Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
Current Problems in Diagnostic Radiology journal homepage: www.cpdrjournal.com
The Value of Accurate Magnetic Resonance Characterization of Posterior Cruciate Ligament Tears in the Setting of Multiligament Knee Injury: Imaging Features Predictive of Early Repair vs Reconstruction Christoper C. Goiney, MD, Jack Porrino, MDn, Bruce Twaddle, MD, Michael L. Richardson, MD, Hyojeong Mulcahy, MD, Felix S. Chew, MD Department of Radiology and Orthopaedic Surgery, University of Washington, Seattle, WA
Multiligament knee injury (MLKI) represents a complex set of pathologies treated with a wide variety of surgical approaches. If early surgical intervention is performed, the disrupted posterior cruciate ligament (PCL) can be treated with primary repair or reconstruction. The purpose of our study was to retrospectively identify a critical length of the distal component of the torn PCL on magnetic resonance imaging (MRI) that may predict the ability to perform early proximal femoral repair of the ligament, as opposed to reconstruction. A total of 50 MLKIs were managed at Harborview Medical Center from May 1, 2013, through July 15, 2014, by an orthopedic surgeon. Following exclusions, there were 27 knees with complete disruption of the PCL that underwent either early reattachment to the femoral insertion or reconstruction and were evaluated using preoperative MRI. In a consensus fashion, 2 radiologists measured the proximal and distal fragments of each disrupted PCL using preoperative MRI in multiple planes, as needed. MRI findings were correlated with what was performed at surgery. Those knees with a distal fragment PCL length of Z 41 mm were capable of, and underwent, early proximal femoral repair. With repair, the distal stump was attached to the distal femur. Alternatively, those with a distal PCL length of r32 mm could not undergo repair because of insufficient length and as such, were reconstructed. If early surgical intervention for an MLKI involving disruption of the PCL is considered, attention should be given to the length of the distal PCL fragment on MRI to plan appropriately for proximal femoral reattachment vs reconstruction. If the distal PCL fragment measures Z41 mm, surgical repair is achievable and can be considered as a surgical option. & 2015 Mosby, Inc. All rights reserved.
Introduction Multiligament knee injury (MLKI) is a rare and clinically challenging injury that may occur in the setting of tibiofemoral dislocation.1 Magnetic resonance imaging (MRI) for MLKI is vital for preoperative planning.2-5 Accurate diagnosis and subsequent management of posterior cruciate ligament (PCL) injuries in the setting of MLKI is of central importance owing to the integral function of the PCL in posterior knee stability. Once diagnosed, a common approach to a highgrade PCL injury in the context of MLKI is to delay surgery for several weeks, with more pressing injuries in a patient with polytrauma taking precedence over injury to the knee. If surgery is delayed, it is inevitable that the PCL will be treated with autograft or allograft reconstruction, as delay results in obscuration of tissue planes and loss of tissue integrity with ligament retraction.1,6-9 However, if early surgical intervention is planned (within 3 weeks from the time of injury) then primary repair of the PCL can be achieved10-12 and has demonstrated positive clinical outcomes comparable with reconstruction.13,14
n
Reprint requests: Jack Porrino, MD, Department of Radiology and Orthopaedic Surgery, University of Washington, 4245 Roosevelt Way NE, Box 354755, Seattle, WA 98105. E-mail address: [email protected] (J. Porrino). http://dx.doi.org/10.1067/j.cpradiol.2015.06.005 0363-0188/& 2015 Mosby, Inc. All rights reserved.
In MLKI, PCL disruption occurs most frequently at the femoral insertion or midsubstance of the ligament, although distal injury typically involves an attached osseous fragment related to either avulsion or coexistent tibial plateau fracture. In this context, with repair, the torn PCL is reattached to the distal femur. As such, the distal PCL stump must be of sufficient length to allow for reattachment. The purpose of our study was to retrospectively identify those patients with a documented MLKI, as encountered and recorded by the orthopedic surgeon (B.T.) at Harborview Medical Center, an allied hospital of the University of Washington, over a 1-year period and to separate those with surgically confirmed injury of the PCL for more detailed review. In those with a surgically documented disruption of the PCL, we sought to identify a possible threshold length on the preoperative MRI of the torn distal PCL fragment, which is necessary to allow for successful proximal femoral repair of the PCL, as opposed to reconstruction.
Materials and Methods Approval from the Investigational Review Board at the University of Washington was obtained for this study. During the period of May 1, 2013, through July 15, 2014, 50 knees with MLKI (47 patients, 3 with multiligament injury to both
2
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
the right and the left knee) were encountered and addressed by the same orthopedic surgeon (B.T.) at Harborview Medical Center, using early surgical intervention as part of their primary algorithm of care. The MLKI was defined as surgically confirmed injury of a cruciate ligament, with additional injury to the other cruciate ligament or collateral ligaments or both, with sufficient instability or injury pattern to have warranted surgical intervention and treatment. These patients were prospectively cataloged by the operating orthopedic surgeon (B.T.) into an MLKI database. Following exclusion of those younger than 18 years at the time of injury (1 patient, with multiligament injury to both the right and the left knee), we retrospectively reviewed each operative report (48 knees from 46 patients) and tabulated the presence or absence of injury to the following structures, as well as the type of repair or reconstruction when applicable: anterior cruciate ligament, PCL, medial collateral ligament or posteromedial corner, and fibular collateral ligament or posterolateral corner. In the operative report, PCL injury was classified as 1. 2. 3. 4.
Grade 1 or partial-thickness sprain. Proximal femoral avulsion. Midsubstance disruption. Distal tibial osseous avulsion or attached to part of a tibial plateau fracture.
