| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Thomas Jefferson University Hospital, Philadelphia,
PA 19066.
2 Present address: Department of Radiology, University Hospital Galway,
Newcastle Rd., Galway, Ireland.
Received October 17, 2007;
accepted after revision January 14, 2008.
Address correspondence to D. Bergin
(dianebergin{at}yahoo.com).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Seventy patients underwent 1.5-T MRI of the knee and arthroscopy within 3 months. MR images were reviewed by two radiologists for parameniscal cysts, bowing and edema around the collateral ligament, meniscal extrusion, cartilage loss, and nonlinear and linear subchondral marrow edema. The findings were correlated with the arthroscopic findings. The positive predictive value (PPV), sensitivity, and specificity of indirect signs were calculated.
RESULTS. Fifty-three medial and 28 lateral meniscal tears were found at arthroscopy. The PPV of indirect signs of meniscal tear was 0.17–1.00 for reader 1 and 0.37–1.00 for reader 2. The PPV of parameniscal cysts was 1.00 for medial and lateral meniscal tears for both readers. The specificity and PPV of periligamentous edema and cruciate ligament bowing for medial meniscal tear were 0.94 and 0.98 for both readers, 0.88 and 0.94 for reader 1, and 0.94 and 0.97 for reader 2. The specificity and PPV of subchondral marrow edema for medial meniscal tear were 0.88 and 0.96 for reader 1 and 0.94 and 0.97 for reader 2 and for lateral meniscal tear were 0.98 and 0.92 for reader 1 and 1.00 and 1.00 for reader 2. The specificity and PPV of linear subchondral marrow edema for medial meniscal tear were 0.94 and 0.97 for reader 1 and 1.00 and 1.00 for reader 2. For lateral meniscal tear, the values were 0.98 and 0.89 for reader 1 and 1.00 and 1.00 for reader 2. The specificity and PPV of nonlinear subchondral marrow edema for medial meniscal tear were 0.94 and 0.89 for reader 1 and 1.00 and 1.00 for reader 2. For lateral meniscal tear, the values were 0.89 and 0.97 for reader 1 and 1.00 and 1.00 for reader 2. The specificity and PPV of cartilage loss for medial meniscal tear were 0.88 and 0.94 for reader 1 and 0.88 and 0.93 for reader 2. For lateral meniscal tear, the values were 0.85 and 0.56 for reader 1 and 0.97 and 0.80 for reader 2.
CONCLUSION. Indirect MRI signs occur in association with meniscal tears and can aid diagnostic confidence when the MRI meniscal appearance is equivocal.
Keywords: indirect MRI signs • meniscal tear • MRI
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The presence of a parameniscal cyst is a well-described secondary sign that has a highly specific association with the presence of a meniscal tear, even in the rare instances in which the tear is not visualized well or at all [9, 10]. However, parameniscal cysts are seen in only 7% of meniscal tears [9]. A number of additional signal intensity abnormalities in the adjacent soft tissues and bone have been anecdotally described in association with meniscal tears.
Using arthroscopy as a reference standard, we sought to determine the frequency of abnormal signal intensity in soft tissues and adjacent osseous structures in association with meniscal tears. Recognition of these signs may serve to increase reader confidence in the diagnosis of meniscal tear, particularly in equivocal cases. Understanding that these secondary signs occur in association with meniscal tears would avoid potential misinterpretation that the findings are attributable to alternative pathologic conditions.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MRI Protocol
All MRI examinations were performed without IV contrast enhancement and
with a 1.5-T unit (Signa, GE Healthcare). Images were obtained with a
dedicated knee coil. Standard sequences were sagittal proton-density spin-echo
(TR/TE, 1,200/20), sagittal T2-weighted fast spin-echo fatsaturated
(4,000/50), coronal T1-weighted spinecho (600/minimal), coronal T2-weighted
fast spin-echo fat-saturated (4,000/60), and axial T2-weighted fast spin-echo
fat-saturated (4,000/60). Other MRI parameters included 4-mm slice thickness,
1-mm gap, 256 x 192 matrix, and 15 cm2 field of view. Fast
spin-echo sequences were performed with an echo-train length of 8.
