T2-weighted image demonstrates high signal within oedematous friable tissue with absence of the ligament in the joint.
these patients showed clinical and arthroscopic evidence of successful reconstruction of the A C L with stable knees. Eight of these patients were imaged in the first postoperative year and two in the second. A few residual low signal strands were seen in two patients (Fig. 3), both of whom were imaged in the first post-operative year. The ligament was not visible in 18 patients (Fig. 3), five of w h o m were imaged in the first post-operative year and 13 in subsequent years. The appearance of the ABC ligament was dependent on the post-operative interval, as well as its structural integrity. In the first post-operative year 15 M R studies were performed with the ligament being clearly seen in eight cases, with two showing residual strands. Failure to visualize the ligament as a low signal structure was present in five patients, four of whom had confirmed ligament failure (Fig. 5) and one had an intact ligament. Soft tissue incorporation into the ligament in the intercondylar region was c o m m o n after 1 year, with only two exceptions. A soft tissue mass of an intermediate signal spanning the intercondylar path of the ligament was seen in 21 patients, of w h o m three had a discernible ligament, two had a few low signal strands, and in l 6 the ligament was not visualized. The obliteration of the prosthetic ligament by soft tissue incorporation did not in itself indicate ligament failure especially after the first post-operative year and
therefore indirect evidence was required to diagnose failure (Figs 3, 8). The position of the intersection of the intercondylar line (Blumensaats) with the tibial tunnel was found to be directly related to the failure rate of the ABC ligaments (Fig. 4). The location of the orifice anterior to the midpoint of the sagittal diameter of the tibia (Figs 4, 5) was associated with a high incidence of failure (P = 0.008) using a multiple box and whisker plot. Non-parametric testing using the Mann-Whitney (U-test) based on the 'ranked' data also proved significant ( P > 0.01). Posterior displacement of the femoral condyles was seen in five cases and was associated with failure in all of them. The posterior cruciate ligament was arcuate in three of these cases. The appearance of the tibial tunnel and its contained ligament also varied. Initially, the tunnel was created by a drill hole of 6 m m diameter but expansion of the drill hole, which was maximal subchondrally at the joint, giving a trumpet appearance, was seen in 21 (70%) of patients (Figs 6, 7). The expansion gave an intermediate signal on Tl-weighted images but was associated with a high T2 signal in some cases. The diameter of the tibial tunnel was increased in 80% of all cases and was seen in cases of -failure and also in patients with an intact ligament. All 10 patients with expansion of the tibial tunnel beyond 12 m m diameter had arthroscopically proven failed ligaments.
95
MRI IN ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION
(b)
(e)
(a)
Fig. 8 (a) The intact ligament at t 5 m o n t h s is shown well posterior to the intercondylar line. (b, c, d) The axial views show the low signal fibres in the tibia1 tunnel, exit and intercondylar notch. The precise point of exit of the ligament and any impingement on the inner condylar wall are best assessed with the axial view.
96
CLINICAL RADIOLOGY
Ca)
Fig. I O - The ligamentum patellae is shown to be crenated in knee extension.
(b) Fig. 9 - Arthroscopic appearances (a) of an intact ligament covered by well organized fibrous tissue. (b) A failed ligament showing extensive disorganized granulation tissue containing carbon debris. Both views are taken just above the point of emergence from the tibia.
M R I provided valuable information with regard to other sources of pain in these symptomatic patients. Meniscal tears were seen in 20 patients (67%): nine involving the medial meniscus, one the lateral meniscus, and in 10 both menisci were damaged. The ABC ligament was intact in 14 o f these cases with meniscal tears and torn in the remaining six (Fig. 11). Osteochondral defects, primarily in the medial compartment (Fig. l 1), with or without loss of joint space and osteophytes, were seen in nine (30%) of all patients and in 56% of patients with medial meniscal tears. These were confirmed arthroscopically. An intra-articular effusion was present in 18 patients (64%), 14 of whom had an intact ABC ligament while in four the ligament replacement had failed. Overall, 16 of the ABC ligament replacements were shown to he intact by M R I and arthroscopy, with a clinically stable knee. Fourteen of these 16 cases also had meniscal damage (87%). Eleven knees were shown to have failed ligaments on M R I , which was confirmed by arthroscopy (Fig. 9) and clinically the knees were unstable. Six of these cases also had associated meniscal tears. In three cases the M R I indicated that the ligament had
failed but clinically the knees appeared stable. Failure of the neo-ligament was not confirmed by arthroscopy in these cases, as this was not performed. In six cases, gadolinium D T P A was injected after the initial imaging sequences. In no case was enhancement demonstrated and no improvement in the imaging information was achieved. The status of the ligament within the intercondylar notch was also assessed in the axial plane. Lateral wall encroachment or narrowing of the transverse diameter of the notch was seen in four of the 11 failures. None of the intact ABC ligaments showed M R evidence of lateral wall encroachment (Fig. 8). Ten patients complained of anterior knee pain postoperatively. O f these eight had M R features related to the patellar tendon. The patellar tendon was significantly thickened in four cases, crenated due to laxity in three (Fig. 10) and showed evidence of tendinitis with inferior bursitis in one patient. In one patient the button at the tibial fixation site had been pulled into the tibial cortex causing pain which settled with time in the presence of a stable knee.
DISCUSSION M R imaging been limited to central third of graft and M R I
after A C L reconstruction has previously the evaluation of autologous grafts. The the patellar tendon was the source of the showed that the ligament was intact in
MRI IN ANTERIOR C R U C I A T E LIGAMENT RECONSTRUCTION
Fig. 11 - D e g e n e r a t i v e articular features are s h o w n on the c o r o n a l view w i t h an effusion in the joint, a n absent m e d i a l meniscus a n d a large s u b c h o n d r a l cyst. Artefact from the tibial fixation p o i n t is seen.