In addition, patient demographics including sex, age at the time of injury, and extremity side were recorded. Those knees with a surgically documented injury of the PCL were identified (41 knees from 40 patients) and were subsequently tabulated separately for more comprehensive operative report and MRI review. There were 6 knees from 6 different patients with a specified Grade 1 or partial-thickness sprain to the PCL per operative report that were not treated with either repair or reconstruction and were not applicable to the study and were therefore excluded, resulting in 35 remaining knees from 35 different patients. In 3 of the 35 remaining knees (3 different patients), the exact injury of the PCL was unclear at the time of operative intervention per operative report (1 due to obscuration by the medial head of the gastrocnemius at the time of repair of the collateral structures, and subsequent refusal by the patient for further intervention at the time of planned cruciate repair; 1 related to complete maceration of the PCL; and the last related to poor visualization of the PCL because of adjacent posterior capsule injury collectively repaired using anchors and sutures in the posterior tibial plateau), and as such, these were excluded from the study, resulting in 32 knees from 32 different patients. There were 4 (4 of 32) knees from 4 different patients with distal tibial osseous avulsion or PCL attached to a part of a tibial plateau fracture. In 1 of these cases of surgically confirmed distal tibial osseous avulsion, there was no accompanying MRI, resulting in 31 knees from 31 different patients. The 3 (3 of 31) remaining knees from 3 different patients with a surgically confirmed distal osseous avulsion from the tibia or PCL attached to a component of a coexistent tibial plateau fracture had an intact PCL and were all treated with bony reattachment, and as such, they were excluded from MRI PCL measurements, resulting in 28 knees from 28 patients. Finally, in 1 knee with intraoperative confirmation of a midsubstance PCL tear, operative intervention was not pursued, as there was sufficient posterior stability intraoperatively despite the presence of injury, and the patient’s age and additional injuries rendered the patient a poor candidate for aggressive intervention. Following this exclusion, there were 27 knees from 27 patients remaining for review. Inclusion and exclusion criteria are demonstrated in Figure 1.
Fig. 1. A flowchart demonstrating stepwise inclusion and exclusion criteria of the study. The left column demonstrates knee and patient inclusions, while the right column demonstrates knee and patient exclusions.
Following exclusions, 2 musculoskeletal trained radiologists (H. M. and J.P.) reviewed the MRI studies in a consensus fashion in an effort to establish the maximal length of the proximal and distal
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
components of the disrupted PCL. Imaging within all 3 planes was used when necessary, as the course of the PCL, in particular when disrupted, can be variable. When measured in the sagittal plane, measurements were acquired along the posterior aspect of the ligament, as this reflects the longest course. Additionally, in certain cases, multiple measurements on a single slice were necessary to obtain the most accurate length, owing to the contour of the ligament. These data were correlated with the operative report description of injury and intervention. All MR studies acquired at our institution were performed on a GE (Signa HDxt) 1.5-T magnet, with the following pulse sequences collectively acquired and reviewed: axial T2 fat suppressed, short tau inversion recovery (STIR); coronal proton density fat suppressed, STIR, T1; and sagittal proton density, T2 fat suppressed, STIR. Two knee MR studies were performed at an outside institution before patient transfer. These studies used a combination of axial proton density fat-suppressed; coronal T1, proton density fatsuppressed, STIR; and sagittal proton density fat-suppressed, T2, T2 fat-suppressed imaging. The magnetic field strength for these 2 examinations was not specified on the available images.
Results Among the remaining 27 knees, there were 4 female and 23 male patients. Age at the time of injury ranged from 18.9-79.9 years. There were 14 right knees and 13 left knees. Based on operative report, there were 12 knees from 12 different patients encountered with proximal femoral PCL avulsion and 15 knees from 15 different patients with a midsubstance PCL disruption. On inspection of each MRI by the 2 musculoskeletal trained radiologists (H.M. and J.P.), there was consensus with the operative description of injury. Specifically, the MRI of each knee with an operative description of proximal femoral avulsion exhibited a disrupted PCL, with tear occurring at the proximal aspect of the ligament,
3
and with no demonstrable proximal fragment (Fig 2). In those with an operative description of midsubstance tear, the MRI demonstrated disruption of the PCL at the middle third of the ligament, with both a proximal and distal fragment evident and measurable (Fig 3). Using the unpaired, 2-tailed t test, the lengths of the distal fragments of the disrupted PCLs of those knees treated with proximal femoral repair were significantly longer than that of those treated with reconstruction (P o 0.000001). Specifically, those MRI studies with a distal fragment of PCL measuring 41 mm or more (n ¼ 12 of 27) underwent repair. The distal fragment in those 12 knees ranged from 41-52 mm in length, with a mean length of 45.9 mm. Alternatively, in those with a distal fragment measuring 32 mm or less (n ¼ 15 of 27), the PCL injury required reconstruction. In those 15 knees, the distal fragment ranged from 16-32 mm in length, with a mean distal fragment length of 27.3 mm. The distributions of the lengths of the distal PCL fragments of those repaired vs those reconstructed are presented in Figure 4. Using the unpaired, 2-tailed t test, the difference in age in those patients who underwent proximal femoral repair (mean age of 45.2 years at the time of injury) was not statistically significantly different (P ¼ 0.20) from that in those who underwent reconstruction (mean age of 36.8 years at the time of injury). Age distribution of the 2 groups is demonstrated in Figure 5. Again using the unpaired, 2-tailed t test, there was no statistically significant difference between sexes for distal PCL fragment length (P ¼ 0.53; Fig 6) nor between the left and the right knee for distal PCL fragment length (P ¼ 0.70; Fig 7). These data were analyzed using R, a free software environment for statistical computing and graphics.15
Discussion The MLKI encompasses a heterogeneous spectrum of injuries, ranging from relatively minor sprain of a cruciate and collateral
Fig. 2. Sagittal proton density fat-suppressed MR images of a 63-year-old man with a multiligament knee injury. The curved arrow in image (A) denotes the site of proximal femoral avulsion of the posterior cruciate ligament. In image (B), measurements “A” and “B” were summed to determine the length of the distal posterior cruciate ligament stump, in this case, 45 mm. The posterior cruciate ligament was repaired.
