| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, University of Wisconsin School of Medicine and Public
Health, E3/311, 600 Highland Ave., Madison, WI 53792.
2 Department of Orthopedics and Rehabilitation Medicine, University of Wisconsin
School of Medicine and Public Health, Madison, WI.
3 Department of Statistics, University of Wisconsin School of Medicine and
Public Health, Madison, WI.
Received May 27, 2008;
accepted after revision August 7, 2008.
Address correspondence to A. A. De Smet.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. From a series of 559 knee MR examinations with arthroscopic correlation, we selected all 16 proven tears isolated to the posterior root of the lateral meniscus for retrospective blinded review, along with 45 cases of arthroscopically intact lateral meniscal posterior roots. The reviewers categorized whether there was a torn, possibly torn, or intact root based on three specific coronal and three specific sagittal image locations.
RESULTS. When all possibly torn roots were considered as torn, the sensitivity and specificity for diagnosis of a root tear were 93% and 89%, respectively. The observers' overall diagnosis of a tear based on all images gave a higher combined sensitivity and specificity than if the diagnosis of a tear had been based on one or any combination of the three coronal and three sagittal locations. Root tears were significantly more common in the presence of an ACL tear (p < 0.0001), but the presence or absence of an ACL tear did not change MR diagnostic accuracy.
CONCLUSION. The standard MR criteria of meniscal distortion and signal to the surface can be used to diagnose lateral meniscal root tears. The presence or absence of an ACL tear did not change diagnostic accuracy.
Keywords: accuracy • meniscus anatomy • meniscus tear • MRI
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Recently, there have been reports as well on the MR appearance of tears of the posterior root of the lateral meniscus. These studies described the association of root tears of the menisci with meniscal extrusion and anterior cruciate ligament (ACL) tears and provided cadaveric correlation of the medial and lateral meniscal roots with their MR appearance [4, 5]. The posterior root of the lateral meniscus has a unique MR appearance in that it does not have the isosceles triangular configuration seen in other regions of the meniscus [5].
Although there has been recent research on the root of the lateral meniscus, to our knowledge, no studies have reported on the ability of MR to diagnose a tear of the posterior root of the lateral meniscus. Accurate preoperative MR diagnosis of a posterior root tear of the lateral meniscus is clinically important because this area is difficult to visualize at arthroscopy, and identification of a root tear may require special arthroscopic portals [4].
We undertook this study of the MR appearance of these tears because we noted that with its different shape, root tears of the lateral meniscus were subjectively more difficult to diagnose. We retrospectively evaluated MR examinations of patients with arthroscopically proven torn and intact posterior roots of the lateral meniscus to determine whether we could diagnose root tears using the standard diagnostic criteria of meniscal distortion and meniscal signal to the surface. We evaluated the root at three specific locations on sagittal and coronal images to determine whether the sensitivity and specificity of tear diagnosis differed in these locations. We also studied the relationship of these tears and their MR diagnosis to the presence or absence of an associated ACL tear.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We reviewed a database of 559 consecutive knee MR examinations performed between October 2001 and June 2004 in patients who had not had prior meniscal surgery and who subsequently underwent correlative knee arthroscopy. This database lists the meniscal tears in all 559 patients classified by the arthroscopic location along with other arthroscopic findings, such as the presence of ACL tears. We identified 16 tears that were localized to the root of the lateral meniscus. We did not include in this study 24 tears of the root that extended into the posterior horn because these tears have typical MR findings of a posterior horn tear on images lateral to the root. We compared the frequency of root tears in patients with and without ACL tears to confirm the previously noted association of root tears with ACL tears.
