DOI:10.2214/AJR.07.2693
AJR 2008; 190:W1-W7
© American Roentgen Ray Society
MRI of Acute Bone Bruises: Timing of the Appearance of Findings in a Swine Model
Donna G. Blankenbaker1,
Arthur A. De Smet1,
Ray Vanderby2,
Ronald P. McCabe2 and
Stephanie A. Koplin3
1 Department of Radiology, University of Wisconsin School of Medicine and Public
Health, E3/311 Clinical Science Center, 600 Highland Ave., Madison, WI
53792-3252.
2 Department of Orthopedics and Rehabilitation, University of Wisconsin School
of Medicine and Public Health, Madison, WI.
3 Department of Pathology, University of Wisconsin School of Medicine and Public
Health, Madison, WI.
Received June 6, 2007;
accepted after revision July 10, 2007.
WEB This is a Web exclusive article.
Address correspondence to D. G. Blankenbaker
(dg.blankenbaker{at}hosp.wisc.edu).
Abstract
OBJECTIVE. We developed an animal model of bone injury to determine
the timing of the MR appearance of bone bruises and to follow these changes
over time.
MATERIALS AND METHODS. We performed serial MRI of 16 knees of eight
swine with one nontraumatized knee as a control and 15 knees traumatized by
direct patellar impact injuries using a force-calibrated device. All knees
were imaged on a 1.5-T scanner using an 8-channel phased-array coil with
T1-weighted, fat-saturated T2-weighted and STIR sequences. Scanning was
performed at 1, 6, 12, and 30 hours after impact injury. Two radiologists
independently reviewed each MR examination to identify MR signal intensity
changes in the patellae and adjacent femoral condyles.
RESULTS. In the 15 traumatized knees, bone bruises were noted in 93%
of the patellae on T1-weighted images and in 87% of the patellae on
fluid-sensitive MR images at 1 hour after injury and in 100% of the patellae
at 6 hours. T1-weighted images were insensitive for detection of bone bruises
in the femoral condyles. Bone bruises in the femoral condyles were seen on
fluid-sensitive MR sequences as early as 1 hour after injury, with an
increasing frequency over the 30-hour period.
CONCLUSION. Bone bruises can be seen as soon as 1 hour after trauma
but may not be seen until 30 hours after trauma. Fluid-sensitive
(fat-saturated, T2-weighted, and STIR) MR sequences are more sensitive than
T1-weighted images in the detection of bone bruises.
Keywords: bone bruise • knee • MRI
Introduction
Posttraumatic bone injuries, often called bone bruises or
contusions, have been noted on MRI since 1988
[1]. A bone bruise is seen on
MRI as an area of low signal intensity on T1-weighted images and high signal
intensity on fluid-sensitive sequences in the marrow
[1-3].
These MRI findings are thought to represent marrow edema and hemorrhage, which
may result from injury to the marrow itself from transmitted shock waves or
from trabecular bone fractures
[4,
5]. The limited histologic
evidence that is available from biopsies of patients confirms the presence of
marrow edema, hemorrhage, and trabecular fractures in these bone bruises
[6]. Although posttraumatic
bone bruises are common, the timing of the appearance of these bruises in the
acute phase has had only limited study.
Recently, we performed MRI on an athlete 2 hours after traumatic hip
subluxation, and the findings revealed a posterior labral tear with hip
effusion but no bone abnormality (Fig.
1A). A subsequent MRI performed 3 months later to confirm healing
of the labral tear revealed an anterosuperior bone marrow abnormality
(Fig. 1B). We were uncertain as
to whether the abnormality at 3 months after the injury was a persistent
delayed-onset bone bruise or an area of osteonecrosis. Clinical and MR
follow-up documented clearing of the MR abnormality within 6 months and
completely normal hip function and radiographic appearance at 2 years after
the injury. Because of the clinical importance of this differential diagnosis,
we developed an animal model for a bone bruise and used this model to
determine the timing of MR changes in the bone marrow after bone injury.