4
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
Fig. 3. Sagittal proton density fat-suppressed images of a 38-year-old man with a multiligament knee injury. The arrow in (A) denotes the site of midsubstance disruption of the posterior cruciate ligament. In (B), measurement “A” reflects the length of the proximal posterior cruciate ligament stump (22 mm), and measurement “B” reflects the length of the distal posterior cruciate ligament stump (27 mm). Reconstruction of the posterior cruciate ligament was required.
ligament to more severe multiligamentous disruption. A severe MLKI is typically the result of tibiofemoral joint dislocation,1 posing challenges both in the early phase related to neurovascular complications and in the delayed phase with a long recovery period often up to 2 years. Although there is potentially good functional outcome, there are near-unanimous residual deficits.7 A variety of treatment options exist for an MLKI depending on additional injuries incurred in the setting of polytrauma and resultant hemodynamic stability of the patient, availability of a surgeon experienced in managing these challenging surgical cases, and his or her particular preference for early vs delayed surgery. At our institution, a level 1 trauma center with a broad catchment area,
an algorithmic approach to early (within 3 weeks of injury), aggressive (repair or reconstruction of all injuries practically possible) management of MLKIs has been devised and is used for the treatment of this complex injury. This form of early and aggressive management of an MLKI has been supported in the literature.16 It should be noted that in those centers with a defined treatment algorithm for MLKI, variability does exist; the purpose of our study is not to advocate a particular MLKI treatment algorithm over the other. As part of our treatment algorithm, injuries of the PCL are managed with reattachment at their site of avulsion if possible, or reconstruction of the PCL is performed if reattachment does not appear technically achievable at the time of surgery.
Fig. 4. A box-and-whisker plot shows median distal PCL length, with repaired and reconstructed PCLs compared. Thick black horizontal lines show median, thin black horizontal lines show interquartile range, whiskers show observed range of nonoutlier data, and the small circle represents an extreme outlier.
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
5
Fig. 5. The box plot distribution of age among those patients who underwent PCL reconstruction vs those who underwent repair.
Evaluation of the preoperative MRI in the setting of an MLKI can be challenging, as there is an inordinate amount of pathology to decipher and separate.5 In the context of such varied pathology and numerous documented surgical approaches, identifying injury to the PCL and other ligaments by MRI is clearly of value to the clinician in preoperative planning.2-4 Quantitative information ascertainable from the preoperative MRI that can reliably predict whether reconstruction or repair of the disrupted PCL will be the likely treatment of choice can influence preparation, positioning of incisions, and graft option availability at the time of surgery. In our series, we retrospectively evaluated the preoperative MRI of 27 knees with a surgically confirmed full-thickness disruption of the proximal or mid-PCL, in the context of an MLKI. The preoperative MRI is used to localize the torn PCL and to also provide an estimate of the length of the distal tibial stump. However, to our knowledge, no standardized measurement of distal tibial stump length that is sufficient for proximal femoral
repair has been reported. With repair to the PCL’s femoral insertion, the distal fragment of the PCL is reattached to the distal femur, which requires sufficient length. On the contrary, if there is an insufficient length of distal tibial PCL to work with, the ligament is reconstructed using an autograft of both the semitendinosus tendons or an Achilles tendon allograft. Based on our series, the ability to perform a proximal femoral PCL repair, as opposed to PCL reconstruction, appears to have a threshold distal tibial PCL stump length. Specifically, all the evaluated MR knee studies in which the distal fragment of the PCL measured 32 mm or less required reconstruction when intervention was undertaken (15 of 27). In those with a distal PCL stump length of 41 mm or more (12 of 27), the ligament was successfully (stability demonstrated with testing at 901 of flexion) repaired or reattached to the distal femur. As one would expect, integrity of the distal PCL fragment is also necessary for repair, as sutures placed in the ligament must maintain their position, and
Fig. 6. The box plot distribution of distal fragment PCL length among female and male patients.
6
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
Fig. 7. The box plot distribution of distal fragment PCL length between left and right knees.
both the ligament and sutures must be able to resist significant force applied by the surgeon before repair, a test that was performed as part of the intraoperative surgical assessment. All 12 knees with a sufficient distal PCL stump length for repair exhibited sufficient quality for repair. Notably, all cases in which the PCL appeared to be avulsed from the distal femur on MRI underwent repair, without exception. Alternatively, those MR studies demonstrating a tear through the middle third of the PCL, resulting in a demonstrable and measurable proximal and distal fragment, required reconstruction. As expected, the more distal the site of PCL disruption, the shorter the distal fragment. One could extrapolate from our series that if a tear occurs within the middle third of the PCL, with a clearly defined proximal and distal fragment present on MRI, reconstruction is required, regardless of the length of the distal PCL fragment. Nevertheless, our data also suggest that there is value in measuring the distal fragment, as there is a clear line of demarcation separating those capable of proximal repair back onto the femur from those requiring reconstruction (repair Z41 mm vs reconstruction r32 mm). Perhaps those cases that fall into the gray zone of 3241 mm can be separated based on the presence or absence of a demonstrable proximal fragment, or a tear that appears to be present at the level of the middle third of the ligament on MRI. A limitation of our study was the lack of intraoperative assessment of PCL fragment lengths to correlate with the aforementioned MRI measurements acquired. Rather, intraoperative findings were described in a qualitative fashion (proximal femoral avulsion vs midsubstance rupture) without a quantitative assessment. Unfortunately, the retrospective nature of the study limited an intraoperative quantitative correlate. In addition to measurement of PCL length, the anatomical location of some PCL fragments on the preoperative MRI was particularly noteworthy. For example, in one MRI of a knee with a surgically confirmed proximal femoral PCL avulsion, the ligament was retracted and flipped along the posterior midline of the tibial plateau (Fig 8). Only on close inspection is this ligament identifiable on the preoperative MRI, as it was markedly displaced from its normal location and could easily be confused for macerated to the point that is no longer demonstrable. During our retrospective review, once identified on the preoperative MRI, the distal component of the torn PCL measured 52 mm, and there was no
appreciable proximal stump, findings that support the ability to perform repair. During operative exploration, based on the appearance of this remnant of the PCL alerting the surgeon to the possibility of repair being achievable, the avulsed PCL was localized through an approach required to repair the complex lateral and posterolateral injury and successfully reattached to the distal femur. The patient achieved excellent functional PCL stability. Furthermore, this particular case highlights that with PCL disruption in the setting of an MLKI, the distance separating the femoral PCL attachment site from the distal PCL fragment does not appear to reflect a reliable means of determining whether the disrupted PCL is or is not capable of being managed with proximal femoral repair. With MLKIs, the resultant displacement of the tibia on the
Fig. 8. A sagittal proton density fat-suppressed image of a 28-year-old man with a multiligament knee injury. The arrow points to the posteriorly and inferiorly displaced distal stump of the proximally avulsed posterior cruciate ligament. In this case, the stump measured 52 mm and was repaired.