50 patients, which correlated in 92% with clinical examination and in 100% of patients at a second look arthroscopy [12]. Moeser et al. [13] described post-operative MR scans in 37 patients with ACL reconstruction using fascia lata which showed that only six ligaments were well defined, 10 ill-defined and 11 indiscernible, despite clinical stability. They concluded that the normal neo-ligament unlike the active ACL has a variable appearance, including non-visualization on M R and that criteria used in evaluating the native ligament will be inadequate to assess the repair. Our findings support this concept, and it is clear from our study that the M R appearance of the neoligament is a function of the type of substance used, the integrity of the prosthesis, the degree and extent of tissue ingrowth, and the time delay between surgical replacement and MRI. The ABC ligament gives a uniform low signal when freshly placed, which is slowly lost by the incorporation of tissue ingrowth, which gives an intermediate signal on T 1weighted sequences. The loss of M R visualization of the ABC ligament is not synonymous with failure and should be interpreted with knowledge of the time interval. An absent ligament in the period up to 1 year is more likely to indicate rupture and failure of the ligament, whereas the bulky tissue ingrowth that develops as time progresses and particularly after 1 year will render the intact ligament's filaments indiscernible. In the normal intercondylar area, the MR contrast of the structures is clearly defined, due to the collagen of the
97
ligament and surrounding fat. Following surgery, however, a wide variety of tissue composition may occur, due to varying proportions of fat, fluid, fibrous tissue and synovium, which leads to an unpredictable MR image and often with a loss of M R contrast. The initial healing process is dependent on synovial and cellular proliferation with revascularization [14,15] and is independent of the type and origin of the graft. However, with patellar tendon grafts the process subsides after approximately 26 weeks [15] and the ligament begins to revert to a normal appearance, which is reflected in the low M R signal at 1 year. On the other hand filamentous carbon in the synthetic ligament continues to be degraded and the fibrous proliferation is encouraged, so that at 1 year, such tissue forms the bulk of the neo-ligament leading to the variable and often isointense signal on M R even in a clinically stable knee. IV gadolinium did not enhance the visualization of the neo-ligament. This may be explained by the fact that all the cases injected were after 1 year, and alterations in vascularity would have substantially subsided and the fibrous ingrowth may also have reduced. Quantitative assessment of the M R signal of autologous grafts, with time, have shown that the unimpinged ACL has a constant and low signal intensity up to 1 year of follow up, compared with an increased signal in the distal two thirds, when the neo-ligament is impinged, and this persists beyond 2 years [16]. These measurements, although useful, are clearly specific for autologous grafts and are unlikely to be of value in prosthetic ligaments, which is highlighted by the fact that 70% of patients in the series had a large soft tissue mass and the majority of these had an indiscernible prosthesis. The results in our study may be affected by the time delay between the assessment and the MRI examination in some patients. However, our results clearly identify the importance with regard to outcome of the positioning of the tibial tunnel and the relationship of the neo-ligament to the intercondylar line. If the orifice is situated posterior to the mid-point, roof impingement is avoided and induced collagen ingrowth will mature and align parallel to the axis of the implant [17,18]. Incorrect placement of the tibial tunnel would be expected to result in impingement and thus abnormal stresses leading to haphazard fibrous proliferation associated with fragmentation of the synthetic fibres which in turn increase tissue proliferation in a chaotic fashion. It is therefore not surprising that in our study the subjects with large intercondylar soft tissue masses and expanded tibial tunnels were associated with a failed repair. Sharp edges of the tibial cortex around the tunnel may also promote shearing tears of the prosthesis, especially in the presence of an anteriorly located orifice and/or abnormal tension of the ligament. The normal cruciate ligament broadens and increasingly fans out as the ligament courses distally towards its tibial insertion, accommodating the contour of the intercondylar roof, preventing acute angulation and impingement as the knee reaches terminal extension [18]. This broad anterior flare of the normal ACL insertion 18 mm deep in the sagittal plane was confirmed by MRI [15], and cannot be copied by a synthetic graft. The presence of this anterior flare has had a direct effect on the choice of the location of the tibial tunnel and also on the incidence of impingement. Initial suggestions of an eccentric placement 5 mm anterior and medial to the centre of the ACL insertion [19] were altered to placement in the centre of
98
CLINICALRADIOLOGY
the A C L insertion [18] and later to a point 3 mm posterior to that [161. The ABC graft inserted through tunnels of 6 mm diameter drilled in the tibia is easily assessed immediately after placement by M R I with the aim of excluding any residual bony impingement. The placement of these bone tunnels is critical to the success of the reconstruction. Our results clearly show that the tibial bone tunnel needs to lie posterior and parallel to the slope of the intercondylar roof with the knee in full extension and with its opening located at 50% of the sagittal depth of tibial joint surface. This correlates with the finding by Howell e t al. [16] that ligaments located at 42% of the sagittal tibial depth did not impinge on the roof of the intercondylar notch. The consequence of impingement of a ligament A C L graft is the production of signs and symptoms which include pain, limited knee extension, synovitis, and these may occur before the ligament fails. At this stage the symptoms may be rectified by a perarthroscopic notchplasty. The success of evaluation of the ligament depends on the quality of the M R technique. A thorough and complete protocol is essential to minimize the incidence of non-visualization of the ligament. Utilization of 5 mm thick sections, in the sagittal plane only [13], reduces the chances of proper visualization of the ligament and imaging in both the obliques of the sagittal and coronal plane must be done to optimize visualization of the ligament. The plane of scans must reflect the inclination and angulation of the line of the ligament track in the tibial bone tunnel and the intercondylar region. This will require a varied number of coronally and sagittally aligned, obliquely orientated scans. The scan width should be 3 mm or less to ensure that the partial volume effect does not obscure residual graft fibres. The ligament will be satisfactorily examined with Tl-weighted sequences but discrimination of fibrous tissue, effusion, meniscal and articular cartilage damage will require the addition of T2-weighted sequences. The knee should be examined in extension as this position highlights the effect of impingement of the A C L graft and the secondary signs of failure such as posterior tibial displacement and arcuate configuration of the PCL. The axial T 1-weighted image allows the dimensions of the orifice and the intercondylar notch to be measured and its position in relation to the AP diameter of the tibial plateau to be accurately measured as the effect of rotation of the tibia on the sagittal view could be discounted. The examination using 2-D image acquisition is time consuming and this can be significantly reduced by 3-D volume acquisition and image reconstruction in the appropriate plane. A further limitation of M R is its static format. Arthroscopic examination via a supra-patellar approach allows the dynamic demonstration of impingement of the ligament. This may, however, be achieved with real time 3-D M R acquisition but considerable computer resource is required which is not generally available. The importance of a full M R examination is highlighted by our study which has shown unequivocally that knee symptoms in a patient with a previous A C L reconstruction may be due to other intra-articular pathology such as meniscal and articular degenerative disease. The high level in our series may be due to the selection o f only symptomatic patients and to the higher proportion of chronic to acute patients but even in the acute presentation, meniscal and ligamentous injuries commonly coexist. It must also be empha-
sized that a successful ligament replacement may allow a return to full athletic pursuits which by its very nature increases the possibility of further intra-articular damage. The presence of an effusion on M R in our series indicated that the cause of symptoms was more likely to be due to other internal mechanical derangement rather than a failure of the ligament replacement. Although the placement of the ligament is the key to success, the length of follow up in any series must also be taken into account as many studies report 80 to 90% success rate at 2 years followed by a significant fall to 40 to 50% at 5 years [2022]. This is reflected in our series which averaged a 4 year follow up. In its present form the prosthetic A C L implant is unlikely to match the performance of the normal A C L but the correct placement in a location which avoids abnormal stresses and mechanical breakdown is crucial in increasing the likelihood of a successful outcome. M R I in the post-operative period provides valuable information on prognosis and detects problems which can be corrected before the development of mechanical damage and re-rupture of the neo-ligament.