Imaging Analysis
Two experienced musculoskeletal radiologists blinded to the original MRI
examination reports, clinical information, and arthroscopic findings reviewed
MR images independently. Each reader assessed the medial and lateral
compartments of each knee on MR images at separate sittings. Each reader noted
the presence or absence of a meniscal tear, the presence of cartilage loss in
the ipsilateral compartment, whether cartilage loss when present was focal or
diffuse, grade of cartilage loss (grade 1, increased T2 signal in cartilage,
indicating chondral softening but with preservation of normal cartilage
contour; grade 2, partial thickness; grade 3, full thick ness), and the
presence of meniscal extrusion more than 3 mm beyond the cortical surface.
Ipsilateral parameniscal soft tissues including the ipsilateral collateral ligament, presence of parameniscal cyst, and parameniscal fat were evaluated. Presence or absence of bowing and peri ligamentous edema of the adjacent collateral ligament was noted. Presence or absence of parameniscal cyst was noted, and when a cyst was present, the largest dimension was recorded.
Adjacent tibial and femoral condyles were evaluated for subchondral marrow edema and linear subchondral marrow edema. It was also specifically noted whether subchondral marrow edema when present was directly adjacent to a meniscal tear. Subchondral marrow edema when present was noted to be linear or nonlinear relative to the articular surface. Linear marrow edema was defined as well-demarcated linear subchondral marrow edema parallel to the articular surface and less than 5 mm deep. Nonlinear subchondral marrow edema had no clearly defined margin. The presence of perivascular edema along the tibial or femoral vascular channels extending into the mar row from the medial and lateral tibial or femoral cortices was recorded. Presence of joint effusion and whether it was small, moderate, or large were recorded.
Arthroscopy Report Review
An independent observer reviewed arthroscopy reports. The presence or
absence of meniscal tears as noted by the orthopedist was recorded.
Statistical Analysis
The incidence of soft-tissue and osseous indirect signs on MRI was
recorded. The positive predictive value (PPV), sensitivity, and specificity of
each indirect MR sign for individual readers for both medial and lateral
menisci were determined. Interobserver agreement for presence of indirect
signs on MRI was determined by calculating the intraclass correlation (ICC)
according to methods developed by Shrout and Fleiss
[11]. Dichotomous variables
(using arthroscopy as the reference standard) were evaluated by Fisher's exact
test and odds ratios. We used Fisher's exact test to determine the
relationship of observed indirect signs on MRI by individual readers with
presence of medial and lateral meniscal tears at arthroscopy. A p
value of less than 0.05 was considered statistically significant. The odds
ratio, defined as the ratio of the odds of an event (indirect sign on MR
observed by combined readers) to the odds of an event in another group
(meniscal tear at arthroscopy), was also calculated because it makes for
concise presentation of the results.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
On MRI, parameniscal cysts (Fig. 1) related to the medial meniscus were identified in seven knees by reader 1 and eight knees by reader 2, for a PPV, sensitivity, and specificity of 1.00 for both readers. Lateral parameniscal cysts were identified in four knees by reader 1 and three knees by reader 2, for a PPV and specificity of 1.00 for both readers. Interobserver agreement for medial and lateral parameniscal cysts was 0.90 and 0.90.The odds ratio for the readers combined between observance of parameniscal cyst on MRI and meniscal tear at surgery was 54 (p < 0.0001) for medial and 56(p < 0.0001) for lateral menisci.
|
|
|
Subchondral marrow edema (Figs. 2 and 3) was seen adjacent to the medial meniscus in 46 MRI examinations by reader 1 and 41 examinations by reader 2. Subchondral marrow edema adjacent to the lateral meniscus was seen in 13 MRI examinations by reader 1 and 12 MRI examinations by reader 2. Linear subchondral marrow edema (Fig. 4) adjacent to the medial meniscus was identified in 38 MRI examinations by reader 1 (p < 0.0001) and 35 examinations by reader 2 (p < 0.0001). Linear subchondral marrow edema adjacent to the lateral meniscus was identified in nine MRI examinations by reader 1 (p < 0.0001) and eight examinations by reader 2 (p < 0.0001). Perivascular marrow edema adjacent to the medial meniscus (Fig. 5A, 5B) was identified in nine MRI examinations by reader 1 (p = 0.09) and 15 examinations by reader 2 (p = 0.013). Peri vascular marrow edema adjacent to the lateral meniscus was identified in five MRI examinations by reader 1 (p = 0.21) and 2 examinations by reader 2 (p = 0.07).
