Cartilage injuries following trauma are a common problem and can lead to
premature, secondary osteoarthritis. Standard magnetic resonance imaging
techniques can detect cartilage breakdown associated with morphological
changes, such as decreases in cartilage thickness and volume, but it cannot
detect early changes of the cartilage matrix. There is a need for a
noninvasive method with which to diagnose articular cartilage abnormalities at
early stages in order to initiate early treatment prior to the macroadaptive
changes seen on radiographs.
Early events in the development of cartilage breakdown include the loss of
proteoglycans, changes in water content, and molecular-level changes in
collagen1,2.
Early diagnosis of cartilage injury requires the ability to detect changes in
proteoglycan concentration and collagen integrity noninvasively before gross
morphological changes occur. Standard cartilage-dedicated magnetic resonance
imaging techniques such as T2-weighted, proton-density-weighted fast-spin-echo
sequences and spoiled gradient-echo sequences are inconclusive in quantifying
early degenerative changes, especially biochemical changes such as
proteoglycan
loss3.
It appears that T1rho-weighted imaging can provide quantitative measures of
pathologic cartilage matrix changes. Previous studies have demonstrated
significant differences in T1rho relaxation times when patients with
symptomatic osteoarthritis have been compared with normal
controls4. In the
present study, we describe two cases in which cartilage injuries were not
detected with standard morphological magnetic resonance imaging but in which
T1rho-weighted imaging was able to detect cartilage abnormalities in vivo that
were confirmed at the time of arthroscopy. The patients were informed that
data concerning these cases would be submitted for publication.
Case 1. A forty-five-year-old man presented with a three-week
history of pain in the right knee following a motor-vehicle accident. Clinical
examination demonstrated a moderate intra-articular knee effusion, a positive
grade-2 Lachman test, and a positive pivot-shift test. The patient also had
mild discomfort on the McMurray test. The patient underwent magnetic resonance
imaging of the right knee, which was performed with use of a 3-Tesla (3-T) GE
Signa EXCITE (Waukesha, Wisconsin) magnetic resonance imaging scanner and a
quadrature transmit/receive knee coil. The T1-weighted spin-echo technique
(repetition time = 700 ms, echo time = 15 ms, field of view = 16 cm, matrix =
288 × 224, slice thickness = 3 mm), T2-weighted fat-saturated
fast-spin-echo technique (repetition time = 3700 ms, echo time = 68 ms, field
of view = 16 cm, matrix = 288 × 224, slice thickness = 3 mm), and
high-resolution three-dimensional fat-saturated spoiled gradient-echo
technique (repetition time = 20 ms, echo time = 7.5 ms, field of view = 16 cm,
matrix = 512 × 512, slice thickness = 1 mm) revealed a small flap tear
of the lateral meniscus, a complete anterior cruciate ligament tear, and an
associated posterolateral tibial bone bruise in the right knee. No articular
cartilage injuries were noted.
To better assess the condition of the articular cartilage, 3-T
T1rho-weighted images were also acquired. These images were acquired with use
of a previously developed sequence based on spin-lock techniques (field of
view = 16 cm, in-plane resolution = 0.6 × 0.6 mm, slice thickness = 3
mm, time of spin-lock [TSL] = 20/40/60/80 ms, spin-lock frequency = 500 Hz,
total acquisition time = approximately thirteen
minutes)4,5.
The images were then postprocessed with use of specialized
software4.
Postprocessing first involved semi-automatic segmentation of the cartilage
from the high-resolution spoiled gradient-echo magnetic resonance images with
use of an in-house-developed spline-based
algorithm6,7.
Next, we generated a color map of the T1rho-calculated images of the cartilage
by fitting the T1rho-weighted images pixel-by-pixel according to the equation
S (TSL) = S0 × exp(-TSL/T1rho). These T1rho maps were then registered
and overlaid on anatomic three-dimensional spoiled gradient-echo images.
The examination demonstrated increased T1rho relaxation times along the
posterolateral aspect of the tibial plateau, overlying the tibial plateau
region with a bone bruise (Fig.