View larger version (83K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —21-year-old male athlete with acute traumatic hip injury. Coronal
T2-weighted fat-saturated image obtained 2 hours after acute injury does not
show any bone marrow edema. Posterior labral tear was noted on axial
T2-weighted imaging (not shown).
|
|
View larger version (73K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —21-year-old male athlete with acute traumatic hip injury. Coronal
T2-weighted fat-saturated image obtained 3 months after injury shows focal
area of abnormally increased T2 (arrow). Sagittal T1-weighted images
(not shown) confirmed anterior location of this signal abnormality.
|
|
Materials and Methods
Experimental Model
Our institutional animal research committee gave protocol approval for this
study. All policies on humane care and use of laboratory animals were
followed.
Ten crossbred female swine from a commercial vendor (University of
Wisconsin Arlington Farms) (mean weight, 51 kg; range, 45-57 kg; mean age, 11
weeks; age range, 10-12 weeks) were used for this study. Adolescent swine were
chosen on the basis of size so they could be moved multiple times while
sedated. We decided to use the swine knee to develop our model for bone
bruises because of the size of this joint in a swine and its morphologic
similarity to the human knee. We specifically selected the patellofemoral
joint for study because striking the anterior surface of the patella allowed
more precise measurement of the impacting force than could be achieved in a
model that attempted to reproduce the clinically common femorotibial bone
bruises such as seen in anterior cruciate ligament injuries. Selection of the
patella as the point of impact was also advantageous because it provided a
site with a direct contusion, the patella, and a site with transmitted
osteochondral forces, the femoral trochlea.
Before the start of our study, two swine knees were scanned without any
knee trauma to learn the normal MR appearance of the nontraumatized swine
knee. In addition, the impacting device was tested on two swine knees for
standardization and calibration.
Method for Impacting Patella
Before beginning the MR portion of the study, we designed a pneumatically
accelerated mass system to impact the swine patella. This concept was our
adaptation of the system used by Atkinson and Haut
[7]. The pneumatic system
allowed more flexibility than a drop-weight system to control the final energy
and velocity of the accelerated mass. By adjusting the acceleration chamber
length (via adjustable stops) and the mass of the impactor, a very wide range
of impact energies and final velocities was available. We used a high-velocity
impact to minimize the possibility of strain-induced fracture of the patella.
The basic system (Fig. 2)
consisted of a cylindric impactor mass (252 g) with polytetrafluoroethylene
bearings (Fig. 3), an
acceleration chamber fitted with optical sensors to measure final velocity, an
electrically operated high-flow-rate valve, and a pressure chamber with a
digital pressure gauge to monitor chamber pressure.
Pressure was applied to the pressure chamber via a pressure regulator and a
ball valve and was read via the digital gauge to ± 0.1 psi (0.69 kPa).
Pressure was released by a small switch and a battery. Once released, the mass
accelerated down the acceleration chamber, where it interrupted two light
beams separated by a precise distance, thereby allowing an accurate measure of
the final velocity. The impactor then hit a custom-fabricated load cell
positioned between the knee and the impactor. Using custom signal
conditioning, load versus time data were relayed to a digital oscilloscope
(TDS1012, Tektronix) and saved for later analysis. Analysis of the data was
done with Microsoft Excel, and the resulting load data were plotted versus
time.
The pneumatically accelerated mass system was then tested on one swine
before beginning the study (two knees) to determine the force that could be
applied without visible fracturing (shattering) of the patella. To place the
device accurately, a skin incision was made directly over the patella, which
allowed palpation and marking of the patella. The size of the impaction device
was approximately one half the size of the swine patella.