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
7
Fig. 9. Sagittal proton density (A) and coronal proton density fat-suppressed (B) images of a 44-year-old woman with a multiligament knee injury. The arrow in (A) denotes the distal stump of the proximally avulsed posterior cruciate ligament, deceptively truncated in the sagittal plane. The arrows in coronal image (B) denote the transverse course of the distal stump, which collectively measured 45 mm and was successfully repaired.
femur, combined with potential displacement of the disrupted PCL, likely renders this distance an unreliable marker in the determination of proximal femoral PCL repair vs reconstruction. In a second case with proximal femoral PCL avulsion, there was sufficient displacement of the torn PCL, simulating a markedly foreshortened distal stump when inspecting the ligament in the sagittal plane alone. On close inspection, using all 3 planes, the PCL had assumed a flattened orientation, coursing partly in a transverse fashion within the intercondylar notch, as opposed to its normal craniocaudal course (Fig 9). If analyzed only in the sagittal plane, this ligament appears markedly truncated. During our retrospective review, the distal component of this torn ligament measured 45 mm, and there was no demonstrable proximal fragment, features suggesting that a successful proximal femoral repair might be attained. During operative exploration, guided by a high index of suspicion based on the preoperative MRI appearance, careful identification of the PCL was achieved as part of the initial arthroscopic assessment, and a miniopen technique to insert 2 robust sutures into the stump of the PCL to facilitate primary proximal femoral repair was performed. Again, this patient achieved excellent PCL stability at subsequent follow-up. In both of these cases, an appreciation of the course and length of the distal PCL stump had direct relevance to the surgical planning and approach. In the context of an MLKI, MRI serves as an important adjunct to physical examination for preoperative planning.2-4 Characterizing the torn PCL length (with attention given to the distal femoral stump) can predict primary proximal femoral repair vs reconstruction. In our series, a distal PCL stump length of Z 41 mm was predictive of proximal femoral repair. In addition, our study demonstrates the ability of the preoperative MRI to elucidate unusual positions and conformations of a torn PCL, which may be amenable to primary proximal femoral repair. By focusing on such details in the setting of an MLKI, MRI can aid in surgical planning and facilitate patient care in the setting of a complex and challenging injury.
References 1. Dodds AL, Narvani A, Williams A. Multiple ligament injuries of the knee. Orthop Trauma 2013;27:106–12. 2. Levy BA, Stuart MJ. How I manage the multiple-ligament injured (dislocated) knee. Oper Tech Sports Med 2011;19:27–33. 3. Seroyer ST, Mushal V, Harner CD. Management of the acute knee dislocation: The Pittsburgh experience. Injury 2008;39:710–8. 4. Twaddle BC, Hunter JC, Chapman JR, et al. MRI in acute knee dislocation: A prospective study of clinical, MRI and surgical findings. J Bone Joint Surg Br 1996;7(4):573–9. 5. Walker RE, McDougall D, Patel S, et al. Radiologic review of knee dislocation: From diagnosis to repair. Am J Roentgenol 2013;201:483–95. 6. Fanelli GC, Edson CJ, Beck JD. How I treat the multiple-ligament injured knee. Oper Tech Sports Med 2010;18:198–210. 7. Howells NR, Brunton LR, Robinson J, et al. Acute knee dislocation: An evidence based approach to the management of the multiligament injured knee. Injury 2011;42:1198–204. 8. Levy BA, Dajani KA, Whelan DB, et al. Decision making in the multiligamentinjured knee: An evidence-based systematic review. Arthroscopy 2009;25 (4):430–8. 9. Merritt AL, Wahl CJ. Rationale and treatment of multiple-ligament injured knee: The Seattle perspective. Oper Tech Sports Med 2011;19:51–72. 10. DiFelice GS, Lissy M, Haynes P. When to arthroscopically repair the torn posterior cruciate ligament. Clin Orthop Relat Res 2012;470:861–8. 11. Richter M, Bosch U, Wippermann B, et al. Comparison of surgical repair or reconstruction of the cruciate ligaments versus nonsurgical treatment in patients with traumatic knee dislocations. The American Journal of Sports Medicine. Am J Sports Med 2002;30(5):718–27. 12. Twaddle BC, Bidwell TA, Chapman JR. Knee dislocation: Where are the lesions? A prospective evaluation of surgical findings in 63 cases. J Orthop Trauma 2003;17(3):198–202. 13. Mariani PP, Santoriello P, Iannone S, et al. Comparison of surgical treatments for knee dislocation. Am J Knee Surg 1999;12:214–21. 14. Owens BD, Neault M, Benson E, et al. Primary repair of knee dislocations: Results in 25 patients (28 knees) at a mean follow-up of four years. J Orthop Trauma. 2007;21(2):92–6. 15. R Development Core Team R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2008. 〈http://www.R-project.org〉 [ISBN: 3-900051-07-0]. 16. Geeslin AG, LaPrade RF. Outcomes of treatment of acute grade-III isolated and combined posterolateral knee injuries: A prospective case series and surgical technique. J Bone Joint Surg Am 2011;93(18):1672–83.