1). The mean T1rho value (and standard deviation) of the cartilage
directly overlying the bone bruise was found to be 60.2 ± 13.7 ms. The
mean T1rho value of the remaining cartilage was found to be 37.5 ± 14.3
ms. Intraoperative evaluation confirmed the diagnosis of an anterior cruciate
ligament tear and a lateral meniscus tear. Evaluation of the lateral
compartment according to the Outerbridge classification system demonstrated
grade-1 changes in the articular cartilage, with focal areas of grade-2
fissures. These changes consisted of softening and linear partial fissures in
the articular cartilage, directly in front of and underneath the posterior
horn of the lateral meniscus (Fig 1,
D). The location of the abnormal cartilage was
directly overlying the associated bone bruise.
Case 2. A nineteen-year-old collegiate basketball player had an
acute onset of anterior pain in the right knee during practice. The patient
had persistent anterior knee pain despite physical therapy. A 3-T magnetic
resonance imaging evaluation (performed with use of the same techniques and
scanner as described in Case 1) demonstrated a small subchondral cyst in the
trochlear region but otherwise showed no cartilage or ligamentous abnormality.
Quantitative T1rho-weighted imaging demonstrated increased T1rho relaxation
times along the trochlea in the region of the subchondral cyst
(Fig. 2). The mean T1rho value
of the cartilage directly overlying the subchondral cyst was found to be 44.2
± 8.2 ms. The mean T1rho value of the remaining cartilage was found to
be 33.4 ± 11.5 ms. Intraoperative findings demonstrated a focal
cartilage defect measuring 4 × 5 mm, with an unstable cartilaginous rim,
at the central portion of the trochlea
(Fig. 2, D).
After débridement, an osteochondral lesion was found to be extending
into the small subchondral cyst. The cyst was mechanically débrided.
After one year of follow-up, the patient was doing well. He was not limited by
the right knee, and the symptoms had resolved.
Standard magnetic resonance imaging techniques are able to demonstrate
morphological defects and surface irregularities of cartilage, but the early
diagnosis of cartilage degeneration requires the ability to noninvasively and
nondestructively detect matrix changes in proteoglycan content and collagen
integrity. In the two cases described here, we employed in vivo 3-T
T1rho-weighted imaging to detect cartilage degeneration that was not seen on
standard magnetic resonance imaging scans. The T1rho parameter describes the
spin-lattice relaxation in the rotating
frame8. T1rho
relaxation probes the slow-motion interactions between motionally restricted
water molecules and their macromolecular environment. Thus, changes to the
extracellular matrix, such as proteoglycan loss, may be reflected in the
measurements of T1rho. All magnetic resonance imaging examinations were
performed with use of a 3-T GE Signa EXCITE magnetic resonance imaging scanner
and a quadrature transmit/receive knee coil. The acquisition times for the
T1rho-weighted images and the three-dimensional spoiled gradient-echo images
were approximately thirteen and eleven minutes, respectively. The images were
then postprocessed with use of specialized
software4. The
semiautomatic postprocessing allowed us to identify the areas of cartilage
that had increased T1rho values. With use of arthroscopy, we were able to
correlate areas that demonstrated increased T1rho values with areas
demonstrating cartilage injury.
T1rho-weighted imaging is one of a few techniques that have shown the
potential of magnetic resonance imaging to reflect changes in the biochemical
composition of cartilage in patients with early osteoarthritis. Other
techniques, including T2 quantification, 23Na magnetic resonance imaging, and
delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC),
also have shown promising results in the imaging of cartilage composition. All
of these techniques are complementary to standardized cartilage-sensitive
images and may provide information regarding cartilage changes (either
proteoglycan or collagen) that may exist prior to changes in cartilage
thickness or surface morphology. However, some of the techniques may have
limitations that can hinder generalized clinical use.