Before animal testing, a certified animal care technician sedated the swine
with tiletamine hydrochloride and zolazepam hydrochloride 7 mg/kg (Telazol IM,
Fort Dodge Animal Health), xylazine 2.2 mg/kg IM (Xyla-Ject, Phoenix
Pharmaceutical), atropine sulfate 0.05 mg/kg IM. IV access was established in
each swine's ear vein. Propofol 2-8 mg/kg was used if additional sedation was
needed for intubation. Maintenance of anesthesia throughout the procedure was
by inhaled 1-3% isoflurane via a 5.0-7.5 endotracheal tube. A dose of
buprenorphine 0.005-0.01 mg/kg IM was given before the start of bone injury
followed by a second dose given at 1-6 hours after trauma as needed to
maintain analgesia at the site of surgery until the animal was sacrificed.
After the initial development of the device for impacting the patella, we
then studied both knees of eight swine, of which 15 were traumatized and one
used as our control. All 15 knees were incised for accurate placement of the
impaction device. Both the patella and distal femur were evaluated for bone
injury due to both direct and indirect injury causing the bone bruise.
MRI Studies
After the patellar impaction injury, the animals were immediately taken to
the MR scanner for imaging. All 16 animal knees were imaged on a 1.5-T magnet
using an 8-channel phased-array coil. All animals were positioned supine with
the coil centered over the knee joint. This was the best position for imaging
because the coil was centered over the knees and because of the anesthesia.
The sedation and anesthesia protocols used for imaging were identical to those
used during the creation of the knee injury. The following sequences were
performed: axial and sagittal T1 (TR/TE, 500/20), fat-saturated fast spin-echo
T2 (2,200/76; echotrain length, 3), and fast spin-echo STIR (3,000/48;
inversion time, 160). Coronal fast spin-echo STIR (3,000/50; inversion time,
160) sequences were also performed along the long axis of the femur. A 16-cm
field of view; matrix, 256-320 x 224-256; number of excitations, 2; and
3-mm thickness with 0.5-mm gap were used for all sequences. Scanning was
performed immediately (completed within 1 hour after injury; average time,
30-40 minutes), 6, 12, and 30 hours after trauma. Because of cost and research
scanner access, we were not able to obtain scans at hourly intervals.
Our initial animal was scanned out to 12 hours with subsequent animals
scanned out to 30 hours. Two animals died after the 12-hour scan; another died
after the 6-hour scan; only the right knee in one animal was imaged at the
6-hour time frame, with the left knee not imaged after the initial 1-hour
scan.
Two musculoskeletal radiologists independently reviewed each MR examination
documenting patellar and femoral signal intensity changes and then resolved
differences in interpretation by consensus reading. Each patella and the
trochlear portions of the distal femoral condyles were subjectively evaluated
for changes in the MR appearance of the marrow from the normal appearance of
adolescent swine knees on the T1-weighted, fatsaturated T2-weighted, and STIR
images. The T1-weighted and fluid-sensitive sequences were read independently
of each other. The readers were not blinded to the timing of the images.
Histologic Studies
After the final MRI, the knee joints were removed after the animals were
sacrificed. After fixation with 10% formalin, the femurs were cut
longitudinally with a band saw slightly anterior to the midline. The femurs
and patellae were subsequently decalcified, and two coronal sections from the
distal femur and two transverse sections from the midportion of the patellae
were submitted for paraffin embedding. Histologic sections were cut at 4 µm
in thickness, stained with H and E, and evaluated for the presence of
trabecular microfractures, hemorrhage, and edema. The location of the sections
used for histologic examination was chosen by estimating the site of the
patellar impact on the distal femur.
The edema in the histologic specimens was measured by estimating the
fraction of edema in the marrow space. This was done with low-power
examination of the slide by estimating the acellular area on the slide because
edema histologically appears as a separation of cells. The edema on each slide
was estimated twice by the same reviewer at different times to ensure internal
consistency. Histologically, edema is represented by a separation of, or space
between, the hematopoietic cells of the bone marrow space. We attempted to
exclude artifact "separation of cells" by identifying edema that
was accompanied by fibrin, proteinaceous debris, and extravasated RBCs.
