MRI of the Knee MALTA MRI in Practice Collagen

Document Sample
MRI of the Knee MALTA MRI in Practice Collagen Powered By Docstoc
					MRI Clinical Applications II                                                            Joseph Castillo

6        MRI of the Knee.

This section addresses learning outcome 1, 2 and 3. The parts on normal and abnormal MRI anatomy of the
knee addresses learning outcome 1. Image optimization, artifacts and response strategy and critical
evaluation addresses learning outcome 2. Protocol change addresses learning outcome 3 and focuses on an
attempt by the candidate to modify the routine protocol. The section also identifies shortcomings in the
present protocol and suggests possible modifications. This part also highlights a possible research question
to assess the value of PD FSE Fat suppressed sagittal sequence with optimized TE and ETL.

Normal MRI Anatomy

The knee joint is made up of three individual joints: two condylotibial and patellofemoral.
Several structures are found in the knee joint and these are describes separately.

The menisci are the embryological remnants of the discs commonly found at condylar
joints where two separate movements can occur. They are crescenteric structures and
the lateral meniscus is more of a complete circle than the medial meniscus. For ease of
diagnostic interpretation the meniscus is divided into three parts (anterior, middle and
posterior thirds). The meniscus is further subdivided radially into the inner edge, middle
and peripheral regions. The menisci are best viewed in the sagittal and coronal planes.
In the sagittal plane, the normal menisci appear as homogenous low signal structure on
all sequence, and have a triangular structure with an outer convex border. The collagen
bundles within each meniscus form two distinct zones. Circumferential fibres are found
in the peripheral third of the meniscus, whereas the transverse collagen fibres connect
the circumferential zone to the meniscal free edge. In the adult, the peripheral 10-20%
of the meniscus receives a blood supply from the capillary plexus that surrounds it, but
the central portion of each meniscus is relatively avascular (Carrino & Schweitzer,

The cruciate ligaments are intracapsular but extrasynovial.                     The anterior cruciate
ligament (ACL) connects the posterior medial aspect of the lateral femoral condyle to
the anterior tibial intercondylar region. It is composed of two functional fibre bundles:
the longer antero medial bundle (AMB) tightens with knee flexion, and the shorter but
bulkier posterolateral bundle which tightens with knee extension. On sagittal images the
ACL is normally parallel to the intercondylar roof. Although an ACL is best seen on a

MRI Clinical Applications II                                                 Joseph Castillo

sagittal section, all three planes are useful to identify an intact ACL. On MRI there is
often high signal intensity interspersed within the fibers of the ACL.

The posterior cruciate ligament (PCL) connects the lateral aspect of the medial femoral
condyle to the posterior intercondylar fossa of the tibia, and passes medial to the ACL.
In relation to the PCL two other ligaments may pass anterior (ligament of Humphrey) or
more commonly posterior (Ligament of Wrisberg). In some both ligaments are present.
Because knee MRI is performed in extension the PCL normally has a sharp bend known
as the genu. The PCL is typically more uniformly hypointense than the ACL, although
there may be a slight increased signal intensity on short TE secondary to magic angle

On each side of the joint collateral ligaments strengthen the capsule.            The medial
collateral ligament extends from the medial condylar region and attaches 4-5cm inferior
to the tibial plateau and posterior to the pes anserinus insertion. Deep to this the medial
capsular ligament is composed of the meniscofemoral and meniscotibial attachments to
the meniscus. The ligament represents the main restraint against valgus strain.
The lateral collateral ligament (LCL) strengthens the capsule laterally, and lies
somewhat posteriorly. It is considered to be made up of two layers and lies deep to the
insertions of the distal iliotibial tract and biceps femoris anteriorly, and to the quadriceps
retinaculum anteriorly.        Deep to the LCL are the meniscofemoral and meniscotibial
attachments.      The intracapsular popliteus tendon passes medial to the LCL, and
posterior fibres of the LCL blend with the deep capsule, which contributes to the arcuate
popliteal ligament. Coronal images are more accurate than axial images for grading
injuries of the medial and lateral collateral ligaments.
The extensor mechanism is a structure that consists of the quadriceps tendon, the
patella, the patellar ligament, Hoffa’s fat pad, and the medial and lateral patellar
retinacula. The quadriceps tendon is made up from the rectus femoris, vastus lateral,
vastus mediali and vastus intermedius, four muscles and hence its name. The extensor
mechanism is best evaluated using sagittal and axial views.
Cartilage and bone marrow is best assessed in all three planes using fat suppressed
sequences and T1 Weighted images. Muscle signal is similar to other parts of the
musculoskeletal system (see Section 1) (ibid).