REFERENCES
l DeHavenKE. Diagnosisof acute knee injuries with haemarthrosis. American Journal of Sports Medicine 1980;8:9-14. 2 NoyesFR, Bassett RW, Grood ES, Butler DL. Arthroscopyin acute traumatic haemarthrosis of the knee: incidence of anterior cruciate tears and other injuries. Journal of Bone and Joint Surgery 1980;62A:687-695. 3 McDaniel WJ Jr, Dameron TB Jr. The untreated ruptures of the anterior cruciate ligament. A follow-upstudy. Journal of Bone and Joint Surgery 1980;62A:696-705. 4 McDanielWJ, Jr, Dameron TB, Jr. The untreated anterior cruciate ligament rupture. Clinical Orthopaedics and Related Research 1983; 172:158 163. 5 FettoJF, Marshall JL. The natural historyand diagnosisof anterior cruciate ligament insufficiency. Clinical Orthopaedics and Related Research 1980;147:29-38. 6 Lynch MA, Henning CE, Glick KR. Knee joint surface changes. Long term follow-up meniscus tear treatment in stable anterior cruciate ligament reconstruction. Clinical Orthopaedics and Related Research 1983;172:148-153. 7 Sachs RA, Daniel DM, Stone ML, Garlein RF. Patellofemoral problems after ACL reconstruction. American Journal of Sports' Medicine 1989;17(6):760 765. 8 Strover AE, Firer P. The use of carbon fibre implants in anterior cruciate ligament surgery. Clinical Orthopaedics and Related Research 1985;196:88 98. 9 Goodship AE, Wilcock SA, Shah JS. The development of tissue around various prosthetic implants used as replacements for ligaments and tendons. Clinical Orthopaedics and Related Research 1985;196:61-68. 10 Lee JK, Yao L, Phelps CT, Wirth CR, CzajkaJ, LozmanJ. Anterior cruciate ligament tears: MR imaging compared with arthroscopy and clinical tests. Radiology 1988;166:861 864. 11 Aichioth PM, Patel DV, Jones CB, Ward JS. Combined inter- and extra-articular reconstructionusingcarbon-dacroncompositeprosthesis for chronic anterior cruciate instability. 2 to 6 year follow-up study. International Orthopaedic (Cycot) 1991; 15:219 227. 12 Rak KM, Gillogly SD, Schaefer RA, Yakes WE, Liljedahl RR. Anterior cruciate ligament reconstruction: evaluation with MR imaging. Radiology 1991;178:553-556. 13 MoeserP, BechtoldRE, Clark T, RovereG, Karstaedt N, Wolfman N. MR imaging of anterior cruciate ligament repair. Journal of Computer Assisted Tomography 1989;13(1):105-109. 14 ArnoczkySP, Tarvin GB, Marshall JL. Anterior cruciate ligament replacement usingpatellar tendon. Journal of Bone and Joint Surge~;v 1982;64A217. 15 Arnoczky SP, Warren RF, Minei JP. Replacement of an anterior cruciate ligamentusinga syntheticprosthesis. An evaluationof graft biology in the dog. American Journal of Sports Medicine1986;4: l 8.
MRI IN ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION 16 Howell SM, Berns GS, Farley TE. Unimpinged and impinged anterior cruciate ligament grafts: MR signal intensity measurements. Radiology 1991; 179:639 643. 17 Fujikawa K, Iseki F, Tomatsu T, Takeda T, Seedhom BB. Microscopic and histological findings after reconstruction of the anterior cruciate ligament by the Leeds-Keio artificial ligament. Knee 1984; 10:35-40. 18 Fujikawa K. Clinical study ofanterior cruciateligament reconstruction with the Leeds-Keio artificial ligament. In: Friedman MJ & Ferkel RD, eds. Prosthetic ligament reconstruction of the knee. Philadelphia: WB Saunders Co, 1988:132-139. 19 Odensten M, Gillquist J. Functional anatomy of the anterior cruciate ligament and a rationale for reconstruction. Journal of Bone and Joint Surgery (America) 1985;67:257-262. 20 ClancyWG, Nelson DA, Reider B, Narechania RG. Reconstruction
99
using one-third of the patella ligament, augmented by extra-articular tendon transfers. Journal of Bone and Joint Surgery 1982;64A:352. 21 Friedman MJ, Sherman OH, Fox JM, Del Pizzo W, Snyder SJ, Ferkel RJ. Autogenic anterior cruciate ligament (ACL) anterior reconstruction of the knee. Clinical Orthopaedics and Related Research 1985;196:9 14. 22 Dickason JM, Fox JM, Del Pizzo W, Friedman M J, Snydar SJ, Blazina M. Stabilisation of the knee joint for anterior instability a long term follow-up. Presented at the American Orthopaedic Society for Sports Medicine (AOSM) Annual Meeting, Williamsburg, VA, USA, 1983. 23 Odensten M, Lysholm J, Gitlquist J. Long-term follow-up study of a distal iliotibial band transfer (DIT) for anterolateral knee instability. Clinical Orthopaedics and Related Research 1983;176:129-135.