Current Problems in Diagnostic Radiology journal homepage: www.cpdrjournal.com
The Value of Accurate Magnetic Resonance Characterization of Posterior Cruciate Ligament Tears in the Setting of Multiligament Knee Injury: Imaging Features Predictive of Early Repair vs Reconstruction Christoper C. Goiney, MD, Jack Porrino, MDn, Bruce Twaddle, MD, Michael L. Richardson, MD, Hyojeong Mulcahy, MD, Felix S. Chew, MD Department of Radiology and Orthopaedic Surgery, University of Washington, Seattle, WA
Multiligament knee injury (MLKI) represents a complex set of pathologies treated with a wide variety of surgical approaches. If early surgical intervention is performed, the disrupted posterior cruciate ligament (PCL) can be treated with primary repair or reconstruction. The purpose of our study was to retrospectively identify a critical length of the distal component of the torn PCL on magnetic resonance imaging (MRI) that may predict the ability to perform early proximal femoral repair of the ligament, as opposed to reconstruction. A total of 50 MLKIs were managed at Harborview Medical Center from May 1, 2013, through July 15, 2014, by an orthopedic surgeon. Following exclusions, there were 27 knees with complete disruption of the PCL that underwent either early reattachment to the femoral insertion or reconstruction and were evaluated using preoperative MRI. In a consensus fashion, 2 radiologists measured the proximal and distal fragments of each disrupted PCL using preoperative MRI in multiple planes, as needed. MRI findings were correlated with what was performed at surgery. Those knees with a distal fragment PCL length of Z 41 mm were capable of, and underwent, early proximal femoral repair. With repair, the distal stump was attached to the distal femur. Alternatively, those with a distal PCL length of r32 mm could not undergo repair because of insufficient length and as such, were reconstructed. If early surgical intervention for an MLKI involving disruption of the PCL is considered, attention should be given to the length of the distal PCL fragment on MRI to plan appropriately for proximal femoral reattachment vs reconstruction. If the distal PCL fragment measures Z41 mm, surgical repair is achievable and can be considered as a surgical option. & 2015 Mosby, Inc. All rights reserved.
Introduction Multiligament knee injury (MLKI) is a rare and clinically challenging injury that may occur in the setting of tibiofemoral dislocation.1 Magnetic resonance imaging (MRI) for MLKI is vital for preoperative planning.2-5 Accurate diagnosis and subsequent management of posterior cruciate ligament (PCL) injuries in the setting of MLKI is of central importance owing to the integral function of the PCL in posterior knee stability. Once diagnosed, a common approach to a highgrade PCL injury in the context of MLKI is to delay surgery for several weeks, with more pressing injuries in a patient with polytrauma taking precedence over injury to the knee. If surgery is delayed, it is inevitable that the PCL will be treated with autograft or allograft reconstruction, as delay results in obscuration of tissue planes and loss of tissue integrity with ligament retraction.1,6-9 However, if early surgical intervention is planned (within 3 weeks from the time of injury) then primary repair of the PCL can be achieved10-12 and has demonstrated positive clinical outcomes comparable with reconstruction.13,14
n
Reprint requests: Jack Porrino, MD, Department of Radiology and Orthopaedic Surgery, University of Washington, 4245 Roosevelt Way NE, Box 354755, Seattle, WA 98105. E-mail address: [email protected] (J. Porrino). http://dx.doi.org/10.1067/j.cpradiol.2015.06.005 0363-0188/& 2015 Mosby, Inc. All rights reserved.
In MLKI, PCL disruption occurs most frequently at the femoral insertion or midsubstance of the ligament, although distal injury typically involves an attached osseous fragment related to either avulsion or coexistent tibial plateau fracture. In this context, with repair, the torn PCL is reattached to the distal femur. As such, the distal PCL stump must be of sufficient length to allow for reattachment. The purpose of our study was to retrospectively identify those patients with a documented MLKI, as encountered and recorded by the orthopedic surgeon (B.T.) at Harborview Medical Center, an allied hospital of the University of Washington, over a 1-year period and to separate those with surgically confirmed injury of the PCL for more detailed review. In those with a surgically documented disruption of the PCL, we sought to identify a possible threshold length on the preoperative MRI of the torn distal PCL fragment, which is necessary to allow for successful proximal femoral repair of the PCL, as opposed to reconstruction.
Materials and Methods Approval from the Investigational Review Board at the University of Washington was obtained for this study. During the period of May 1, 2013, through July 15, 2014, 50 knees with MLKI (47 patients, 3 with multiligament injury to both
2
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
the right and the left knee) were encountered and addressed by the same orthopedic surgeon (B.T.) at Harborview Medical Center, using early surgical intervention as part of their primary algorithm of care. The MLKI was defined as surgically confirmed injury of a cruciate ligament, with additional injury to the other cruciate ligament or collateral ligaments or both, with sufficient instability or injury pattern to have warranted surgical intervention and treatment. These patients were prospectively cataloged by the operating orthopedic surgeon (B.T.) into an MLKI database. Following exclusion of those younger than 18 years at the time of injury (1 patient, with multiligament injury to both the right and the left knee), we retrospectively reviewed each operative report (48 knees from 46 patients) and tabulated the presence or absence of injury to the following structures, as well as the type of repair or reconstruction when applicable: anterior cruciate ligament, PCL, medial collateral ligament or posteromedial corner, and fibular collateral ligament or posterolateral corner. In the operative report, PCL injury was classified as 1. 2. 3. 4.
Grade 1 or partial-thickness sprain. Proximal femoral avulsion. Midsubstance disruption. Distal tibial osseous avulsion or attached to part of a tibial plateau fracture.