Results
The one control knee had normal signal within the patellae and femurs on
the initial and serial MR scans. The normal adolescent swine knee has
prominent patellar and distal femoral ossification centers that have high
signal intensity on T1-weighted images and low signal intensity on
fat-saturated T2-weighted or STIR images (Fig.
4A,
4B,
4C).
View larger version (95K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A —Control nontraumatized swine knee. Patellar and femoral ossification
centers show high signal intensity on sagittal T1-weighted image (A)
and low signal intensity on T2-weighted fat-saturated (B) and STIR
(C) images.
|
|
View larger version (91K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B —Control nontraumatized swine knee. Patellar and femoral ossification
centers show high signal intensity on sagittal T1-weighted image (A)
and low signal intensity on T2-weighted fat-saturated (B) and STIR
(C) images.
|
|
View larger version (74K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4C —Control nontraumatized swine knee. Patellar and femoral ossification
centers show high signal intensity on sagittal T1-weighted image (A)
and low signal intensity on T2-weighted fat-saturated (B) and STIR
(C) images.
|
|
Impact Load
A plot of eight impacts on the traumatized knees
(Fig. 5) shows the
reproducibility of the impact loading. Peak loads observed were around 15,000
N and occurred at around 115 milliseconds after the first measured contact.
Average peak compressive load was calculated to be 14,900 N, average impact
energy was 35.6 J, and final velocity was 16.8 m/s. Loads are quite high but
very short in duration, keeping total impact energy below the mean value
reported by Atkinson and Haut
[7] for bone fracture-producing
impacts to human knees (60 ± 25 J).
MRI Findings
In all MR scans in which both a fat-saturated T2-weighted and STIR series
were obtained, when a high-signal-intensity abnormality was identified, it was
always seen on both sequences except in the femur of one knee on the 1-hour
scan in which the STIR images were positive but the T2-weighted images were
negative. Overall, the findings were slightly more conspicuous (higher signal
and more obvious) on the STIR images. In view of their similar edema
sensitivity, the following data consider the findings on either the
fat-saturated T2-weighted or STIR sequence as representing an abnormality on a
fluid-weighted MR sequence.
Patella—Both the T1-weighted and fluid-weighted sequences
were sensitive in identifying findings of a patellar bone bruise soon after
the injury (Fig. 6). Low-signal
abnormality on T1-weighted images was seen in 93% (13/14) of the patellae and
high-signal abnormality in 87% (13/15) of the patellae on fluid-weighted MR
sequences obtained within 1 hour of the injury (Fig.
7A,
7B). All patellae had both
T1-weighted and fluid-weighted MR abnormalities at 6 hours after the injury.
These changes persisted at 12 and 30 hours after the injury except for one
patella at 30 hours in which the signal abnormality was no longer seen on
either T1-weighted or fluid-sensitive sequences. An increase in low signal
intensity over time was seen in six of the patellae and an increase in high
signal on the fluid-sensitive sequences in 10 patellae. An overview of the
results is provided in Table
1.
View larger version (72K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7A —Traumatized swine patella. Sagittal STIR MR images at initial 1-hour
scan (A) show edemalike signal intensity (arrow) within
patella and at 30-hour scan (B) show increased signal intensity
(arrow) within patella.
|
|
View larger version (73K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7B —Traumatized swine patella. Sagittal STIR MR images at initial 1-hour
scan (A) show edemalike signal intensity (arrow) within
patella and at 30-hour scan (B) show increased signal intensity
(arrow) within patella.
|
|
Femoral condyles—Unlike the situation in the patellae, a
low-signal abnormality in the femoral condyles was noted on T1-weighted images
in only one knee at the 1-, 6-, and 12-hour intervals after the injury and in
29% (2/7) of the knees imaged at 30 hours after the injury. In contrast, the
fluid-weighted MR sequences had a high-signal abnormality in 33% (5/15) on the
examination performed within 1 hour of the injury (Fig.