MRI of the Knee Following Prosthetic Anterior Cruciate Ligament Reconstruction V. N. CASSAR-PULLICINO, I. W. McCALL and A. E. STROVER*
Department of Diagnostic Imaging and *Institute of Orthopaedics, The Robert Jones and Agnes Hunt Orthopaedic Hospital, Oswestry, Shropshire The MRI appearances of the knee in 30 patients with residual symptoms following active biocomposite (ABC) reconstruction of the anterior cruciate ligament were prospectively correlated with the clinical and arthroscopic findings. The variable MR features of the neo-ligament depend on the integrity of the prosthesis, the degree and extent of tissue ingrowth, and the time delay between surgical placement and MRI. The ABC ligament gives a uniform low signal when freshly placed, which is slowly lost by the incorporation of tissue ingrowth and therefore the loss of MR visualization of the ABC ligament is not synonymous with failure. Analysis of the imaging features suggests that a neo-ligament emerging through the tibial tunnel either at or posterior to the mid point of the AP diameter of the tibia is likely to be intact (P = 0.008) compared with one placed anteriorly. The presence of an effusion (64 %) indicates that symptoms are more likely to be due to other internal mechanical derangement rather than failure of the neoligament. Meniscal tears were seen in 67% while osteo-chondral defects were noted in 30% of all patients. Excellent correlation with arthroscopic findings was established confirming intact neoligaments in 16 replacements. MRI in the post-operative period detects rectifiable problems before the development of irreversible mechanical damage and re-rupture of the neo-ligament, but a thorough MRI technique needs to be utilized in examining the entire knee. Cassar-
Pullicino, V.N., McCall, I.W. & Strover, A.E. (1994). Clinical Radiology 49, 89 99. MRI of the Knee Following Prosthetic Anterior Cruciate Ligament Reconstruction
Accepted for Publication 19 July 1993
The anterior cruciate ligament (ACL) is an important part of the stable knee. In the active sportsman, incompetence of the ACL gives symptoms of recurrent and usually painful giving way with repetitive haemarthoses often leading to damage of other intra-articular structures [ 1,2]. The natural history of untreated cruciate ligament tears over a 10 year period indicates that although the majority of patients returned to active sports, few are asymptomatic, 67% showing early radiographic evidence of degenerative changes, while 6% showed frank osteoarthritis [3,6]. Eighty-five per cent required meniscetomy during the 10 year follow-up period [4]. Orthopaedic opinion is increasingly in agreement that the majority of cases of ruptured anterior cruciate ligament in the sportsman require early diagnosis and surgical repair or reconstruction with reinforcement [5]. Reconstruction has also been advocated for the old or chronic cruciate deficient knee [3,4,6]. Methods of ligament reconstruction include autogenous grafts, allografts and synthetic ligaments. Allograft and synthetic ligament reconstructions are particularly attractive by virtue of eliminating donor site problems [7], and the move has been towards arthroscopic surgical introduction, which has dramatically reduced the morbidity associated with this kind of surgery. The synthetic ABC (active bio-composite) ligament is composed of bundles of carbon and polyester fibres combined together in an open partially braided arrangement (Fig. 1) which allows the ingrowth of fibrovascular tissue between the fibre bundles with the development of a 'neoligament' [8]. In the short term the artificial scaffold Correspondence to: Dr V. N. Cassar-Pullicino, Department of Diagnostic Imaging, The Robert Jones and Agnes Hunt Orthopaedic Hospital, Oswestry, Shropshire SY11 7AG.
(a)
(b) Fig. ! (a, b) The active bio-composite ligament. The button and loop at the ends of the two braided components depict the tibial and femoral attachments respectively. The connecting filaments are composed of intertwined carbon and polyester fibres. The construction is designed to promote fibrous ingrowth with neo-ligament formation.
90
CLINICAL RADIOLOGY
(a)
(c)
O)
(d)
Fig. 2 - An intact active bio-composite neo-ligament at 9 months is demonstrated throughout its full length in the relevant projections. (a) The sagittal view shows the ligament exiting the tibia in the correct position just over the 50% ratio. The intercondylar line lies anterior to the ligament with a satisfactory gap between the lower tip of the intercondylar line and the tibia. (b, c) The coronal view shows the ligament in the intercondylar notch curving around the lateral condyle, which is inserted into the lateral cortex of the femur by a carbon pin (d).
91
MR[ IN ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION
(a)
(b)
Fig. 3 - The coronal views show (a) an intact rico-ligament at 12 months, with ingrowth of intermediate signal fibrous tissues which has partly obliterated the low signal of the prosthesis. (b) An intact neo-ligament at 2 years post-operation in which the prosthesis has been completely obscured by fibrous ingrowth. The tibial tract is normal in width and position.