MRI Parameters
Each of the 559 original MR examinations was performed using the same
protocol on a 1.5-T magnet (Signa, GE Healthcare) using a 4-channel knee
phased-array coil and fast spin-echo imaging. A field of view of 14 cm, slice
thickness of 3 mm with a 1.5-mm interslice gap, bandwidth of 20 kHz, and
matrix of 256 x 192 were used for all four sequences except for a matrix
of 256 x 224 for the coronal T1-weighted images. The parameters for the
coronal T1-weighted images were TR range/TE, 600–700/17, 1 signal
average, and echo-train length of 3. The parameters for the coronal
fat-saturated proton density–weighted images were 1,800–2,000/17,
1 signal average, and echo-train length of 4. The parameters for the sagittal
proton density–weighted images were 2,000–2,200/17, 1 signal
average, and echo-train length of 4. The parameters for the sagittal
T2-weighted fat-saturated sequence were 3,000–3,400/60, 1 signal
average, and echo-train length of 6.
Arthroscopic Technique
Each knee arthroscopy was performed with a consistent protocol followed by
three sports medicine fellowship-trained academic orthopedic surgeons who have
from 5 to 32 years of experience with knee arthroscopy. In all knees, the
meniscus was visualized from the standard anterior portals and probed to
expose under and upper surface tears and to evaluate the mobility of the
meniscus. When visualization proved difficult, or especially when MRI
indicated a posterior meniscal or root tear, the posterior compartment was
inspected with the arthroscope placed through the contralateral portal and
passed through the intercondylar notch to look down on the meniscal root. The
surgeons were aware of the MR findings in each case, and the MRI findings were
often reviewed in the operating room immediately before or during the
case.
Case Selection
In addition to the patients with proven root tears, we also selected for
further study from the database of 559 patients, the MR examinations of 45
patients who had intact lateral menisci at arthroscopy—23 of these
patients with an ACL tear and 22 with an intact ACL. These latter 45 studies
were selected using consecutive medical record numbers to minimize case
selection bias. For this case selection, all patients with an intact lateral
meniscus were divided into two groups on the basis of the presence or absence
of an ACL tear. Each group was then sorted by the medical record numbers.
Starting with the lowest medical record number, the first 23 patients with an
ACL tear and the first 22 with an intact ACL were selected for the study. We
included patients with and without ACL tears because we had observed
clinically and it has been reported that lateral root tears are often
associated with an ACL tear
[4]. Because of this observed
association, we wished to determine whether the hemorrhage and soft-tissue
change associated with an ACL tear might cause more false-positive diagnoses
in patients with an intact root of the lateral meniscus.
Retrospective MR Review
These MR examinations were then read in consensus by two fellowship-trained
academic musculoskeletal radiologists with 5 and 6 years of experience in
interpreting knee MR examinations. The radiologists evaluated the root of the
lateral meniscus in each of the 61 cases, which were blinded as to both the
original MR interpretation and the arthroscopic findings. They assessed the
root on the coronal and sagittal images in three locations: on the lateral
slope of the tibial eminence, at the level of the lateral intercondylar
tubercle, and between the lateral and medial intercondylar tubercles, which
are also called the tibial spines. These three areas of the root were selected
for evaluation on the basis of a review of the literature
[4,
5] and a review before the
study of MR examinations with arthroscopically proven torn and intact
roots.
Along the lateral border of the tibial eminence, the intact root appeared as a gently curving low-signal-intensity rectangle on the coronal image (Fig. 1A) and a discrete low-signal-intensity triangle on the sagittal image (Fig. 1B). When the root was adjacent to the lateral intercondylar tibial tubercle, the intact root appeared as a low-signal-intensity crescent on the coronal image (Fig. 1C) and a slightly flattened, elongated low-signal-intensity triangle on the sagittal image (Fig. 1D). Finally, when the root was between the intercondylar tubercles, it appeared as a linear low-signal-intensity structure on the coronal image (Fig. 1E) and a thin low-signal-intensity triangle on the sagittal image (Fig. 1F).
|
|
|
|
|
|
Review of Arthroscopic Diagnoses
The arthroscopic reports, hand-drawn maps of the meniscal pathology, and
arthroscopic photographs of the menisci were reviewed by the orthopedic
surgeon to confirm that the original database assignment of the presence or
absence of a root tear was correct. Four menisci with an original operative
diagnosis of a root tear but with an operative report description of only
fraying and photographs confirming the absence of a tear were considered to be
intact for this study.