MRI Clinical Applications II                                                Joseph Castillo

Clinical Indications

         Meniscal degeneration and tears
         Anterior Cruciate Ligament disruption
         Posterior cruciate ligament disruption
         Collateral ligaments injury
         Chondromalacia patellae
         Extensor mechanism abnormalities - Patellar tendonitis, Patella alta and baja,
          Patella bursae
         Ganglion cyst, meniscal cyst, synovial cyst
         Fractures.

Case Study - Anterior Cruciate Ligament Tear

The ACL can be damaged by a variety of different mechanisms. Common to all forms
of injury is rotation of the femur on the tibia at the time of injury, commonly associated
with varus or valgus stress. For this reason, it is unusual for an ACL injury to exist in
isolation, and in particular meniscal tears are involved in up to 68% of acute injuries or in
up to 91% of chronic ACL deficient knees. Collateral ligament injury is also a common
finding but may predate the ACL injury (Shahabpour et al., 1997).
The ACL is best visualized on sagittal images, but is also seen in coronal and axial
The normal ACL is represented as one or more bands of low signal intensity, and the
separate bundles can be distinguished near the point of insertion due to interposition of
fat between them.        It may have variable thickness but is usually about 3 to 4mm.
Therefore the normal ACL is depicted in one sagittal image. Primary signs of an ACL
tear depend on the time elapsed since the injury. Changes in morphology as well as
signal intensity determine whether it is intact.        In a complete tear of the ACL
discontinuity of fibers and increased signal intensity on short and long TE is present.
Partial or incomplete disruptions may be associated with attenuation of some fascicles,
which may appear indistinct due to adjacent oedema and haemorrhage.               Secondary

MRI Clinical Applications II                                                Joseph Castillo

signs include bunched up PCL, a wavy patella ligament, and anterior translation of the
femur on the tibial condyles although none of these are specific (Irizarry & Recht, 1997).
However absence of ACL in the sagittal and coronal plane is diagnostic for ACL tear and
the accuracy is in excess of 95%. Bone bruising involving the posterolateral tibial
plateau (segond fracture) and lateral femoral condyle has been shown to be a relatively
specific sign of an acute complete ACL tear (ibid).

Image Optimization in MRI of the Knee

MRI of the knee is one of the most frequently requested exams in musculoskeletal
imaging. This is because of its inherent accuracy in depicting internal derangements
and allowing the orthopaedic surgeons to use this study as a road map for therapeutic
arthroscopy. Because of the various pathology, mainly due to trauma and osteoarthritic
degeneration, the most important consideration is a proper protocol. Proton Density
with or without fat suppression, T1-weighted SE, T2-weighted FSE, STIR and Gradient
echo pulse sequences are most commonly used in investigating the various structures
forming the knee joint and as such there is no standard protocol. Various sites use
different sequences but the choice of slice thickness, field of view, coil, and imaging
matrix play an important role in determining spatial resolution(Irizarry & Recht, 1997;
Rubin, 1997; Kaplan et al., 2001). Attention must be given to these technical details as
improved spatial resolution provides the radiologist with data to present interpretation
with confidence.