In addition, patient demographics including sex, age at the time of injury, and extremity side were recorded. Those knees with a surgically documented injury of the PCL were identified (41 knees from 40 patients) and were subsequently tabulated separately for more comprehensive operative report and MRI review. There were 6 knees from 6 different patients with a specified Grade 1 or partial-thickness sprain to the PCL per operative report that were not treated with either repair or reconstruction and were not applicable to the study and were therefore excluded, resulting in 35 remaining knees from 35 different patients. In 3 of the 35 remaining knees (3 different patients), the exact injury of the PCL was unclear at the time of operative intervention per operative report (1 due to obscuration by the medial head of the gastrocnemius at the time of repair of the collateral structures, and subsequent refusal by the patient for further intervention at the time of planned cruciate repair; 1 related to complete maceration of the PCL; and the last related to poor visualization of the PCL because of adjacent posterior capsule injury collectively repaired using anchors and sutures in the posterior tibial plateau), and as such, these were excluded from the study, resulting in 32 knees from 32 different patients. There were 4 (4 of 32) knees from 4 different patients with distal tibial osseous avulsion or PCL attached to a part of a tibial plateau fracture. In 1 of these cases of surgically confirmed distal tibial osseous avulsion, there was no accompanying MRI, resulting in 31 knees from 31 different patients. The 3 (3 of 31) remaining knees from 3 different patients with a surgically confirmed distal osseous avulsion from the tibia or PCL attached to a component of a coexistent tibial plateau fracture had an intact PCL and were all treated with bony reattachment, and as such, they were excluded from MRI PCL measurements, resulting in 28 knees from 28 patients. Finally, in 1 knee with intraoperative confirmation of a midsubstance PCL tear, operative intervention was not pursued, as there was sufficient posterior stability intraoperatively despite the presence of injury, and the patient’s age and additional injuries rendered the patient a poor candidate for aggressive intervention. Following this exclusion, there were 27 knees from 27 patients remaining for review. Inclusion and exclusion criteria are demonstrated in Figure 1.
Fig. 1. A flowchart demonstrating stepwise inclusion and exclusion criteria of the study. The left column demonstrates knee and patient inclusions, while the right column demonstrates knee and patient exclusions.
Following exclusions, 2 musculoskeletal trained radiologists (H. M. and J.P.) reviewed the MRI studies in a consensus fashion in an effort to establish the maximal length of the proximal and distal
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
components of the disrupted PCL. Imaging within all 3 planes was used when necessary, as the course of the PCL, in particular when disrupted, can be variable. When measured in the sagittal plane, measurements were acquired along the posterior aspect of the ligament, as this reflects the longest course. Additionally, in certain cases, multiple measurements on a single slice were necessary to obtain the most accurate length, owing to the contour of the ligament. These data were correlated with the operative report description of injury and intervention. All MR studies acquired at our institution were performed on a GE (Signa HDxt) 1.5-T magnet, with the following pulse sequences collectively acquired and reviewed: axial T2 fat suppressed, short tau inversion recovery (STIR); coronal proton density fat suppressed, STIR, T1; and sagittal proton density, T2 fat suppressed, STIR. Two knee MR studies were performed at an outside institution before patient transfer. These studies used a combination of axial proton density fat-suppressed; coronal T1, proton density fatsuppressed, STIR; and sagittal proton density fat-suppressed, T2, T2 fat-suppressed imaging. The magnetic field strength for these 2 examinations was not specified on the available images.
Results Among the remaining 27 knees, there were 4 female and 23 male patients. Age at the time of injury ranged from 18.9-79.9 years. There were 14 right knees and 13 left knees. Based on operative report, there were 12 knees from 12 different patients encountered with proximal femoral PCL avulsion and 15 knees from 15 different patients with a midsubstance PCL disruption. On inspection of each MRI by the 2 musculoskeletal trained radiologists (H.M. and J.P.), there was consensus with the operative description of injury. Specifically, the MRI of each knee with an operative description of proximal femoral avulsion exhibited a disrupted PCL, with tear occurring at the proximal aspect of the ligament,
3
and with no demonstrable proximal fragment (Fig 2). In those with an operative description of midsubstance tear, the MRI demonstrated disruption of the PCL at the middle third of the ligament, with both a proximal and distal fragment evident and measurable (Fig 3). Using the unpaired, 2-tailed t test, the lengths of the distal fragments of the disrupted PCLs of those knees treated with proximal femoral repair were significantly longer than that of those treated with reconstruction (P o 0.000001). Specifically, those MRI studies with a distal fragment of PCL measuring 41 mm or more (n ¼ 12 of 27) underwent repair. The distal fragment in those 12 knees ranged from 41-52 mm in length, with a mean length of 45.9 mm. Alternatively, in those with a distal fragment measuring 32 mm or less (n ¼ 15 of 27), the PCL injury required reconstruction. In those 15 knees, the distal fragment ranged from 16-32 mm in length, with a mean distal fragment length of 27.3 mm. The distributions of the lengths of the distal PCL fragments of those repaired vs those reconstructed are presented in Figure 4. Using the unpaired, 2-tailed t test, the difference in age in those patients who underwent proximal femoral repair (mean age of 45.2 years at the time of injury) was not statistically significantly different (P ¼ 0.20) from that in those who underwent reconstruction (mean age of 36.8 years at the time of injury). Age distribution of the 2 groups is demonstrated in Figure 5. Again using the unpaired, 2-tailed t test, there was no statistically significant difference between sexes for distal PCL fragment length (P ¼ 0.53; Fig 6) nor between the left and the right knee for distal PCL fragment length (P ¼ 0.70; Fig 7). These data were analyzed using R, a free software environment for statistical computing and graphics.15
Discussion The MLKI encompasses a heterogeneous spectrum of injuries, ranging from relatively minor sprain of a cruciate and collateral
Fig. 2. Sagittal proton density fat-suppressed MR images of a 63-year-old man with a multiligament knee injury. The curved arrow in image (A) denotes the site of proximal femoral avulsion of the posterior cruciate ligament. In image (B), measurements “A” and “B” were summed to determine the length of the distal posterior cruciate ligament stump, in this case, 45 mm. The posterior cruciate ligament was repaired.