Statistical Analysis
The difference in the frequency of a root tear noted at arthroscopy in
patients with and without an ACL tear was tested using a z-test for
subgroup analysis. For the z-test, z is calculated as the
difference in sensitivities divided by the square root of the sum of the
standard error (SE) squared for each sensitivity. The p value for the
two-sided z-test is obtained assuming that the z-value
follows standard normal distribution.
Sensitivity, specificity, and SE of the mean were calculated for the overall diagnosis of a root tear and the diagnosis of a tear at each of the three locations for both coronal and sagittal images. The kappa statistic was used to compare the value of each of the three locations in the two imaging planes, both individually and in multiple combinations, to determine the agreement between one or a combination of the images and the overall diagnosis.
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
When a possible tear was considered a definite tear for the overall diagnosis, the sensitivity of detection of a root tear was 94% ± 3% SE (15/16) and specificity was 89% ± 4% (40/45). If a possible tear was considered not a tear, the sensitivity decreased to 75% ± 6% (12/16) with only a slight increase in specificity to 93% ± 3% (42/45). All four menisci with fraying and without a discrete tear at arthroscopy were interpreted as intact on MRI.
When a possible tear was considered a definite tear for patients with and without ACL tears, there was a significant difference in sensitivity for detection of a root tear, with a sensitivity of 100% ± 0% (13/13) with an ACL tear and 67% ± 9% (2/3) without an ACL tear (p = 0). However, the specificity was not significantly different at 83% ± 6% (19/23) and 96% ± 4% (21/22), respectively (p = 0.15).
The sensitivity and specificity of a diagnosis of a tear on the basis of the appearance of the root at each of the six locations and all combinations of locations were compared when a possible diagnosis was considered a tear and when it was not considered a tear. Similar to the overall consensus reading, the sensitivity was higher when a possible tear was considered a tear, with only a slight reduction in specificity, compared with when a possible tear was not considered a tear. For this reason, only the data considering a possible tear as a tear are given in the analysis.
Analysis of Root Appearance at Individual Locations
Analysis of interpretation of the presence of a tear on the six individual
locations revealed that the highest sensitivity for diagnosis of a root tear
was found by assessing the coronal and sagittal images on the lateral margin
of the tibial eminence (88% ± 4% and 87% ± 4%, respectively)
with a specificity of 73% ± 6% and 89% ± 4%, respectively. The
sensitivity for each of the other four locations ranged from 29% to 50% with
specificity ranging from 68% to 89%. The agreement between the overall
diagnosis and each individual location was 0.72 and 0.58, respectively, for
the sagittal and coronal images on the lateral margin of the tibial eminence.
The kappa values for the other four locations were less than 0.5.
Analysis of the six locations alone and combinations of the six locations revealed that no single location or combination of locations was better than the readers' overall consensus. Of the multiple possible combinations, diagnosis of a tear on the sagittal images lateral to the tibial eminence alone or in combination with diagnosis of a tear on the sagittal or coronal images of the root between the intercondylar tubercles gave a sensitivity ranging from 87% to 88% and a specificity ranging from 80% to 89%. The kappa values comparing these combinations with the overall diagnosis ranged from 0.68 to 0.74. A sensitivity of 100% was noted with a tear diagnosed on one or both of the coronal and sagittal images of the root lateral to the tibial eminence, but specificity of diagnosis of a root tear decreased to 67%.
|
|
|
MR Appearance of the Torn Root
Tears of the root were manifested by either distortion (Fig.
2A,
2B,
2C) of the root or
intrameniscal signal contacting the root (Fig.