To date, using super conductive units, the study of the knee joint requires a dedicated
extremity coil, which is often transmit receive coils ensuring optimum and uniform signal
coverage. Most MR system manufacturers now offer dedicated knee coils as standard.
These coils use a cylindrical configuration, similar to the head coil, to provide a
homogenous imaging volume, and a quadrature design that provide improved signal to
noise ratio. Flexible surface coils are used as an alternative when the knee joint is too
large to fit in the rigid knee coil or when the patient is unable to extend the knee, usually
following trauma. Spatial resolution is necessary to image the small structures that are
found in the knee and the FOV should be kept small ideally not exceed the length of the
coil. The range of FOV in the supero-inferior length is of 160-200mm. Using a matrix of

MRI Clinical Applications II                                               Joseph Castillo

256, allows an in-plane resolution of less than 1mm. Thin slices in the range of 3-4mm
and gap of not more then 0.5mm ensure high spatial resolution (. 3D acquisition using
isotropic matrix is useful to provide high resolution visualization of anatomy in any plane.
The bandwidth is another important factor that may affect imaging especially at high
field strengths.     The signal-to-noise ratio can also be improved by using a narrow
receive bandwidth. One disadvantage of narrow bandwidth is the increase in chemical
shift artifact and spatial misregistration are increased at the interfaces between cortical
bone, hyaline cartilage and fibrocartilage. This effect may adversely affect diagnostic
accuracy (Fitzgerald, 1994) as these artifacts can simulate or obscure thinning of the
cortex or overlying hyaline cartilage. Fat-suppression techniques eliminate chemical
shift misregistration and so narrow bandwidth can be used.
In addition to these technical considerations, magnetic field strength is another
important factor, which is not under the practitioners’ control. Diagnostic quality in knee
imaging has been researched and reported with field strengths ranging from 0.2 to 1.5T
(Vellet et al.,1995, Cotten et al., 2000, Maubon et al., 1999). The levels of confidence in
reporting was better in high field strength than low field strength (Rand et al, 1999) and
in general, lower field strengths units must use more signal averages to provide
equivalent quality as that obtained with high field strengths. This results in increased
scan times.

Artifacts and response strategy

The main source of artefact in knee imaging is from the popliteal vessel pulsation and
patient movement. Pre-saturation pulses placed S and I to the FOV eliminate this
problem. Phase-ghosting in sagittal view is eliminated by putting the phase encoding
direction S to I. However, oversampling is necessary to eliminate aliasing from the thigh
and lower leg. This phase encoding direction (S-I) could create truncation artifacts,
which could mimic tears. This is reduced by using high matrix above 192 (Turner et al.,
1991). Applying immobilization pads, putting a Velcro strap over the thigh, reduces
patient movement.

Scan Plane

MRI Clinical Applications II                                                  Joseph Castillo

Despite the advances and the development of new sequences, the planes used to
assess the “routine knee” have changed minimally over the years. Optimal evaluation of
the knee joint includes sagittal, coronal and axial planes. The axial planes are obtained
from the superior surface of the patella to the tibial tuberosity using a sagittal localizer.
This view allows depiction of the extensor tendon mechanism, including the
patellofemoral joint. It also shows the medial and lateral retinacula coursing from the
patella to the medial and lateral collateral ligaments respectively. The axial images then
serve as a localizer for prescribing the coronal and sagittal oblique sections.
The coronal sections are graphically prescribed on an axial image from the patella to the
posterior surfaces of the femoral condyles. The planes is oriented parallel to the
anterior/posterior surfaces of the femoral condyles. This view allows assessment of the
medial and lateral collateral complexes, medial and lateral menisci, and anterior and
posterior cruciate ligaments.
The sagittal sections are graphically prescribed from the lateral to the medial collateral
ligament and aligned parallel with the anterior cruciate ligament. This done by orienting
the planes parallel to the medial border of the femoral condyle, unless the ACL could be
visualized. Vahey et al. (1994) suggest the use of a coronal image and orient the planes
20 - 30 degrees parallel to the ACL.

Pulse sequence

Short TE spin echo sequences such as proton density or T1 weighted are preferable for
meniscal and ligamentous imaging. The T2 weighted images are useful for ACL tears if
the short TE sequences are indeterminate. STIR or PD fat suppressed in the coronal or
sagittal planes are useful to demonstrate better the bone bruises and osteochondral
fractures that are associated with ACL tears. T2*-weighted gradient echo sequence is
equivalent to spin echo imaging for the diagnosis of meniscal pathology. Gradient
echoes have been suggested to provide higher sensitivity for meniscal tears but lower
diagnostic specificity when compared to spin echoes sequences (Fitzgerald, 1994,
Crues III et al., 1999). The increased sensitivity occurs because of the reduced volume
averaging of gradient echoes (thinner slices especially with 3D) and better SNR, whilst
the lower specificity occurs due to the high signal from the articular surface. As in other
areas of the musculoskeletal system, gradient echo are useless for bone marrow
pathology and in areas containing metal. According to Guckel et al., (1995) 3D gradient