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Established indirect MRI signs of meniscal tears described in the literature [9, 10, 12] include meniscal and parameniscal cysts, joint effusion, and medial collateral ligament edema. Meniscal and parameniscal cysts are believed to occur as a result of leakage of fluid due to meniscal tear [9]. In our study population, the sensitivity of parameniscal cyst was low, for readers 1 and 2, 0.06 and 0.11 for medial tears and 0.14 and 0.32 for lateral tears, but with a specificity of 100% for all meniscal tears. This indirect MRI sign has a reported [9] PPV of 100% for the presence of medial and lateral meniscal tear at arthroscopy, which corresponds with our findings.
Adjacent collateral ligament edema may be seen as a result of primary ligamentous injury [12]. Medial collateral ligament edema also has been described as an abnormality in patients with medial compartment osteoarthritis [13]. Medial collateral ligament edema in this setting occurs as a result of chronic friction and shearing forces between the ligament and adjacent osteophytes and as a result of meniscal extrusion [13]. In osteoarthritis, meniscal extrusion is facilitated by laxity and microtears of the meniscofemoral and meniscotibial attachments that generate local softtissue edema [13]. To our knowledge, edema of the adjacent collateral ligament associated with meniscal tears has not been described in the literature. We found nearly 100% specificity for the presence of adjacent collateral ligament edema on MRI in relation to the presence of meniscal tear at arthroscopy. Of interest was that the sensitivity was higher for medial meniscal tears than for lateral meniscal tears, reflecting the closer apposition of the peripheral margin of the medial meniscus to the medial collateral ligament than of the lateral meniscus and lateral collateral ligament. Patients with grade 4 chondromalacia were excluded from our patient population, reducing the potential of coexistent osteophytes, which can cause collateral ligament edema or collateral ligament bowing.
The presence of parameniscal soft-tissue edema likely reflects a combination of factors, including local increase in fluid formation secondary to tear, increase in blood flow in the region secondary to inflammatory mediators, and reactive local synovitis secondary to meniscal tear. The situation is analogous to the edema seen around the popliteal meniscal fascicle in the setting of a lateral meniscal tear [14–18].This edema, unlike edema related to osteoarthritis, is probably secondary to reactive inflammation and hyperemia due to adjacent meniscal tear and coexistent meniscal extrusion causing secondary mass effect on the collateral ligament. Readers 1 and 2 had observations of bowing of the ipsilateral collateral ligament of 0.88 and 0.94 specificity for the presence of medial meniscal tear and 1.00 and 1.00 specificity for the presence of lateral meniscal tear.
Loss of hoop strength secondary to disruption of collagen fibers of the meniscus in the presence of a tear leads to extrusion of the meniscal substance beyond the tibial or femoral articular cortex of the knee joint [19]. Meniscal extrusion has been described in association with age-related degenerative changes, meniscal tears, and osteoarthritis [20, 21]. In our study, specificity of observation of meniscal extrusion by both readers was 0.94 for medial tears and 0.88 and 0.97 for lateral tears, with a sensitivity of 0.64 and 0.49 for medial and 0.03 and 0.14 for lateral tears. This finding reflects the larger size of the medial meniscus, higher intrinsic hoop strength, and tighter appositionof the medial meniscus to the ipsilateral tibial margin compared with the lateral meniscus. Loss of this hoop strength with meniscal tear causes greater relative extrusion of meniscal substance of the medial meniscus than of the lateral meniscus.
Periarticular marrow edema on MRI may be associated with cartilage loss or direct trauma [22, 23]. We evaluated three types of marrow edema in our population group. Perivascular edema, defined as not extending to the articular surface and primarily occurring around the vascular channels, extending from the tibial and femoral medial and lateral cortices, was relatively infrequently observed but had high specificity for the presence of meniscal tear. We believe this finding reflects increased blood flow in reaction to the adjacent primary soft-tissue injury, that is, the meniscal tear, but also reflects secondary increased blood flow in response to increased stress in the adjacent osseous trabeculae (e.g., collateral ligament and trabeculae) as a result of loss of the normal buttressing effect of the intact meniscus to weight-bearing force. To our knowledge, this report is the first description of perivascular marrow edema on MRI.