8A,
8B). The frequency of MR
abnormality then steadily increased until the 30-hour examination when 100%
(6/6) of the femurs showed a high-signal abnormality. An increase in high
signal intensity over time was seen in nine femurs on the fluid-sensitive
sequences. No change in T1 low signal intensity was detected within the
femurs. An overview of the results is provided in
Table 1.
View larger version (81K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8A —Traumatized swine patella. Sagittal STIR MR image at 1 hour after
injury shows subtle subchondral edema (arrow). This area of
abnormality increases in signal intensity by 12-hour sagittal image (B)
(arrow). In B, there is also extensive high-signal abnormality
in patella.
|
|
View larger version (79K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8B —Traumatized swine patella. Sagittal STIR MR image at 1 hour after
injury shows subtle subchondral edema (arrow). This area of
abnormality increases in signal intensity by 12-hour sagittal image (B)
(arrow). In B, there is also extensive high-signal abnormality
in patella.
|
|
Histologic Findings
Table 2 summarizes our
histologic analysis of 16 knees. The histology was normal only in the control
left knee of swine 2. All 15 traumatized patellae and femurs had varying
degrees of edema, with the greatest amounts of edema within the patellae.
Hemorrhage was seen in two patellae and three femurs
(Fig. 9). There were no
trabecular fractures.
Discussion
MRI is sensitive in the detection of occult stress and posttraumatic
fractures in and around the knee joint
[4,
5]. Bone bruises were first
described in association with major knee injuries
[8,
9]. The bone bruise may be the
only obvious sign of injury and prompts a closer inspection for more subtle
lesions, such as ligamentous and meniscal tears
[5]. Detection of bone bruises
on MRI depends on imaging the marrow and its evolution in response to an acute
injury. The MRI appearance of a bone bruise likely represents the underlying
pathology thought to be trabecular disruption, hemorrhage, and edema
[4,
10]. Microfractures of the
subchondral trabeculae may lead to local cancellous increased fluid content
from hemorrhage [10]. Previous
studies have described the natural history of bone bruises
[11,
12]; however they did not
describe when the bone bruise first occurs at MRI. In the study by Miller et
al. [11], follow-up MRI after
initial MR was completed at various time intervals ranging from 1.5 to 12
months in 24 (83%) of 29 patients with isolated medial collateral ligament
injuries. Complete resolution of bone bruises was observed in all cases. In
the study by Escalas and Curell
[12] using the rabbit model,
an experimental occult bone injury was produced in the rabbit knee and MRI was
performed after killing the animals at 1, 3, and 9 weeks after impact. At 1
and 3 weeks similar changes were noted on MRI, and at 9 weeks MRI showed less
intense image patterns and one knee did not show any abnormal imaging
findings.
The clinical significance of our study goes back to the question in the
evaluation of our patient who had a traumatic injury to the hip. We had
originally assumed that a bone bruise should be seen on an MR image obtained 2
hours after the injury. Thus we thought that the signal abnormality noted on
the MR image had to represent osteonecrosis. However, the location of the MR
abnormality on the anterosuperior aspect of the femoral head was not a typical
appearance or location for osteonecrosis. In fact, this location suggested a
bone injury suffered during reduction of a presumed posterior hip subluxation
similar to the humeral head Hill-Sachs injury in an anterior shoulder
dislocation. On the basis of our animal study showing that a bone bruise may
not be visible on MRI for more than 12 hours after injury, we now think that
our patient had a slowly resolving bone bruise rather than osteonecrosis.
Our results indicate that after a high-impact direct bone injury such as
the patellae in our model, a bone bruise is present approximately 90% of the
time on either T1-weighted or fluid-sensitive MR sequences done within 1 hour
after the injury and at 100% at 6 hours after injury. However, with
lesser-impact-force injuries across an osteochondral surface, such as the
femoral condyle in our model, a bone bruise may take more than 12 hours to
appear on a fluid sequence. In our study, all six swine femoral condyles that
were imaged at 30 hours showed a bone bruise pattern on a fluid-weighted MR
sequence.