MRI Clinical Applications II                                                Joseph Castillo

echo sequences lack of specificity but high sensitivity for meniscal pathology, lack of
sensitivity and specificity for ACL tears due to lesser soft-tissue contrast, which prevents
the detection of intervening soft-tissue collection, and higher delineation of the hyaline
Fast spin echo imaging acquires multiple lines of k-space data within a single TR and
thus scan time is decreased. Attention must however be paid to the selection of scan
parameters with FSE. Kojima et al. (1996) found that when the TE was increased above
19ms the intrameniscal signal changes became less conspicuous. This was mentioned
by Crues III et al. (1999). At 19ms, increase in TR did not affect image quality. Image
blurring was noted above ETL of 4, but ETL of 2 provided the best quality especially
sharpness of menisci. Fat suppression increased the conspicuity of tears and bone
marrow pathology.

Use of fat suppression

Fat suppression is an important technique in musculoskeletal imaging. This technique
is particularly used to eliminate the bright signal from fat and increase the conspicuity of
bone marrow pathology, meniscus and cartilage. There are two sequences, which use
fat suppression. One technique for fat suppression is frequency selective presaturation
(or chemical shift). This method uses the chemical shift difference between lipid proton
and water proton resonance frequency, which is around 220Mhz at 1.5T. With this
technique radio frequency selectively saturates the fat protons in each slice and then
dephased before acquisition. This technique however depends on uniform magnetic
fields and high field strengths, where spectral separation is wider. Therefore it may be
problematic where magnetic field is inhomogeneous due to metallic screws or at low
magnetic fields where spectral separation is too narrow to allow selective excitation.
Other disadvantages of fat saturation include reduced number of slices and unfamiliar
tissue contrast. A T1 weighted separate sequence in the same plane may be required
to help localize signal abnormalities
An alternative method to suppress fat is STIR (Short Tau Inversion Recovery). STIR
images are useful for evaluating marrow pathology, osteomylitis, bone bruises and
infiltrative neoplasms.        This method relies on the application of a 180 degree
radiofrequency pulse at the beginning of the sequence, followed by a time delay (TI).

MRI Clinical Applications II                                             Joseph Castillo

The 180 degree pulse inverts longitudinal magnetization along the negative z axis. T1
recivery then occurs during TI. As the longitudinal magnetization of the various tissue
recovers towards the positive z axis they will pass through zero net magnetization at a
characteristic time depending on its T1 relaxation time. If the pulse sequence is applied
at this time no signal is received from that particular tissue. With STIR, the inversion
time is chosen so that fat is nullified. At 1.5T the TI is 150ms. The disadvantage of
STIR includes the long acquisition time, low SNR and cannot be used with gadolinium-
enhanced sequences. The effect of gadolinium that of shortening the T1 relaxation time
is actually nullified with STIR (Tsao & Mirowitz, 1997)

Use of Contrast Medium

IV contrast is virtually never indicated in knee joint imaging although it may be required
in the delination of bone tumours. Magnetic resonance arthrography is practiced in many
imaging centres for the diagnosis of meniscal tears and chondral defects. This is done
by introducing a dilute solution of gadolinium in saline (1:1000) into the joint capsule.
The knee is then imaged in three planes using T1W or PD weighted both fat
suppressed. Indirect arthrography could be used where the synovial fluid within the joint
enhances 15 minutes post IV administration of gadolinium. This occurs due to slow
spread of the contrast from the synovial vascular network to the intrarticular surface of
the membrane and thus into the intrarticular cavity where its concentration remains high
for about 1 hour. This increases conspicuity of meniscal tears (Andreisek et al, 2001).


