This issue of Applied Imaging addresses the MR evaluation of the
painful knee. Whether the pain is chronic (e.g., osteoarthritis) or
acute (e.g., posttraumatic internal derangement), MRI is the
preferred imaging modality for further evaluation. MRI very
accurately assesses tears of the menisci as well as the cruciate
and collateral ligaments. MR is also very sensitive for the
presence of bone-marrow edema, which can be seen in the acute
setting ("bone bruise") or in the chronic setting from
William G. Bradley, Jr., MD, PhD, FACR
Although plain films (x-rays) have traditionally been the first
diagnostic imaging study performed in the evaluation of the painful
knee, today they are useful only for evaluating joint space
narrowing, alignment, and major trauma. Over the past 15 years,
magnetic resonance imaging (MRI) has become the premier, first-line
imaging study that should be performed in the evaluation of the
painful knee. This issue of Applied Imaging describes the MR
evaluation of internal derangement (tears of the menisci and the
cruciate and collateral ligaments), osteochondral abnormalities
(chondromalacia, osteoarthritis, and osteochondral defects),
synovial cysts, and bone bruises.
In order to evaluate internal derangement adequately, the knee
must be scanned in the axial, coronal, and sagittal planes using
thin sections (3 to 5 mm thick) with a combination of T
- and T
-weighted techniques, including at least one T
-weighted image with fat suppression.
The anterior and posterior horns of the medial and lateral
menisci appear as wedge-shaped, low-intensity structures pointing
toward each other on sagittal and coronal views.
High signal within the meniscus, extending to the superior or
inferior articular surface (Figure 1), constitutes a meniscal tear.
This is best seen on images with narrowed "meniscal" windows
(Figure 1). A caveat: in the setting of prior surgery, it can be
difficult to distinguish an old scar from an acute tear without
injecting gadolinium into the joint space. Internal meniscal signal
not extending to the articular surfaces or extending only to the
base of the "wedge" does not constitute a tear. In children, this
represents the normal vascular pedicle of the meniscus; in adults
it represents myxoid degeneration, ie, normal aging.
Horizontal-oblique tears are considered results of myxoid
degeneration and are called "degenerative" tears. Those oriented in
the vertical direction, involving both the superior and inferior
articular surfaces, are considered "traumatic" tears.
Tears do not need to be seen in both the sagittal and coronal
planes, as they are only visualized when they are perpendicular to
the plane of section. In fact, tears running oblique to both the
coronal and sagittal planes may well be visualized only by a notch
on the articular surface. This "notch sign"
has served us very well over the last decade in diagnosing subtle
meniscal tears (Figure 2). Truncation of the tip of the meniscus
represents a "parrot beak" tear, ie, fraying of the free margin of
Occasionally, a large piece of meniscal fibrocartilage will
become detached from its capsular attachment and flip back-to-front
or centrally into the intercondylar notch. These "bucket-handle" or
are recognized on the basis of two findings: an absence of the
normal meniscus in its expected position and abnormal
meniscal-intensity material in an unexpected location (Figure 3).
Bucket-handle tears flipping from posterior to anterior appear as
two arrowheads pointing posteriorly in the position of the anterior
horn of the meniscus, with nothing at all in the expected position
of the posterior horn. This pattern is more common for
bucket-handle tears of the lateral meniscus. Tears flipping from
the periphery into the intracondylar region of the knee appear as
an extra structure adjacent to the cruciate ligaments, with an
absence of meniscal signal in the expected position of the
posterior horn. Typically, this pattern is seen with bucket-handle
tears of the medial meniscus. When the displaced meniscal fragment
lies parallel to the posterior cruciate ligament, it may produce
the "double PCL sign."
The anterior cruciate ligament may be seen on sagittal images
(Figure 4); however, it is always seen as an upside -down "V" on
coronal images (Figure 5). It is most commonly injured with valgus
stress to the knee while the leg is in external rotation.
MR studies performed shortly after the injury will fail to
demonstrate the separate fascicles of the ACL, which are replaced
by an ill-defined mass of tissue that is bright on T
-weighted images (reflecting edema). MR imaging performed much
later in a chronically "ACL-deficient" knee also fails to
demonstrate the well-defined fibers of the ACL, but lacks the T
hyperintensity of an acute tear.
With the medial knee joint distraction resulting from valgus
stress, the medial collateral ligament (MCL) as well as the medial
meniscus may also be torn (known as O'Donohue's triad). Acute
injuries of the MCL are best evaluated in the coronal plane
demonstrating discontinuity of the normally black line (Figure 5)
running along the medial aspect of the knee. With acute injuries,
there is also a linear fluid collection (which is bright on a T
-weighted image) superficial to the torn MCL. This is usually
associated with reticulation of the overlying subcutaneous tissues,
indicating soft-tissue swelling. Chronic MCL tears result in
thickening without associated fluid.
In addition to the medial distraction resulting from valgus
stress during an ACL injury, there is lateral impaction. This may
lead to bone marrow contusions of both the femoral and tibial
surfaces of the lateral compartment (Figure 6). At the time of the
MR examination, the leg is in a neutral position (rather than being
externally rotated as it was at the time of ACL injury); therefore,
the "kissing contusions" are not directly opposite each other.
Bone marrow contusions (also called "bone bruises" and
"trabecular fractures") cannot be seen on plain films and can be a
major source of pain.