3A,
3B,
3C). In two patients, the root
was given an overall diagnosis of a tear, and the root was considered as
having a tear on all three sagittal and all three coronal images but the root
was diagnosed as intact at arthroscopy (Fig.
4A,
4B). Review of the operative
notes and map of the intraarticular findings and operative photographs in both
patients confirmed the original operative diagnosis of an intact root.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our sensitivity of 94% was higher than and our specificity of 89% was similar to the results of the study by Oei et al. [1] in which the authors found a mean pooled sensitivity of 80% and specificity of 90% for diagnosing lateral meniscal tears in all locations. Our improved sensitivity compared with the overall reported sensitivity for MR diagnosis of lateral meniscal tears may be due to two factors. First, the reviewers paid particular attention to the root, which may have increased their sensitivity to subtle abnormalities. Second, for analysis, we considered as intact the lateral meniscus root in four patients in whom fraying was found at arthroscopy based on retrospective review of the arthroscopic photographs and descriptions of the surgery. If we had used the original operative diagnosis of a root tear and not considered these as frayed, our sensitivity would have been lower because all four roots were considered intact on MRI. Slight fraying of a meniscus is usually not detectable on MRI, and it is controversial whether such fraying may be a cause for knee pain [6].
We found no surgical series with which to compare our frequency of root tears because in large surgical series analyzing the frequency of lateral meniscal tears, tears isolated to the posterior root have not been listed separately from other posterior horn tears [7–9]. We found that tears isolated to the posterior root of the lateral meniscus were uncommon in our patients because such tears were identified at arthroscopy in only 16 (2.9%) of 559 knee arthroscopies in our series. This frequency is lower than reported in a recent study [10]. However, in that study, root tears with extension into the posterior horn were included as root tears (review of prior study data by current study lead author). We excluded root tears with extension into the posterior horn because the purpose of our study was to evaluate MR diagnosis for tears localized to the root.
As suspected on the basis of our clinical experience and as previously reported [4], we found that root tears of the lateral meniscus are highly associated with ACL tears because they were present in 8% of patients with ACL tears but in only 0.8% of patients without ACL tears. The incidence of lateral root tears in our study is similar to the 9.8% found in a study of 264 patients with ACL tears [4]. In that study, the diagnosis of a root tear was based on the MR appearance because only four of the 26 root tears diagnosed on MRI were confirmed at arthroscopy.
We speculated that the presence of an ACL tear might reduce our accuracy in correctly identifying an intact root if the hemorrhage from the ligament tear caused loss of definition of the root on MRI. However, we found that the specificity was not significantly lower in the presence of an ACL tear. In fact, the sensitivity was 100% (13/13) in the presence of an ACL tear but statistically significantly lower at 67% (2/3) without a tear. However, by statistical convention, when one test achieves perfect sensitivity and the other does not, the p value for the difference is 0.0. With only three root tears in patients without an ACL tear, we do not think the lower sensitivity in the absence of an ACL tear is clinically relevant.
We evaluated the root in three locations: between the intercondylar tubercles, at the level of the lateral tubercle, and on the lateral edge of the tibial eminence adjacent to the lateral tubercle. We selected these locations because they provided the best MR assessment of the root in a review of root anatomy before beginning this study. In a study of several cadavers, Brody et al. [5] found that the posterior root of the lateral meniscus inserted from the posterior slope of the lateral intercondylar tubercle to the crest between the intercondylar tubercles and correlated the anatomic appearance with the MR appearance. In their anatomic study of 92 cadaver knees, Kohn and Moreno [11] found that the root attached for a variable length ranging from 32 to 263 mm with a mean of 115 mm. They similarly found that the attachment extended medially from the crest between the intercondylar tubercles to the back of the posterior slope of the lateral tubercle and provided a diagram showing slight extension of the root onto the upper edge of the lateral slope of the tibial eminence.