A young man referred for MRI after he sustained a knee injury while playing football

Patient Consideration, equipment used and method

MRI Clinical Applications II                                                 Joseph Castillo

MRI was performed on a GE Signa MR/i (GE, Milwaukee) using a transmit/receive coil
also known as chimney coil. The patient completed the metal screening questionnaire.
The procedure was explained to the patient and then asked to change into a hospital
gown. The patient was given earplugs, instructed on the use of the panic buzzer. The
patient was positioned supine, feet first, with the knee in a relaxed slightly flexed position
within the coil. The coil is offset to mark 50 and the other leg is positioned comfortably
at the side on a foam pad. The laser light was positioned over a mark on the coil, which
corresponds to a point below the lower border of the patella.

A clear display of the ACL is essential in knee imaging. The ligament is best seen in
oblique sagittal scans. Using an axial image as localizer the slices are oriented to the
lateral femoral condyle, which runs parallel to the ACL (Vahey et al., 1994) (See end of
section). If the equipment is not capable of doing oblique imaging, then the knee is
externally rotated 15 degrees. This brings the ACL perpendicular (Westbrook, 1999).

Some patients may find it difficult to extend their knee, and the use of the extremity coil
becomes impossible. In this case a flexible coil wrapped around the knee is sufficient.
Alternatively, the Torso Phased Array coil is worth considering. Patients with metal
screws may experience some discomfort and should be asked to inform the practitioner
if this happens.


The site protocol (Table 6.1) used in routine knee imaging is the following
Three plane localizer, Axial PD, Coronal PD/T2, Sagittal PD/T2, Sagittal PD -fs, Coronal

           TR       TE         TI   F   Matrix       N    ET    FO    THK     SA     Time

MRI Clinical Applications II                                                 Joseph Castillo

            mse     msec           A                 E   L     V              T
            c                                        X
Ax PD       2775  20                    384x224      2   16    16    3/0.4    S,I    2:46
Cor         2500  25/102                384x256      2   8     16    4/0.4    S,I    5:25
Sag        2500 25/100                  384x256      2   8     16    4/0.4    S,i    5:47
Sag PD 3275 20                          384x224      3   8     16    3/0.4    S,I    4:41
fs                                                                            Fat
Cor        4775 80           150        256x192      2   10    16    4/1      S,I    5:05
Table 6.1 – Routine Knee protocol as used in Malta


There is a tear in the posterior horn of the medial meniscus and an extensive tear of the
anterior cruciate ligament. The medial collateral ligament appears intact. There is a
small intrarticular effusion. There is a contusion in the lateral tibial plateau (see images
at end of section).

Critical Evaluation.

The PD fat suppressed sequence clearly demonstrates a high linear signal in the
posterior horn of the medial meniscus indicating a tear. The same sequence clearly
demonstrates an ACL tear and the effusion. The Axial PD and Coronal PD confirm the
ACL tear.     A bone bruise in the lateral tibial plateau is clearly seen with the STIR
sequence. This protocol seems to address most common requests. This protocol is a
modification version of a larger extensive protocol, which is still being used in the private
health sector. Our modified protocol has now been accepted by radiologists and is now

Protocol Change.

The protocol currently used at St.Luke’s Hospital, was modified from the adopted one
used at St.Jame’s Hospital. The change was necessary because now the equipment is
a 1.5T using a dedicated extremity coil instead of the 0.5T using a flexible surface coil.

MRI Clinical Applications II                                              Joseph Castillo

In addition more patients are being referred for MRI and we needed to shorten the scan
time without affecting diagnosis.

The protocol at St.James Hospital included an Axial T2 TSE, Sagittal and Coronal T1
SE, Sagittal and Coronal Dual Echo TSE, and Sagittal and Coronal STIR. The scanning
time is in the range of 45 minutes and the whole procedure usually takes one hour. The
change to the current protocol was gradual because our radiologists wanted to be sure
that any change would not affect diagnostic accuracy.          This involved analysis of
literature and liaising with all consultant radiologists.