should be strong enough at the time of implantation to allow early functional rehabilitation of the knee, eventually maturing to become a permanent and durable reconstruction of the anterior cruciate ligament. Due to the fibrogenic properties of carbon, prosthetic degradation becomes an asset, rather than a deficiency [9]. Magnetic resonance scanning is now established as an accurate and non-invasive method of evaluating the integrity of the anterior cruciate ligament [10]. The entire length of the ligament can be visualized on a sagittal view, which is slightly oblique in the coronal plane. Magnetic resonance has also been used to evaluate autologous grafts with considerable success [i2,13] and should therefore provide a valuable method to evaluate postoperative problems in scaffold types of synthetic ligaments. This study was undertaken to assess the post-operative status of ACL reconstructions using the ABC ligament in patients who continued to have symptoms of pain, swelling and instability, and to correlate the M R appearances with clinical and arthroscopic findings. The postoperative integrity of the synthetic ligament and internal structures of the knee were studied by MRI to investigate ligament failure. MATERIALS AND METHODS
The prospective study involved 30 patients who had
POSTERIOR =~ 75 -o ~._ 65
+ +
"6 ~ 55 50%
~ 45 c = -~ '~ 35 ,= "6 g ~ 25 _o ANTERIOB
q + +
I Failed
P = 0.008
I Intact
L i g a m e n t status
Fig. 4 'Box and whisker' plot display of one-tailed 't'-test showing the effect of tibial tunnel location on eventual outcome of ligament status. The control box covers the middle 50~ of data values, between the upper and lower quartiles. The 'whiskers' extend to 1.5 times the interquartile range and the control line is the median. Extreme values are plotted as separate plots.
92
CLINICAL RADIOLOGY
(a)
(b) Fig. 5 - (a) The sagittal view of a normal knee with an intact natural anterior cruciate ligament demonstrating the relationship of the intercondylar line to the AP diameter o f the tibia (arrow). (b) An intact ligament shown well posterior to the lower tip of the intercondylar line. (c) A failed ligament shown emerging from the tibia anterior to the intercondylar line with a narrow gap between the lower tip and the tibia.
(c)
returned to the 'problem knee clinic' at the Robert Jones and Agnes Hunt Orthopaedic Hospital complaining of symptoms of pain, swelling or instability following reconstruction of the anterior cruciate ligament using the active biological composite (ABC) ligament. It was ascertained at the outset that the carbon and polyester fabric of the ABC ligament has a low signal on MRI by undertaking an M R study of the ligament with a phantom. MRI of the whole knee was performed utilizing a 0.5 T G E Max scanner, with a dedicated knee surface coil. The knee was positioned in full extension as this was thought to provide a better assessment of the whole length of the neo-ligament through both the tibial tunnel and within the joint. The entire knee was scanned in the sagittal plane, using Tl-weighted spin-echo (TR 800/TE 30), and T2-weighted Gradient-echo (TR 300/TE 20/20 ~ sequences, followed by coronal S.E. Tl-weighted scans and axial S.E. Tl-weighted scans of the patello-femoral articulation to include the neo-ligament, the intercondylar notch and its contents. The optimum coronal view was used to measure accurately the inclination of the neoligament to the sagittal plane. A Tl-weighted S.E. sequence (TR 800 TE 30) with thin (3 mm) continuous
MRI IN ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION
(a)
93
(b)
Fig. 6 - (a, b) The braided portion of the neo-ligament lies in the tibial canal. The tibial exit is anterior to the mid point of the AP tibial diameter and the process of impingement has resulted in enlargement of tibial orifice prior to ligament failure. The stabilizing button has been drawn into the tibia due to the excess tension of the ligament within this joint.
sections was directed in the oblique sagittal plane along the axis of the neo-ligament. F r o m this sequence, the optimum oblique sagittal image was used as a further scout view to measure the inclination of the neo-ligament to the coronal plane. Further oblique 3 m m contiguous coronal images were obtained through the axis of the neoligament. The total examination time was 50 min. Longitudinal sectional images were thus obtained of the neo-ligament from its point of entry into the tibial drill hole through the intercondylar notch of the knee to its point of exit via the 'over-the-top' route (Fig. 2). The first six patients were also imaged after intravenous Gadolinium enhancement (0.1 mmol/kg body weight) using T1 S.E. sequences. Preoperative and post-operative clinical assessment, anterior cruciate ligament reconstruction and arthroscopic assessment were undertaken by the same orthopaedic surgeon (AES). The M R images were independently assessed by two experienced radiologists (IWM and VCP). The radiologists were blind to the clinical findings, operative details and post-operative status of the patients. At final follow-up the knees were classified as either stable or unstable. A stable knee had to demonstrate the following findings: extension to at least 0 ~ both Lachman and drawer tests demonstrating no more than grade 1 laxity with a solid end point and an absent pivot shift. Tibial translation was measured using the West-
minster cruciometer [11] and stability in these cases was defined as an increased tibial translation of not more than 2.5 m m of difference when compared with the intact knee. All patients had arthroscopic 'second look' assessments 3 months following the initial operation, confirming an intact prosthetic ligament. Ninety per cent of patients had at least one other operative arthroscopy at various intervals during the next two years. Three patients (10%) refused a further follow-up arthroscopy, and were evaluated clinically. The interval between either arthroscopy or clinical evaluation and the M R I examination varied from 3 to 6 months. Four patients had more than one M R examination. A one-tailed 't'-test was performed to compare the effect of the location of the tibial tunnel expressed as a percentage of the tibial sagittal distance, with ligament status in failed and intact reconstructions. RESULTS The study group consisted of 30 patients with a mean age of 29 years (range 20-42 years). There were 26 males and four females. Surgery was performed for chronic knee instability in 25 patients and for an acute A C L rupture in five. The neo-ligament was seen as a thick, distinct structure with a uniformly low signal in 10 patients (Fig. 2). All
94
CLINICAL RADIOLOGY
(a)
(b)
Fig, 7 - (a) Enlargement of the tibial orifice is again shown on Tl-weighted image. (b) The T2-weighted image demonstrates high signal within oedematous friable tissue with absence of the ligament in the joint.