Subchondral marrow edema adjacent to the meniscal tear was identified as being linear or nonlinear. Both readers found linear subchondral marrow edema immediately adjacent and parallel to the articular surface, with medial meniscal tears in more than 60% and lateral meniscal tears in almost 90% of examinations. To our knowledge, this description is the first in the literature of linear subchondral marrow edema on MRI. This type of marrow edema was identified immediately adjacent to the meniscus and probably is reflective of hyperemia and increased fluid at the cortical–chondral–meniscal interface. Further study of this population will determine whether this edema resolves or persists depending on the evolution of the meniscal tear and associated cartilage abnormalities and subsequent response to treatment.
Nonlinear subchondral marrow edema is well described in the radiology literature. This finding is frequently associated with cartilage abnormalities on the adjacent articular surface. The sensitivity of this nonlinear type of subchondral marrow edema for both medial and lateral meniscal tears was moderate, but the specificity was high for both types of tear. The presence of this marrow edema probably reflects coincident cartilage loss, chronicity, and extent of meniscal tear.
The standard MRI knee protocol at our institution includes fluid-sensitive fat-saturated sequences in at least two planes and frequently in three planes. This technique improves the conspicuity of edema and fluid that generates many of the indirect signs of meniscal tears described. We concede that MRI knee protocols at other institutions may not include fat-saturated fluid-sensitive sequences in all planes. Nonetheless, these indirect MRI signs when detected, even in a single plane, may aid final diagnosis.
Interobserver agreement between readers for the presence of indirect MRI signs of meniscal tears in most cases was moderate to high, indicating the utility of identification of these signs. Interobserver agreement was poor for observation of cartilage loss near the lateral meniscus and of edema around the lateral collateral ligament. This poor interobserver agreement was primarily attributable to the small number of cases in which these indirect signs were present. In addition, the MRI protocol instituted was not optimized for detection of cartilage loss, inherently limiting the sensitivity for cartilage defects. Detection of cartilage defects, however, was not a primary aim of this study and therefore did not affect our overall conclusions.
This study was limited by the lack of a control group. The ideal study would include a series of MRI knee examinations of patients who do not have meniscal tears proved at arthroscopy. However, because one of the most common indications for arthroscopy is meniscal repair, it would be difficult to enroll such a control group. To reduce the effect of this potential weakness, the MRI examinations in this study were reviewed by each reader at two different sittings, one for the medial meniscus and one for the lateral meniscus. Because most of the patients in this study had only one meniscal tear in a knee, the unaffected meniscus and compartment served as an inherent modified control group for this study. It was also not possible to determine the temporal association between presence of the described indirect MRI signs and the onset of meniscal tear. Future study of this patient population, who may undergo follow-up MRI, would be useful to assess for resolution or progression of the parameniscal findings. The challenge would be that patients typically do not undergo repeated imaging unless they have persistent or new symptoms, creating inherent bias.
We describe MRI signal abnormalities that can occur in the parameniscal soft tissues and bone marrow in association with meniscal tears. Although these indirect signs vary in sensitivity and specificity for meniscal tear, we believe their greatest value may be in directing attention to the meniscus and increasing diagnostic confidence in cases in which the primary MRI meniscal appearance is equivocal for absence or presence of a tear. In the setting of evident meniscal tear on MRI, it is also important to recognize that these soft-tissue and osseous MRI findings are associated with primary meniscal tears and should not automatically be interpreted as additional primary abnormalities. The indirect signs described vary considerably in frequency and specificity with regard to meniscal tears. To our knowledge, this description is the first of linear and perivascular edema associated with ipsilateral meniscal tears. Linear subchondral marrow edema may be a particularly useful indirect sign in patients who have otherwise equivocal MRI findings of meniscal tear. These preliminary findings suggests that this MRI sign has high sensitivity and specificity for meniscal tear.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  B. J. Laundre, M. S. Collins, J. R. Bond, D. L. Dahm, M. J. Stuart, and J. N. Mandrekar MRI Accuracy for Tears of the Posterior Horn of the Lateral Meniscus in Patients With Acute Anterior Cruciate Ligament Injury and the Clinical Relevance of Missed Tears Am. J. Roentgenol., August 1, 2009; 193(2): 515 - 523. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