The first step involved an analysis of the disadvantages of using T1/PD pulse
sequences in Spin echo, Fast Spin echo and gradient echo sequences.              Literature
demonstrates that a controversy exists about the utility of fast spin echo proton density
in the evaluation of tissues with short T2 relaxation time such as menisci. In proton
density fast spin echo T2 relaxation results in image blurring as the echo train
approaches the edges of k-space for short T2 tissues. Some argued that Fast spin
echo sequence has decreased sensitivity for meniscal tears (Kaplan et al., 2001, Creus
III et al.1999), whilst others have shown that with improved sequences which limits the
inter echo spacing together with a low ETL has increased the sensitivity for meniscal
tears (Cheung et al., 1997, Kojima et al, 1996). FSE-XL a new version of FSE by
General Electric Medical Systems is able to achieve shorter echo spacing than FSE
through the use of narrower RF pulses. One of the advantages of FSE-XL reported by
Maier and Verkatesan (GE applicationists) is that image blurring from T2 decay during
the echo train can be greatly reduced for the same ETL used in FSE. Thus higher ETL
could be used with FSE-XL without image degradation. After analyzing the sensitivities
reported in the above literature, it was decided to eliminate the Sagittal and Coronal T1
SE, and introduce a Fast Spin Echo Proton Density sequence. Fat Suppression was
added to the sequence and thus we eliminated the sagittal STIR sequence.                    Fat
suppression makes tears more conspicuous against the suppressed marrow signal in
the femur and tibia. Although all radiologists were informed of this change, at first there
were some mixed reactions with main one being that now they had less sequences to
compare with. However, after two months of running the new protocol, each radiologist
was handed an examination and each were asked to comment on the quality of the
sequences and whether there is scope to address further changes. Although no audit

MRI Clinical Applications II                                                 Joseph Castillo

was carried out to see whether the introduced sequence affected diagnostic accuracy
for meniscal tear, radiologists all commented that the PD –fat sat images were of good
diagnostic quality and did not feel that the new protocol hindered there level of
confidence in reporting. These remarks must be evaluated with caution as no attempt
has been done to evaluate statistically the diagnostic accuracy of this change as was
done at the Oxford MRI Group (Moore, 1992).
In addition, no attempt was made to assess how this change affected report turn around
time. That is are radiologists taking longer or less time to report individual cases?

Currently the Sagittal PD FSE –fatsat is run with intermediate weighting (TEeff between
30 and 36ms). This maximizes contrast between fluid, fibrocartilage and articular
cartilage. The articular cartilage is if intermediate signal whilst fluid is relatively bright
due to magnetization transfer effect inherent in the FSE technique. With FSE
acquisitions, the train of slice selective 180 degrees pulses acts as off resonant pulses
for other slices. As the TEeff is higher than 26msec there is a possibility of missing
meniscal tears, but the Sagittal PD/T2 with a TE of 25 reduce this risk. Thus a possible
modification could be to optimize the Sagittal PD FSE-fs by reducing the TE and ETL
and thus eliminate the Sagittal PD/T2 and reduce scanning time. (This could be a
possible research question for my MSc Dissertation). Again this would require strict
audit of diagnostic accuracy and therapeutic impact, both of which affect patient
management (see section 9).

Although all radiologists accepted the sagittal PD FSE with fat suppression, the axial
plane received various comments. Some preferred the axial plane as it completes the
whole exam with three planes, whilst others commented that the axial did not offer any
diagnostic value. Evaluation of literature has shown that the axial PD, even with high
resolution, is important for the examination of the patellofemoral compartment, articular
cartilage, the collateral ligaments and cruciate ligaments
Articular cartilage requires a high matrix range 512x384. In our protocol, the matrix has
been reduced as axial images are mainly utilized to visualize the anterior cruciate
ligament and serves as a localizer to plane the sagittal PD FSE-fs.
The coronal plane is used to examine the collateral ligaments and serves as another
plane to examine the cruciate. The PD echo depicts clearly the ligaments whilst the
long TE echo is important to diagnose meniscocapsular separation. As the fat in this

MRI Clinical Applications II                                                Joseph Castillo

sequence returns a bright signal it may be difficult to differentiate fat from fluid.
However radiologists never commented on this issue, because the protocol includes a
Coronal STIR that make up for this deficiency. Maybe adding fat suppression with the
Coronal FSE PD/T2 could eliminate the STIR and reduce further scanning time which
currently is approx 24minutes and thus within the appointment time slot of 40minutes.