these patients showed clinical and arthroscopic evidence of successful reconstruction of the A C L with stable knees. Eight of these patients were imaged in the first postoperative year and two in the second. A few residual low signal strands were seen in two patients (Fig. 3), both of whom were imaged in the first post-operative year. The ligament was not visible in 18 patients (Fig. 3), five of w h o m were imaged in the first post-operative year and 13 in subsequent years. The appearance of the ABC ligament was dependent on the post-operative interval, as well as its structural integrity. In the first post-operative year 15 M R studies were performed with the ligament being clearly seen in eight cases, with two showing residual strands. Failure to visualize the ligament as a low signal structure was present in five patients, four of whom had confirmed ligament failure (Fig. 5) and one had an intact ligament. Soft tissue incorporation into the ligament in the intercondylar region was c o m m o n after 1 year, with only two exceptions. A soft tissue mass of an intermediate signal spanning the intercondylar path of the ligament was seen in 21 patients, of w h o m three had a discernible ligament, two had a few low signal strands, and in l 6 the ligament was not visualized. The obliteration of the prosthetic ligament by soft tissue incorporation did not in itself indicate ligament failure especially after the first post-operative year and
therefore indirect evidence was required to diagnose failure (Figs 3, 8). The position of the intersection of the intercondylar line (Blumensaats) with the tibial tunnel was found to be directly related to the failure rate of the ABC ligaments (Fig. 4). The location of the orifice anterior to the midpoint of the sagittal diameter of the tibia (Figs 4, 5) was associated with a high incidence of failure (P = 0.008) using a multiple box and whisker plot. Non-parametric testing using the Mann-Whitney (U-test) based on the 'ranked' data also proved significant ( P > 0.01). Posterior displacement of the femoral condyles was seen in five cases and was associated with failure in all of them. The posterior cruciate ligament was arcuate in three of these cases. The appearance of the tibial tunnel and its contained ligament also varied. Initially, the tunnel was created by a drill hole of 6 m m diameter but expansion of the drill hole, which was maximal subchondrally at the joint, giving a trumpet appearance, was seen in 21 (70%) of patients (Figs 6, 7). The expansion gave an intermediate signal on Tl-weighted images but was associated with a high T2 signal in some cases. The diameter of the tibial tunnel was increased in 80% of all cases and was seen in cases of -failure and also in patients with an intact ligament. All 10 patients with expansion of the tibial tunnel beyond 12 m m diameter had arthroscopically proven failed ligaments.
95
MRI IN ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION
(b)
(e)
(a)
Fig. 8 (a) The intact ligament at t 5 m o n t h s is shown well posterior to the intercondylar line. (b, c, d) The axial views show the low signal fibres in the tibia1 tunnel, exit and intercondylar notch. The precise point of exit of the ligament and any impingement on the inner condylar wall are best assessed with the axial view.
96
CLINICAL RADIOLOGY
Ca)
Fig. I O - The ligamentum patellae is shown to be crenated in knee extension.
(b) Fig. 9 - Arthroscopic appearances (a) of an intact ligament covered by well organized fibrous tissue. (b) A failed ligament showing extensive disorganized granulation tissue containing carbon debris. Both views are taken just above the point of emergence from the tibia.
M R I provided valuable information with regard to other sources of pain in these symptomatic patients. Meniscal tears were seen in 20 patients (67%): nine involving the medial meniscus, one the lateral meniscus, and in 10 both menisci were damaged. The ABC ligament was intact in 14 o f these cases with meniscal tears and torn in the remaining six (Fig. 11). Osteochondral defects, primarily in the medial compartment (Fig. l 1), with or without loss of joint space and osteophytes, were seen in nine (30%) of all patients and in 56% of patients with medial meniscal tears. These were confirmed arthroscopically. An intra-articular effusion was present in 18 patients (64%), 14 of whom had an intact ABC ligament while in four the ligament replacement had failed. Overall, 16 of the ABC ligament replacements were shown to he intact by M R I and arthroscopy, with a clinically stable knee. Fourteen of these 16 cases also had meniscal damage (87%). Eleven knees were shown to have failed ligaments on M R I , which was confirmed by arthroscopy (Fig. 9) and clinically the knees were unstable. Six of these cases also had associated meniscal tears. In three cases the M R I indicated that the ligament had
failed but clinically the knees appeared stable. Failure of the neo-ligament was not confirmed by arthroscopy in these cases, as this was not performed. In six cases, gadolinium D T P A was injected after the initial imaging sequences. In no case was enhancement demonstrated and no improvement in the imaging information was achieved. The status of the ligament within the intercondylar notch was also assessed in the axial plane. Lateral wall encroachment or narrowing of the transverse diameter of the notch was seen in four of the 11 failures. None of the intact ABC ligaments showed M R evidence of lateral wall encroachment (Fig. 8). Ten patients complained of anterior knee pain postoperatively. O f these eight had M R features related to the patellar tendon. The patellar tendon was significantly thickened in four cases, crenated due to laxity in three (Fig. 10) and showed evidence of tendinitis with inferior bursitis in one patient. In one patient the button at the tibial fixation site had been pulled into the tibial cortex causing pain which settled with time in the presence of a stable knee.
DISCUSSION M R imaging been limited to central third of graft and M R I
after A C L reconstruction has previously the evaluation of autologous grafts. The the patellar tendon was the source of the showed that the ligament was intact in
MRI IN ANTERIOR C R U C I A T E LIGAMENT RECONSTRUCTION
Fig. 11 - D e g e n e r a t i v e articular features are s h o w n on the c o r o n a l view w i t h an effusion in the joint, a n absent m e d i a l meniscus a n d a large s u b c h o n d r a l cyst. Artefact from the tibial fixation p o i n t is seen.