We also found that both T1-weighted imaging and fluid-sensitive imaging
(fat-saturated T2-weighted or STIR) were nearly equally sensitive in the
detection of high-impact-injury bone bruises in the patellae. However,
fluid-sensitive MR sequences were more sensitive than T1-weighted images in
the detection of bone bruises within the femoral condyles. We hypothesize that
the MRI findings were more apparent in the patella than in the femoral
condyles because the patellar injuries were more severe due to the direct bone
impact on to the patella. In contrast, the injuries of the condyles were
likely of a lower impact force because the chondral surfaces of the patella
and femur diffused the forces impacting the femoral condyles. The greater
sensitivity of fluid-weighted sequences for marrow abnormality when compared
with T1-weighted images is well documented
[1,
3-5,
13,
14].
To our knowledge, there has been one previous study on an animal model of
bone bruises [15] in swine. In
this study, a small metal hammer was used to create a bone bruise of the
medial aspect of the proximal tibia. The authors did not quantify the force on
the tibia. Each tibia was struck with a hammer until the cortex of the tibia
was mildly depressed. The animals were imaged immediately after surgery and
biweekly until they were sacrificed. They did not determine when bone bruises
first appeared.
We found that fat-saturated T2-weighted and STIR imaging were comparable in
detection of bone marrow edema. In only one femoral condyle did we find edema
on the STIR images that was not detected on the fat-saturated T2-weighted
images. This finding agrees with a previous study evaluating bone contusions
of the knee comparing fat-saturated T2-weighted and STIR images in which a
comparable sensitivity and specificity were found for the two pulse sequences
[1]. In that study, STIR
imaging had slightly increased lesion conspicuity, whereas T2-weighted
fat-saturated images showed slightly improved overall subjective image
quality. We agree with their assessment that when rapid detection of bone
contusions is desired or when a lesion is suspected, STIR imaging is
preferable over fat-saturated T2-weighted imaging. However, we did not find
that fat-saturated T2-weighted imaging had improved subjective image quality.
This difference between the findings of the previous study and ours likely
reflects improvement in MR image quality in the 10 years since their
study.
Histologic analysis of the knees showed marrow edema within all of the
traumatized patellae and adjacent femurs but no marrow abnormalities in the
nontraumatized knee. We were surprised to find that hemorrhage was noted
histologically in only two of the traumatized patellae and three of the
traumatized femurs. In addition, no trabecular microfractures were found. Our
histologic findings support the findings reported in the literature of the
histology of a bone bruise. At histopathologic examination in MR-documented
bone bruises, microfractures of cancellous bone, edema, and hemorrhage were
also identified [6,
15].
Our study has a number of limitations. First, we were not able to image all
of the animals at all intervals because of the death of several animals during
the study. We kept the animals under analgesic sedation after the impact
injury per the requirement of our animal care committee. This continuous
sedation likely contributed to the deaths, which occurred before the scheduled
delayed scanning at 30 hours. Second, because of the practical aspects of
research scanner access and the need for anesthesia during MRI, we were not
able to obtain scans at hourly intervals. Thus when an MR abnormality was not
noted at 1 hour but was seen at 6 hours, we cannot be certain when the MR
abnormality appeared during the 1- to 6-hour interval. Third, we did not
evaluate whether there might be additional MR abnormalities that occurred
later than the 30-hour MR image. Fourth, we did not obtain a pretrauma MR
image in every knee; however, it is unlikely that the juvenile swine would
have preexisting bone contusions.
In conclusion, our study showed that in an animal model, MRI can depict
early changes in bone marrow resulting from direct and indirect injury to
bone. Both direct and indirect bone bruises are usually present on MRI
performed at 12 hours after trauma but may be delayed up to 30 hours.
Clinically, if an MR examination is performed too soon after trauma and is
negative for injury, with the patient having continued pain or limitations, a
follow-up MRI may be useful 5-7 days after trauma.