They should be suspected in the patient with the acutely injured
knee without a meniscal or ligamentous injury in the absence of
plain-film findings of fracture. Because their water content is
higher than that of adjacent fatty bone marrow, they are best
detected on fat-suppressed, T
At high field, we use fat-saturated T
-weighted fast spin echo images and at any field strength, STIR
In one study of 73 consecutive patients with suspected internal
derangement of the knee, fat-saturated T
-weighted fast spin echo imaging detected one or more bone bruises
in 30% of patients in whom lesions were not seen on conventional
In evaluating bone marrow contusions it is important to
determine if they extend to the articular surface, as this is
usually an indication to rest the athlete for 30 days. Continued
activity in this setting could convert a bone bruise with intact
cartilage (type I) (Figure 6) to one in which the cartilage becomes
disrupted (type II) (Figure 7A), leading to early osteoarthritis.
Bone bruises can be seen whenever abnormal forces are applied to
normal tissue or essentially normal forces are applied to weakened
tissue, eg, osteoporotic bones. Thus, milder versions of the same
forces that lead to tibial plateau fractures will lead to bone
marrow contusions, resulting in normal plain films and high signal
on STIR or fat-saturated T
-weighted fast spin echo images.
An interesting pattern of bone marrow contusion can be seen with
traumatic dislocation of the patella. After the patella dislocates
laterally, the quadriceps contracts, jamming the medial patellar
facet against the lateral femoral condyle, resulting in bone marrow
edema on both surfaces (Figure 7B). Usually, this is also
associated with disruption of the black line of the medial
retinaculum, which, acutely, may have associated bleeding or edema.
Bone marrow contusions or fractures associated with cortical
disruption can leak marrow into the joint space, leading to a
lipohemarthrosis. Since sagittal images are acquired with the
patient supine, fat rises to the top, joint fluid is in the middle,
and intact red cells are at the bottom (Figure 8).
Lesions of the posterior cruciate ligament and the lateral
collateral ligament are much less common than their anterior
counterparts. The PCL is recognized as a dark C-shaped structure on
sagittal images and is visualized as a dark rounded structure seen
en face on coronal images (Figure 5).
The coronal images are best for visualizing partial PCL tears
(which show increased signal in a normally black structure), while
the sagittal view may better demonstrate complete disruption.
The earliest of the osteochondral abnormalities is simple
cartilaginous thinning as seen in chondromalacia patellae
and osteoarthritis. Although cartilage can be evaluated on our
routine sequences, we tend to use a higher resolution (512 * 512
matrix) thin-slice (1.5 mm sections) gradient echo technique when
subtle abnormalities of cartilage are being sought.
With advancing osteoarthritis, the articular cartilage is
completely denuded and subchrondral sclerosis can be seen as low
intensity on all sequences. With continued bone-on-bone irritation,
fibrovascular reaction can lead to bone-marrow edema, which appears
bright on STIR or fat-saturated T
-weighted fast spin echo images.
Osteochondral defects arise from varying degrees of traumatic
and ischemic insults to the articular surface of the knee (Figure
9). Early osteochondral defects appear as localized areas of
rounded bone marrow edema without frank fluid collection.
With advancing disease, a ring of frank fluid can be seen
surrounding the osteochondral defect, although the overlying
articular cartilage remains intact. Subsequently, the articular
cartilage also becomes disrupted and, eventually, the fragment
separates, becoming "free" within the joint space, leaving a divot
Cystic Disease of the Knee
Synovial cysts are very common in the knee given the large
number of synovium-lined surfaces. The most common is a Baker's
cyst (Figure 10), which generally results from a knee effusion that
forces synovium through the space between the tendons of the medial
head of the gastrocnemius and the semimembranosus.
Baker's cysts may persist following resolution of the knee effusion
and are a cause of medial joint line tenderness.
Another, less common, cause of medial joint line tenderness is
which is inflammation of the synovium of the pes anserinus (the
confluence of tendons from the sartorius, gracilis, and
Synovial cysts can be seen in conjunction with torn menisci,
more commonly involving the lateral than the medial meniscus. These
cysts appear to result from a ball valve mechanism that allows
fluid to flow from the torn meniscus into the cyst but not in the
MRI is now the preferred imaging technique for the evaluation of
the painful knee. Not only can it detect the fractures that are
missed on plain films, but it can also detect soft-tissue
abnormalities (meniscal and cruciate/collateral ligament tears)
that cannot be detected by plain film. These findings are critical
for the therapeutic decisions to be made by the orthopedic
Clinical Quiz: True or False
1. Areas of high signal intensity within the menisci on MRI
scans of the knee represent meniscal tears.
2. A plain film of the knee is a pretty good way to screen for
3. MRI is recommended for all patients with knee pain.
4. MRI is good for diagnosing tears of the anterior cruciate
5. MRI can differentiate among different causes of medial joint
line tenderness in the knee.
1. False. Areas of high signal generally represent myxoid
degeneration; in order to diagnose a tear, the high signal must
extend to the superior or inferior articular surfaces.
2. False. Frank cortical disruptions, which are frequently
missed on plain films, can be detected easily using MRI due to its
thin slice, tomographic capability. More importantly, bone bruises
(which can be a major source of pain) cannot be diagnosed at all by
plain films and are seen easily with fat-suppressed T
-weighted MRI studies.
3. False. Patients with cardiac pacemakers, ferromagnetic
intracranial aneurysm clips, and neurostimulators cannot have an MR
study, even of the knee.
4. True. The normal ACL can be visualized on all coronal images
of the knee and on most sagittal images. If it is not seen, it is
torn or deficient. Acute tears are associated with abnormal fluid
collections in the intercondylar notch, often with "kissing
contusions" of the anterior lateral tibial plateau and the
posterior lateral femoral condyle.
5. True. MRI can distinguish medial meniscal tears from medial
bone marrow contusions (bone bruises) from Baker's cysts (ruptured
or not) from anserine bursitis.