50 patients, which correlated in 92% with clinical examination and in 100% of patients at a second look arthroscopy [12]. Moeser et al. [13] described post-operative MR scans in 37 patients with ACL reconstruction using fascia lata which showed that only six ligaments were well defined, 10 ill-defined and 11 indiscernible, despite clinical stability. They concluded that the normal neo-ligament unlike the active ACL has a variable appearance, including non-visualization on M R and that criteria used in evaluating the native ligament will be inadequate to assess the repair. Our findings support this concept, and it is clear from our study that the M R appearance of the neoligament is a function of the type of substance used, the integrity of the prosthesis, the degree and extent of tissue ingrowth, and the time delay between surgical replacement and MRI. The ABC ligament gives a uniform low signal when freshly placed, which is slowly lost by the incorporation of tissue ingrowth, which gives an intermediate signal on T 1weighted sequences. The loss of M R visualization of the ABC ligament is not synonymous with failure and should be interpreted with knowledge of the time interval. An absent ligament in the period up to 1 year is more likely to indicate rupture and failure of the ligament, whereas the bulky tissue ingrowth that develops as time progresses and particularly after 1 year will render the intact ligament's filaments indiscernible. In the normal intercondylar area, the MR contrast of the structures is clearly defined, due to the collagen of the
97
ligament and surrounding fat. Following surgery, however, a wide variety of tissue composition may occur, due to varying proportions of fat, fluid, fibrous tissue and synovium, which leads to an unpredictable MR image and often with a loss of M R contrast. The initial healing process is dependent on synovial and cellular proliferation with revascularization [14,15] and is independent of the type and origin of the graft. However, with patellar tendon grafts the process subsides after approximately 26 weeks [15] and the ligament begins to revert to a normal appearance, which is reflected in the low M R signal at 1 year. On the other hand filamentous carbon in the synthetic ligament continues to be degraded and the fibrous proliferation is encouraged, so that at 1 year, such tissue forms the bulk of the neo-ligament leading to the variable and often isointense signal on M R even in a clinically stable knee. IV gadolinium did not enhance the visualization of the neo-ligament. This may be explained by the fact that all the cases injected were after 1 year, and alterations in vascularity would have substantially subsided and the fibrous ingrowth may also have reduced. Quantitative assessment of the M R signal of autologous grafts, with time, have shown that the unimpinged ACL has a constant and low signal intensity up to 1 year of follow up, compared with an increased signal in the distal two thirds, when the neo-ligament is impinged, and this persists beyond 2 years [16]. These measurements, although useful, are clearly specific for autologous grafts and are unlikely to be of value in prosthetic ligaments, which is highlighted by the fact that 70% of patients in the series had a large soft tissue mass and the majority of these had an indiscernible prosthesis. The results in our study may be affected by the time delay between the assessment and the MRI examination in some patients. However, our results clearly identify the importance with regard to outcome of the positioning of the tibial tunnel and the relationship of the neo-ligament to the intercondylar line. If the orifice is situated posterior to the mid-point, roof impingement is avoided and induced collagen ingrowth will mature and align parallel to the axis of the implant [17,18]. Incorrect placement of the tibial tunnel would be expected to result in impingement and thus abnormal stresses leading to haphazard fibrous proliferation associated with fragmentation of the synthetic fibres which in turn increase tissue proliferation in a chaotic fashion. It is therefore not surprising that in our study the subjects with large intercondylar soft tissue masses and expanded tibial tunnels were associated with a failed repair. Sharp edges of the tibial cortex around the tunnel may also promote shearing tears of the prosthesis, especially in the presence of an anteriorly located orifice and/or abnormal tension of the ligament. The normal cruciate ligament broadens and increasingly fans out as the ligament courses distally towards its tibial insertion, accommodating the contour of the intercondylar roof, preventing acute angulation and impingement as the knee reaches terminal extension [18]. This broad anterior flare of the normal ACL insertion 18 mm deep in the sagittal plane was confirmed by MRI [15], and cannot be copied by a synthetic graft. The presence of this anterior flare has had a direct effect on the choice of the location of the tibial tunnel and also on the incidence of impingement. Initial suggestions of an eccentric placement 5 mm anterior and medial to the centre of the ACL insertion [19] were altered to placement in the centre of
98
CLINICALRADIOLOGY
the A C L insertion [18] and later to a point 3 mm posterior to that [161. The ABC graft inserted through tunnels of 6 mm diameter drilled in the tibia is easily assessed immediately after placement by M R I with the aim of excluding any residual bony impingement. The placement of these bone tunnels is critical to the success of the reconstruction. Our results clearly show that the tibial bone tunnel needs to lie posterior and parallel to the slope of the intercondylar roof with the knee in full extension and with its opening located at 50% of the sagittal depth of tibial joint surface. This correlates with the finding by Howell e t al. [16] that ligaments located at 42% of the sagittal tibial depth did not impinge on the roof of the intercondylar notch. The consequence of impingement of a ligament A C L graft is the production of signs and symptoms which include pain, limited knee extension, synovitis, and these may occur before the ligament fails. At this stage the symptoms may be rectified by a perarthroscopic notchplasty. The success of evaluation of the ligament depends on the quality of the M R technique. A thorough and complete protocol is essential to minimize the incidence of non-visualization of the ligament. Utilization of 5 mm thick sections, in the sagittal plane only [13], reduces the chances of proper visualization of the ligament and imaging in both the obliques of the sagittal and coronal plane must be done to optimize visualization of the ligament. The plane of scans must reflect the inclination and angulation of the line of the ligament track in the tibial bone tunnel and the intercondylar region. This will require a varied number of coronally and sagittally aligned, obliquely orientated scans. The scan width should be 3 mm or less to ensure that the partial volume effect does not obscure residual graft fibres. The ligament will be satisfactorily examined with Tl-weighted sequences but discrimination of fibrous tissue, effusion, meniscal and articular cartilage damage will require the addition of T2-weighted sequences. The knee should be examined in extension as this position highlights the effect of impingement of the A C L graft and the secondary signs of failure such as posterior tibial displacement and arcuate configuration of the PCL. The axial T 1-weighted image allows the dimensions of the orifice and the intercondylar notch to be measured and its position in relation to the AP diameter of the tibial plateau to be accurately measured as the effect of rotation of the tibia on the sagittal view could be discounted. The examination using 2-D image acquisition is time consuming and this can be significantly reduced by 3-D volume acquisition and image reconstruction in the appropriate plane. A further limitation of M R is its static format. Arthroscopic examination via a supra-patellar approach allows the dynamic demonstration of impingement of the ligament. This may, however, be achieved with real time 3-D M R acquisition but considerable computer resource is required which is not generally available. The importance of a full M R examination is highlighted by our study which has shown unequivocally that knee symptoms in a patient with a previous A C L reconstruction may be due to other intra-articular pathology such as meniscal and articular degenerative disease. The high level in our series may be due to the selection o f only symptomatic patients and to the higher proportion of chronic to acute patients but even in the acute presentation, meniscal and ligamentous injuries commonly coexist. It must also be empha-
sized that a successful ligament replacement may allow a return to full athletic pursuits which by its very nature increases the possibility of further intra-articular damage. The presence of an effusion on M R in our series indicated that the cause of symptoms was more likely to be due to other internal mechanical derangement rather than a failure of the ligament replacement. Although the placement of the ligament is the key to success, the length of follow up in any series must also be taken into account as many studies report 80 to 90% success rate at 2 years followed by a significant fall to 40 to 50% at 5 years [2022]. This is reflected in our series which averaged a 4 year follow up. In its present form the prosthetic A C L implant is unlikely to match the performance of the normal A C L but the correct placement in a location which avoids abnormal stresses and mechanical breakdown is crucial in increasing the likelihood of a successful outcome. M R I in the post-operative period provides valuable information on prognosis and detects problems which can be corrected before the development of mechanical damage and re-rupture of the neo-ligament.