Reflecting on this protocol change, I would never have imagined that a simple change
would require this amount of work. And yet, considering the implications that a protocol
change might have on diagnostic accuracy, therapeutic impact and patient
management, we have not yet assessed the efficacy of this change. At least we need
the comments of the referring orthopaedic surgeons to assess whether there was any
change in the MRI findings when compared to their patient outcome.
The whole change process, however, helped me to understand more the effects of the
scan parameters on pathology. With hindsight, I have missed an opportunity to use this
change process as a learning instrument to my colleagues whom I have been training
for the last 8 months. Asking radiologists to give constructive feedback on the quality of
the films and how it could be improved has probably helped to reduce resistance for
In conclusion, the educational experience gained by a protocol change process,
together with the met objectives must be balanced with a critical audit to assess the
impact it has in practice in terms of patient care, patient throughput, patient outcome
and level of confidence in reporting. All these could be adversely affected and would
hinder and future change process if the shortcomings are not addressed immediately
through quality circles in which a group of people analyse each step of the change
process through objective quantitative data. But this is another subject.
Thus although I met the short term objective to reduce scanning time, I failed to prove
that diagnostic accuracy was maintained.

The UK Experience.

The routine protocol for Knee is totally different and relies heavily on gradient echo
sequences. The protocol includes a Sagittal 3D GRASS, Coronal STIR and Sagittal T2*

MRI Clinical Applications II                                                  Joseph Castillo

            TR      TE         TI    F   Matrix      N    ET    FO    THK      SA     Time
            mse     msec             A               E    L     V              T
            c                                        X
Sag 3D      36      15               3   256x224     1          20    1.8      S,I    6:30
GRASS                                5
Sag T2*     600     15               2   256x224     2          16    3.5/0           4:32
Cor         3575    24         120       256x192     2    8     18    4/1             2:58

The Sagittal 3D Grass produce slice thickness of 1.8mm and thus allows the radiologist
for multiplanar reconstruction. However gradient echo sequences have been found to
be less accurate than spin echo sequences as regards cruciate ligaments (Irizarry &
Recht, 1997), but they are favoured by radiologists in the UK hospital and it seems are
unlikely to change. The Sag T2* GRE is meant to assess menisci, but again gradient
echo although equally sensitive to spin echo are less specific to tears (Carrino &
Schweitzer, 2002). The Coronal STIR is run to evaluate for bone bruise and collateral
ligaments. The scan time is roughly 14 minutes. Despite this lower scan time, I would
hesitate to adopt this protocol, as it is less specific for the common pathologies affecting
the knee joint.


Accurate diagnosis and ultimate treatment is of utmost importance, because if untreated
this will lead to progressive deterioration of the medial collateral ligament leading to
meniscal tears and damage to the articular cartilage. Treatment involves an ACL
reconstruction. Some patients do extremely well. Ronaldo, the famous world top scorer
is an example. However, there are others in which the reconstruction is not successful
due to abnormalities of the femoral and tibial tunnels with subsequent instability,
impingement ad graft failure. Evaluation of ACL reconstruction with MRI is common

Alternative Imaging Modalities

MRI Clinical Applications II                                                    Joseph Castillo

Clinical examination provides information regarding instability of the ACL. In the injured
knee, with traumas to various structures, clinical examination may be complicated and
difficult due to pain.     There are several tests to assess the instability of ACL, with
sensitivities ranging from 78% to 89%.             Arthroscopy is highly accurate, but it is
expensive and has been associated with 8% complications. General Radiography and
CT scan is limited in clearly showing these structures.


       Andreisek G., Klausner A., Schroder R., Maurer J. (2001) Indirect MR Arthrography of
        the knee joint: Technique, Value and limitations, Imaging Decisions MRI, 5(2):2

       Carrino J.A., Schweitzer M.E. (2002) Imaging of sports-related knee injuries, Radiologic
        Clinics of North America, 40(2):181-202.

       Caudana R., Ternullo S., Bergamo Andreis I.A., Cerini R., Grazioli A., Sommavilla E.
        (1997) The Knee, La Radiologia Medica, 93:156-167.