4
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
Fig. 3. Sagittal proton density fat-suppressed images of a 38-year-old man with a multiligament knee injury. The arrow in (A) denotes the site of midsubstance disruption of the posterior cruciate ligament. In (B), measurement “A” reflects the length of the proximal posterior cruciate ligament stump (22 mm), and measurement “B” reflects the length of the distal posterior cruciate ligament stump (27 mm). Reconstruction of the posterior cruciate ligament was required.
ligament to more severe multiligamentous disruption. A severe MLKI is typically the result of tibiofemoral joint dislocation,1 posing challenges both in the early phase related to neurovascular complications and in the delayed phase with a long recovery period often up to 2 years. Although there is potentially good functional outcome, there are near-unanimous residual deficits.7 A variety of treatment options exist for an MLKI depending on additional injuries incurred in the setting of polytrauma and resultant hemodynamic stability of the patient, availability of a surgeon experienced in managing these challenging surgical cases, and his or her particular preference for early vs delayed surgery. At our institution, a level 1 trauma center with a broad catchment area,
an algorithmic approach to early (within 3 weeks of injury), aggressive (repair or reconstruction of all injuries practically possible) management of MLKIs has been devised and is used for the treatment of this complex injury. This form of early and aggressive management of an MLKI has been supported in the literature.16 It should be noted that in those centers with a defined treatment algorithm for MLKI, variability does exist; the purpose of our study is not to advocate a particular MLKI treatment algorithm over the other. As part of our treatment algorithm, injuries of the PCL are managed with reattachment at their site of avulsion if possible, or reconstruction of the PCL is performed if reattachment does not appear technically achievable at the time of surgery.
Fig. 4. A box-and-whisker plot shows median distal PCL length, with repaired and reconstructed PCLs compared. Thick black horizontal lines show median, thin black horizontal lines show interquartile range, whiskers show observed range of nonoutlier data, and the small circle represents an extreme outlier.
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
5
Fig. 5. The box plot distribution of age among those patients who underwent PCL reconstruction vs those who underwent repair.
Evaluation of the preoperative MRI in the setting of an MLKI can be challenging, as there is an inordinate amount of pathology to decipher and separate.5 In the context of such varied pathology and numerous documented surgical approaches, identifying injury to the PCL and other ligaments by MRI is clearly of value to the clinician in preoperative planning.2-4 Quantitative information ascertainable from the preoperative MRI that can reliably predict whether reconstruction or repair of the disrupted PCL will be the likely treatment of choice can influence preparation, positioning of incisions, and graft option availability at the time of surgery. In our series, we retrospectively evaluated the preoperative MRI of 27 knees with a surgically confirmed full-thickness disruption of the proximal or mid-PCL, in the context of an MLKI. The preoperative MRI is used to localize the torn PCL and to also provide an estimate of the length of the distal tibial stump. However, to our knowledge, no standardized measurement of distal tibial stump length that is sufficient for proximal femoral
repair has been reported. With repair to the PCL’s femoral insertion, the distal fragment of the PCL is reattached to the distal femur, which requires sufficient length. On the contrary, if there is an insufficient length of distal tibial PCL to work with, the ligament is reconstructed using an autograft of both the semitendinosus tendons or an Achilles tendon allograft. Based on our series, the ability to perform a proximal femoral PCL repair, as opposed to PCL reconstruction, appears to have a threshold distal tibial PCL stump length. Specifically, all the evaluated MR knee studies in which the distal fragment of the PCL measured 32 mm or less required reconstruction when intervention was undertaken (15 of 27). In those with a distal PCL stump length of 41 mm or more (12 of 27), the ligament was successfully (stability demonstrated with testing at 901 of flexion) repaired or reattached to the distal femur. As one would expect, integrity of the distal PCL fragment is also necessary for repair, as sutures placed in the ligament must maintain their position, and
Fig. 6. The box plot distribution of distal fragment PCL length among female and male patients.
6
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
Fig. 7. The box plot distribution of distal fragment PCL length between left and right knees.
both the ligament and sutures must be able to resist significant force applied by the surgeon before repair, a test that was performed as part of the intraoperative surgical assessment. All 12 knees with a sufficient distal PCL stump length for repair exhibited sufficient quality for repair. Notably, all cases in which the PCL appeared to be avulsed from the distal femur on MRI underwent repair, without exception. Alternatively, those MR studies demonstrating a tear through the middle third of the PCL, resulting in a demonstrable and measurable proximal and distal fragment, required reconstruction. As expected, the more distal the site of PCL disruption, the shorter the distal fragment. One could extrapolate from our series that if a tear occurs within the middle third of the PCL, with a clearly defined proximal and distal fragment present on MRI, reconstruction is required, regardless of the length of the distal PCL fragment. Nevertheless, our data also suggest that there is value in measuring the distal fragment, as there is a clear line of demarcation separating those capable of proximal repair back onto the femur from those requiring reconstruction (repair Z41 mm vs reconstruction r32 mm). Perhaps those cases that fall into the gray zone of 3241 mm can be separated based on the presence or absence of a demonstrable proximal fragment, or a tear that appears to be present at the level of the middle third of the ligament on MRI. A limitation of our study was the lack of intraoperative assessment of PCL fragment lengths to correlate with the aforementioned MRI measurements acquired. Rather, intraoperative findings were described in a qualitative fashion (proximal femoral avulsion vs midsubstance rupture) without a quantitative assessment. Unfortunately, the retrospective nature of the study limited an intraoperative quantitative correlate. In addition to measurement of PCL length, the anatomical location of some PCL fragments on the preoperative MRI was particularly noteworthy. For example, in one MRI of a knee with a surgically confirmed proximal femoral PCL avulsion, the ligament was retracted and flipped along the posterior midline of the tibial plateau (Fig 8). Only on close inspection is this ligament identifiable on the preoperative MRI, as it was markedly displaced from its normal location and could easily be confused for macerated to the point that is no longer demonstrable. During our retrospective review, once identified on the preoperative MRI, the distal component of the torn PCL measured 52 mm, and there was no
appreciable proximal stump, findings that support the ability to perform repair. During operative exploration, based on the appearance of this remnant of the PCL alerting the surgeon to the possibility of repair being achievable, the avulsed PCL was localized through an approach required to repair the complex lateral and posterolateral injury and successfully reattached to the distal femur. The patient achieved excellent functional PCL stability. Furthermore, this particular case highlights that with PCL disruption in the setting of an MLKI, the distance separating the femoral PCL attachment site from the distal PCL fragment does not appear to reflect a reliable means of determining whether the disrupted PCL is or is not capable of being managed with proximal femoral repair. With MLKIs, the resultant displacement of the tibia on the
Fig. 8. A sagittal proton density fat-suppressed image of a 28-year-old man with a multiligament knee injury. The arrow points to the posteriorly and inferiorly displaced distal stump of the proximally avulsed posterior cruciate ligament. In this case, the stump measured 52 mm and was repaired.