REFERENCES
l DeHavenKE. Diagnosisof acute knee injuries with haemarthrosis. American Journal of Sports Medicine 1980;8:9-14. 2 NoyesFR, Bassett RW, Grood ES, Butler DL. Arthroscopyin acute traumatic haemarthrosis of the knee: incidence of anterior cruciate tears and other injuries. Journal of Bone and Joint Surgery 1980;62A:687-695. 3 McDaniel WJ Jr, Dameron TB Jr. The untreated ruptures of the anterior cruciate ligament. A follow-upstudy. Journal of Bone and Joint Surgery 1980;62A:696-705. 4 McDanielWJ, Jr, Dameron TB, Jr. The untreated anterior cruciate ligament rupture. Clinical Orthopaedics and Related Research 1983; 172:158 163. 5 FettoJF, Marshall JL. The natural historyand diagnosisof anterior cruciate ligament insufficiency. Clinical Orthopaedics and Related Research 1980;147:29-38. 6 Lynch MA, Henning CE, Glick KR. Knee joint surface changes. Long term follow-up meniscus tear treatment in stable anterior cruciate ligament reconstruction. Clinical Orthopaedics and Related Research 1983;172:148-153. 7 Sachs RA, Daniel DM, Stone ML, Garlein RF. Patellofemoral problems after ACL reconstruction. American Journal of Sports' Medicine 1989;17(6):760 765. 8 Strover AE, Firer P. The use of carbon fibre implants in anterior cruciate ligament surgery. Clinical Orthopaedics and Related Research 1985;196:88 98. 9 Goodship AE, Wilcock SA, Shah JS. The development of tissue around various prosthetic implants used as replacements for ligaments and tendons. Clinical Orthopaedics and Related Research 1985;196:61-68. 10 Lee JK, Yao L, Phelps CT, Wirth CR, CzajkaJ, LozmanJ. Anterior cruciate ligament tears: MR imaging compared with arthroscopy and clinical tests. Radiology 1988;166:861 864. 11 Aichioth PM, Patel DV, Jones CB, Ward JS. Combined inter- and extra-articular reconstructionusingcarbon-dacroncompositeprosthesis for chronic anterior cruciate instability. 2 to 6 year follow-up study. International Orthopaedic (Cycot) 1991; 15:219 227. 12 Rak KM, Gillogly SD, Schaefer RA, Yakes WE, Liljedahl RR. Anterior cruciate ligament reconstruction: evaluation with MR imaging. Radiology 1991;178:553-556. 13 MoeserP, BechtoldRE, Clark T, RovereG, Karstaedt N, Wolfman N. MR imaging of anterior cruciate ligament repair. Journal of Computer Assisted Tomography 1989;13(1):105-109. 14 ArnoczkySP, Tarvin GB, Marshall JL. Anterior cruciate ligament replacement usingpatellar tendon. Journal of Bone and Joint Surge~;v 1982;64A217. 15 Arnoczky SP, Warren RF, Minei JP. Replacement of an anterior cruciate ligamentusinga syntheticprosthesis. An evaluationof graft biology in the dog. American Journal of Sports Medicine1986;4: l 8.
MRI IN ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION 16 Howell SM, Berns GS, Farley TE. Unimpinged and impinged anterior cruciate ligament grafts: MR signal intensity measurements. Radiology 1991; 179:639 643. 17 Fujikawa K, Iseki F, Tomatsu T, Takeda T, Seedhom BB. Microscopic and histological findings after reconstruction of the anterior cruciate ligament by the Leeds-Keio artificial ligament. Knee 1984; 10:35-40. 18 Fujikawa K. Clinical study ofanterior cruciateligament reconstruction with the Leeds-Keio artificial ligament. In: Friedman MJ & Ferkel RD, eds. Prosthetic ligament reconstruction of the knee. Philadelphia: WB Saunders Co, 1988:132-139. 19 Odensten M, Gillquist J. Functional anatomy of the anterior cruciate ligament and a rationale for reconstruction. Journal of Bone and Joint Surgery (America) 1985;67:257-262. 20 ClancyWG, Nelson DA, Reider B, Narechania RG. Reconstruction
99
using one-third of the patella ligament, augmented by extra-articular tendon transfers. Journal of Bone and Joint Surgery 1982;64A:352. 21 Friedman MJ, Sherman OH, Fox JM, Del Pizzo W, Snyder SJ, Ferkel RJ. Autogenic anterior cruciate ligament (ACL) anterior reconstruction of the knee. Clinical Orthopaedics and Related Research 1985;196:9 14. 22 Dickason JM, Fox JM, Del Pizzo W, Friedman M J, Snydar SJ, Blazina M. Stabilisation of the knee joint for anterior instability a long term follow-up. Presented at the American Orthopaedic Society for Sports Medicine (AOSM) Annual Meeting, Williamsburg, VA, USA, 1983. 23 Odensten M, Lysholm J, Gitlquist J. Long-term follow-up study of a distal iliotibial band transfer (DIT) for anterolateral knee instability. Clinical Orthopaedics and Related Research 1983;176:129-135.