       Cheung L.P., Li K.C., Hollett M.D., Bergman A.G., Herfkens R.J. (1997) Meniscal tears of
        the knee: accuracy of detection with FSE MR imaging and arthroscoping correlation in
        293 patients, Radiology, 203(2):508-512.

       Cotton A., Delfaut E., Demondion X., Lapeque F., Boukhelifa M.., Boutry N et al, (2000)
        MR Imaging of the knee at 0.2 and 1.5T:correlation with surgery, American Journal of
        Roentgenology, 174(4):1093-1097.

       Creuss III J.V., Manji S.A., Shellock F.G. (1999) Knee In:Magnetic Resonance Imaging
        (eds. Stark DD & Bradley WG Jr.), 3 edn St.Louis,USA, Mosby.

       Fitzgerald S.W. (1994) Magnetic Resonance Imaging of the meniscus, advanced
        concepts, MRI Clinics of North America, 2(3):349-364.

       Maier C.F., Venkatesan R. “MRI Techniques of Orthopaedics Imaging”.online.
        General ElectricMedical Systems. Available:

MRI Clinical Applications II                                                      Joseph Castillo

        3.jsp (1/11/02).

       Guckel C., Jundt G., Schnabel K., Gachter A. (1995) Spin-echo and 3D gradient echo
        imaging of the knee joint: a clinical and histopathological comparison, European journal of
        Radiology, 21:25-33.

       Irizarry J.M. Recht M.P. (1997) MR Imaging of the knee ligaments and the postoperative
        knee, Radiologic Clinics of North America, 35(1):45-76.

       Kojima K.Y., Demlow T.A., Szumowski J. Quinn S.F. (1996) Coronal fat suppression fast
        spin echo images of the knee, evaluation of 202 patients with arthroscopic correlation,
        Magnetic Resonance Imaging, 14(9):1017-1022.

       Maubon A.J., Ferru J.M., Berger V., Soulage M.C., DeGraef M., Aubas P., Coupeau P.,
        Dumont E., Rouanet J.P. (1999) Effect of field strength on MR images:comparison of
        same subject at 0.5T, 1.0T and 1.5T, Radiographics, 19(4):1057-1067.

       Moore N.R., Golding S.J. (1992) Increasing patient throughput in magnetic resonance
        imaging: a practical approach. Oxford MRI Group. British Journal of Radiology,

       Rand T., Imhof H., Turetschek K., Schneider B., Vogel T., Gable C., Tratting S. (1999)
        Comparison of low field 0.2T and hiegh field 1.5T MR Imaging in the differentation of
        intact meniscus, European Journal of Radiology, 30:22-27.

       Shahabpour M., Handelberg F., Casteleyn P.P., Machiels F., Osteaux M. (1997) Imaging
        in sportsmedicine-knee, European Journal of Radiology, 26(1): 23-45.

       Tsao L.Y., Mirowitz S.A. (1997) MR Imaging of the Shoulder Imaging Techniques,
        Diagnostic Pitfalls and normal variants, MRI Clinics of North America, 5(4):683-704.

       Turner D.A., Rapoport M.I., Erwin W.D., McGould M., Silvers R.I. (1991) Truncation
        artifact:potential pitfall in MR Imaging of the meniscus of the knee, Radiology, 179(3):629-

MRI Clinical Applications II                                                      Joseph Castillo

       Vahey T.N., Meyer S.F., Shelbourne K.D., Klootwyk T.E. (1994) MR Imaging of Anterior
        cruciate ligament injuries, MRI Clinics of North America, 2(3):365-380.

       Vellet A.D., Lee D.H., Munk P.L., Hewett L., Eliasziw M., Dunlavy S., Vidito L., Fowler
        P.J., Miniaci A., Amendola A. (1995) Anterior Cruciate Ligament tear: prospective
        evaluation of diagnostic accuracy of middle- and high-field strength MR Imaging at 1.5T
        and 0.5T, Radiology, 197(3):826-30.

        Westbrook C. (1999) Handbook of MRI technique, 2 ed, Oxford, Blackwell Science.


Shared By:
Description: MRI of the Knee MALTA MRI in Practice Collagen