C.C. Goiney et al. / Current Problems in Diagnostic Radiology ] (2015) ]]]–]]]
7
Fig. 9. Sagittal proton density (A) and coronal proton density fat-suppressed (B) images of a 44-year-old woman with a multiligament knee injury. The arrow in (A) denotes the distal stump of the proximally avulsed posterior cruciate ligament, deceptively truncated in the sagittal plane. The arrows in coronal image (B) denote the transverse course of the distal stump, which collectively measured 45 mm and was successfully repaired.
femur, combined with potential displacement of the disrupted PCL, likely renders this distance an unreliable marker in the determination of proximal femoral PCL repair vs reconstruction. In a second case with proximal femoral PCL avulsion, there was sufficient displacement of the torn PCL, simulating a markedly foreshortened distal stump when inspecting the ligament in the sagittal plane alone. On close inspection, using all 3 planes, the PCL had assumed a flattened orientation, coursing partly in a transverse fashion within the intercondylar notch, as opposed to its normal craniocaudal course (Fig 9). If analyzed only in the sagittal plane, this ligament appears markedly truncated. During our retrospective review, the distal component of this torn ligament measured 45 mm, and there was no demonstrable proximal fragment, features suggesting that a successful proximal femoral repair might be attained. During operative exploration, guided by a high index of suspicion based on the preoperative MRI appearance, careful identification of the PCL was achieved as part of the initial arthroscopic assessment, and a miniopen technique to insert 2 robust sutures into the stump of the PCL to facilitate primary proximal femoral repair was performed. Again, this patient achieved excellent PCL stability at subsequent follow-up. In both of these cases, an appreciation of the course and length of the distal PCL stump had direct relevance to the surgical planning and approach. In the context of an MLKI, MRI serves as an important adjunct to physical examination for preoperative planning.2-4 Characterizing the torn PCL length (with attention given to the distal femoral stump) can predict primary proximal femoral repair vs reconstruction. In our series, a distal PCL stump length of Z 41 mm was predictive of proximal femoral repair. In addition, our study demonstrates the ability of the preoperative MRI to elucidate unusual positions and conformations of a torn PCL, which may be amenable to primary proximal femoral repair. By focusing on such details in the setting of an MLKI, MRI can aid in surgical planning and facilitate patient care in the setting of a complex and challenging injury.
References 1. Dodds AL, Narvani A, Williams A. Multiple ligament injuries of the knee. Orthop Trauma 2013;27:106–12. 2. Levy BA, Stuart MJ. How I manage the multiple-ligament injured (dislocated) knee. Oper Tech Sports Med 2011;19:27–33. 3. Seroyer ST, Mushal V, Harner CD. Management of the acute knee dislocation: The Pittsburgh experience. Injury 2008;39:710–8. 4. Twaddle BC, Hunter JC, Chapman JR, et al. MRI in acute knee dislocation: A prospective study of clinical, MRI and surgical findings. J Bone Joint Surg Br 1996;7(4):573–9. 5. Walker RE, McDougall D, Patel S, et al. Radiologic review of knee dislocation: From diagnosis to repair. Am J Roentgenol 2013;201:483–95. 6. Fanelli GC, Edson CJ, Beck JD. How I treat the multiple-ligament injured knee. Oper Tech Sports Med 2010;18:198–210. 7. Howells NR, Brunton LR, Robinson J, et al. Acute knee dislocation: An evidence based approach to the management of the multiligament injured knee. Injury 2011;42:1198–204. 8. Levy BA, Dajani KA, Whelan DB, et al. Decision making in the multiligamentinjured knee: An evidence-based systematic review. Arthroscopy 2009;25 (4):430–8. 9. Merritt AL, Wahl CJ. Rationale and treatment of multiple-ligament injured knee: The Seattle perspective. Oper Tech Sports Med 2011;19:51–72. 10. DiFelice GS, Lissy M, Haynes P. When to arthroscopically repair the torn posterior cruciate ligament. Clin Orthop Relat Res 2012;470:861–8. 11. Richter M, Bosch U, Wippermann B, et al. Comparison of surgical repair or reconstruction of the cruciate ligaments versus nonsurgical treatment in patients with traumatic knee dislocations. The American Journal of Sports Medicine. Am J Sports Med 2002;30(5):718–27. 12. Twaddle BC, Bidwell TA, Chapman JR. Knee dislocation: Where are the lesions? A prospective evaluation of surgical findings in 63 cases. J Orthop Trauma 2003;17(3):198–202. 13. Mariani PP, Santoriello P, Iannone S, et al. Comparison of surgical treatments for knee dislocation. Am J Knee Surg 1999;12:214–21. 14. Owens BD, Neault M, Benson E, et al. Primary repair of knee dislocations: Results in 25 patients (28 knees) at a mean follow-up of four years. J Orthop Trauma. 2007;21(2):92–6. 15. R Development Core Team R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2008. 〈http://www.R-project.org〉 [ISBN: 3-900051-07-0]. 16. Geeslin AG, LaPrade RF. Outcomes of treatment of acute grade-III isolated and combined posterolateral knee injuries: A prospective case series and surgical technique. J Bone Joint Surg Am 2011;93(18):1672–83.