Acknowledgments
The authors gratefully thank Kimberly Mauer, Experimental Surgery,
University of Wisconsin School of Medicine and Public Health for the expert
assistance in sedation and care of our swine during this project. The authors
also thank Kelli Hellenbrand, MR research technologist, Department of
Radiology, University of Wisconsin School of Medicine and Public Health for
assistance in scanning our subjects. We also thank Ben Graf, Professor of
Orthopedics and Rehabilitation, University of Wisconsin School of Medicine and
Public Health, for his advice and input before beginning this project.
References
- Yao L, Lee JK. Occult intraosseous fracture: detection with MR
imaging. Radiology 1988;167
: 749-751[Abstract/Free Full Text]
- Arndt WF 3rd, Truax AL, Barnett FM, Simmons GE, Brown DC. MR
diagnosis of bone contusions of the knee: comparison of coronal T2-weighted
fast spin-echo with fat saturation and fast spin-echo STIR images with
conventional STIR images. AJR 1996;166
: 119-124[Abstract/Free Full Text]
- Rosen MA, Jackson DW, Berger PE. Occult osseous lesions documented
by magnetic resonance imaging associated with anterior cruciate ligament
ruptures. Arthroscopy 1991;7
: 45-51[Medline]
- Mink JH, Deutsch AL. Occult cartilage and bone injuries of the
knee: detection, classification, and assessment with MR imaging.
Radiology 1989;170
: 823-829[Abstract/Free Full Text]
- Newberg AH, Wetzner SM. Bone bruises: their patterns and
significance. Semin Ultrasound CT MR1994; 15:396
-409[CrossRef][Medline]
- Rangger C, Kathrein A, Freund MC, Klestil T, Kreczy A. Bone bruise
of the knee: histology and cryosections in 5 cases. Acta Orthop
Scand 1998; 69:291
-294[Medline]
- Atkinson PJ, Haut RC. Subfracture insult to the human cadaver
patellofemoral joint produces occult injury. J Orthop
Res 1995; 13:936
-944[CrossRef][Medline]
- Kaplan PA, Gehl RH, Dussault RG, Anderson MW, Diduch DR. Bone
contusions of the posterior lip of the medial tibial plateau (contrecoup
injury) and associated internal derangements of the knee at MR imaging.
Radiology 1999;211
: 747-753[Abstract/Free Full Text]
- Schweitzer ME, Tran D, Deely DM, Hume EL. Medial collateral
ligament injuries: evaluation of multiple signs, prevalence and location of
associated bone bruises, and assessment with MR imaging.
Radiology 1995;194
: 825-829[Abstract/Free Full Text]
- Lynch TCP, Cruses JV III, Morgan FW, et al. Bone abnormalities of
the knee: prevalence and significance at MR imaging.
Radiology 1989;171
: 761-766[Abstract/Free Full Text]
- Miller M, Osborne JR, Gordon WT, Hinkin DT, Brinker MR. The natural
history of bone bruises: a prospective study of magnetic resonance
imaging-detected trabecular microfractures in patients with isolated medial
collateral ligament injuries. Am J Sports Med1998; 26:15
-19[Abstract/Free Full Text]
- Escalas F, Curell R. Occult posttraumatic bone injury.
Knee Surg Sports Traumatol Arthrosc 1994;2
: 147-149[CrossRef][Medline]
- Kapelov SR, Teresi LM, Bradley WG, et al. Bone contusions of the
knee: increased lesion detection with fast spin-echo MR imaging with
spectroscopic fat saturation. Radiology1993; 189:901
-904[Abstract/Free Full Text]
- Stafford SA, Rosenthal DI, Gebhardt MC, Brady TJ, Scott JA. MRI in
stress fracture. AJR 1986;147
: 553-556[Free Full Text]
- Ryu KN, Jin W, Ko YT, et al. Bone bruises: MR characteristics and
histological correlation in the young pig. Clin
Imaging 2000; 24:371
-380[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