The dGEMRIC technique, which has been validated in multiple studies to
allow for the assessment of the proteoglycan component of articular
cartilage9,10,
requires a sixty to ninety-minute wait after either an intravenous or
intra-articular injection of the contrast agent (Gd-DPTA) for effective
penetration. Furthermore, joint exercise is required by the patient and long
acquisition times are needed for T1
quantification11.
Sodium 23 nuclear magnetic resonance imaging, which uses sodium concentrations
as a marker for proteoglycan
loss12, is of
limited clinical use because of the inherent low sensitivity of sodium signal
and the limited availability of sodium magnetic resonance imaging, which
requires special coils and
hardware13. Like
T1rho mapping, T2 mapping does not require the use of special hardware, coils,
or contrast agents. However, the T2 parameter seems to be most sensitive to
the collagen component of
cartilage14,
whereas the T1rho parameter seems to be more sensitive to proteoglycan
concentrations15,16.
This is of importance because it has been suggested that loss of proteoglycan
is an initiating event in early osteoarthritis, whereas neither the content
nor the type of collagen is altered in early
osteoarthritis17.
Furthermore, T1rho relaxation times do not seem to be affected by the
orientation of collagen that can affect T2 relaxation
techniques13. Other
reports have suggested that T1rho relaxation may prove to be a more sensitive
clinical tool than T2 for early osteoarthritis
detection4.
Our study was implemented with the use of a 3-T magnetic resonance imaging
scanner because of the advantages afforded by the use of a higher field
strength (such as increased signal-to-noise ratio and higher
resolution)18,19.
However, T1rho-weighted magnetic resonance images also can be acquired on more
readily available 1.5-T
scanners20.
Clinical in vivo studies comparing T1rho values of cartilage in normal
patients with those in patients with symptomatic osteoarthritis have been
conducted at both 1.5 and 3 T, with similar
results3,4.
Regatte et al.3
compared the T1rho values for six patients who had symptomatic osteoarthritis
with those for eight controls at 1.5 T and found a significant elevation in
T1rho values in the osteoarthritis group. The osteoarthritis group had T1rho
values that ranged from 63 to 95 ms, whereas the control group had values that
ranged from 45 to 55 ms. In a previous study, we obtained similar findings at
3 T, with a significant difference (p = 0.002) in the average T1rho value
between nine patients with osteoarthritis (53.06 ± 4.60 ms) and ten
control subjects (45.04 ± 2.59
ms)4.
One limitation of the present report was that the resolutions of the
T2-weighted fat-saturated fast spin-echo images (0.56 × 0.71 mm) were
not optimized because of the restriction of the total scan time per patient
and the relatively long echo time that was used (68 ms), which may cause loss
of signal from cartilage components (such as those in the radial zone) with a
very short T2. Thus, the morphological cartilage aberrations likely could have
been detected with fast-spin-echo parameters optimized for cartilage. A second
limitation related to T1rho-weighted imaging is that, compared with
conventional T1 and T2-weighted imaging sequences, T1rho-weighted sequences
have a higher specific absorption rate because of the application of spin-lock
pulses. However, previous studies have demonstrated the feasibility of
implementing T1rho-weighted techniques both at 1.5 and 3 T that do not exceed
the specific absorption rate for human
imaging4,20.
The specific absorption rate in the present study did not exceed the safety
guidelines for imaging human subjects, with acquisition parameters that were
comparable with clinical imaging sequences (in-plane resolution = 0.6 ×
0.6 mm, slice thickness = 3mm, and field of view = 16 cm). This sequence can
be easily incorporated into current clinical protocols with an acquisition
time of approximately thirteen minutes. Techniques to further reduce the
specific absorption rate and acquisition time are in development.
In conclusion, in vivo T1rho-weighted imaging has demonstrated the
feasibility of detecting early trauma-induced articular cartilage changes that
are not seen with standard magnetic resonance imaging. Quantitative imaging
may enhance our ability to detect subtle, early matrix changes associated with
articular cartilage injuries when used in conjunction with standardized
cartilage-sensitive imaging. ?