We found that for MR diagnosis of a tear, the lateral meniscal root needs to be assessed on both the coronal and sagittal images at each of the three locations. For individual location evaluation, our retrospective observers had their highest sensitivity (100%) by detecting an abnormality in the root on the lateral slope of the tibial eminence on one or both of the coronal or sagittal images of this area. However, their specificity dropped to 67% using an abnormality in this location as the criterion for a root tear. The reviewers' overall subjective diagnosis based on all six areas of evaluation of the root gave the highest combination of sensitivity and specificity (94% and 89%, respectively).
Another finding of our study was that if we considered an overall diagnosis of a possible tear or diagnosis of a possible tear at a specific location as indicating a definite tear, the sensitivity was considerably increased, with only a slight decrease in specificity compared with considering all possible tears as an intact meniscus. Thus, when our retrospective reviewers suspected there was a root tear of the lateral meniscus, the root was usually found to be torn at arthroscopy.
Our criteria for the diagnosis of a root tear were the same as used in a previous MR study of four root tears confirmed by arthroscopy [4]. As in that study, we diagnosed a lateral meniscal root tear when there was distortion or increased signal intensity in the root that extended to the surface of the meniscus [4]. No other study has specifically addressed the appearance of a torn lateral meniscal root. In a pictorial review of radial tears of the menisci, the authors indicated that, in their opinion, coronal images were the most important in diagnosing lateral root tears [12]. They noted that the central portion of the posterior horn of the lateral meniscus should cover the most medial part of posterior tibial plateau on at least one image [12].
The posterior root of the lateral meniscus can be difficult to assess on MRI for multiple reasons including pulsation artifact from the popliteal artery, volume averaging and magic angle effect because of the slope of the meniscus on the tibial eminence, and the complex anatomy related to the origin of the meniscofemoral ligaments [10, 12]. Despite these possible causes for diagnostic errors, we achieved both a high sensitivity and specificity for the MRI diagnosis of a root tear.
A limitation of our study is the small number of isolated tears of the lateral meniscal root. Because these are not common tears, we identified only 16 such tears in a review of more than 500 patients who underwent an MR examination and correlative knee arthroscopy. Perhaps with a larger series, our findings on the sensitivity and specificity for MR diagnosis of a root tear might be different from what we achieved; however, we expect that our findings would be similar.
A second limitation is the retrospective use of knee arthroscopy as the reference standard. Although knee arthroscopy is currently the best available reference standard, there are variations in arthroscopic terminology and arthroscopists' visualization of the root that may limit the use of this technique as a reference standard. In a recent study of assessment of meniscal tear type and location using a video of knee arthroscopies, the kappa values for interobserver variation for defining the type and location of a meniscal tear ranged from 0.3 to 0.63 [13]. A prospective study in which each tear and each apparent MR abnormality is carefully identified and photographed at arthroscopy would be preferable to a retrospective study.
A final limitation is our assumption that an MR abnormality on the lateral slope of the tibial eminence indicated a root tear. The root of the lateral meniscus extends anatomically from between the intercondylar tubercles to the posterior slope of the lateral intercondylar tubercle [11]. The lateral edge of the tibial eminence was considered as part of the root because, with volume averaging on MRI and variation in image location relative to the lateral tubercle, an image including the lateral slope of the tibial eminence did include part of the root. Thus, in some cases, the observers may have been assessing the most central aspect of the posterior horn and not the root. However, because all tears were isolated to the root on the basis on arthroscopic localization, an MR abnormality noted on the lateral slope of the tibial eminence should indicate root abnormality.
In summary, we found that the standard MR criteria for a meniscal tear can be used to diagnose a tear of the posterior lateral meniscal root with high sensitivity and specificity. Overall subjective diagnosis of a tear using all images had higher sensitivity and specificity than diagnosis of a root tear on the basis of a specific MR image. Root tears are significantly more common when there is an ACL tear, but diagnosis of a root tear is not affected if there is an ACL tear.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
