| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, University of California, San Francisco, 400
Parnassus Ave., A-367, San Francisco, CA 94143.
2 School of Medicine, University of California, San Francisco, San Francisco,
CA.
Received January 22, 2008;
accepted after revision April 29, 2008.
Address correspondence to T. M. Link
(tmlink{at}radiology.ucsf.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. MRI studies obtained at 1.5 T were analyzed in 145 patients with pelvic insufficiency fractures. In 64 of 145 patients, MRI and multidetector CT (MDCT) findings were compared. Imaging studies were analyzed by two radiologists; combined clinical history, findings from all imaging studies, and follow-up imaging studies served as the standard of reference.
RESULTS. In the subgroup undergoing both imaging techniques, MRI detected 128 of 129 (99%) fractures in 63 of 64 (98%) subjects, whereas CT detected only 89 of 129 (69%) fractures in 34 of 64 (53%) subjects. In particular, fractures at the femoral head and acetabulum were better detected with MRI. In the complete population, two or more fractures were found in 70.3% (102/145) of patients, and 89.2% (33/37) of patients with pubic insufficiency fractures had concomitant fractures at other locations. In 63 of 145 (43.4%) patients, a previous malignancy was found; in only 93 of 145 (64.1%) patients, the leading symptom responsible for the MRI examination was pain.
CONCLUSION. This study showed that MRI was substantially better than CT in detecting insufficiency fractures. In addition, two or more insufficiency fractures were frequently present, typical fracture combinations were found, and insufficiency fractures were frequently associated with malignant disease.
Keywords: comparative studies • CT • MRI • pelvic insufficiency fractures • proximal femur insufficiency fractures
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MRI is sensitive in the detection of these fractures. However, MRI findings may not always show a fracture line, and abnormalities on MRI may sometimes be interpreted as metastatic disease if the radiologist is not familiar with the appearance of these imaging findings [1, 12]. An incorrect diagnosis may initiate invasive and costly procedures that could be harmful to the patient or that could misguide therapy. Malignant lesions are frequently suspected in patients who have undergone radiation therapy and chemotherapy of the pelvis. However, irradiated bone is especially weakened and prone to insufficiency fractures, as shown previously [13].
CT is also used as an imaging procedure for insufficiency fractures in the pelvis. CT has been shown to be sensitive in depicting fracture lines and in the assessment of biomechanical stability [14]. It may be the more specific technique to diagnose insufficiency fractures [12], but at present it is not clear which imaging technique is better suited to this purpose. It has also been found that insufficiency fractures tend to occur in multiple locations, and concomitant soft-tissue abnormalities are frequently shown [7, 15, 16]. However, exact data from a larger patient population are not available.
Thus, a better understanding of insufficiency fractures of the pelvis is required, and further study is needed to determine whether CT or MRI is better suited to investigate these fractures. The aims of our study were to compare the sensitivities of CT and MRI in detecting insufficiency fractures of the pelvis and proximal femur; to analyze the typical location, morphology, and combinations thereof in these fractures; to analyze typical imaging morphology; and to analyze associated clinical findings, such as prior radiation therapy or chemotherapy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As a criterion for the comparison between CT and MRI, we restricted analysis to studies obtained within 3 months of the first cross-sectional study that showed the fracture. The rationale for this cutoff value was the altered bone metabolism in patients with insufficiency fractures that also induced delayed fracture healing, with a healing period of at least several months [17, 18]. In all patients, the clinical history was analyzed and potential risk factors for insufficiency fractures were recorded.
MRI
In all 145 subjects, MRI studies obtained at 1.5 T (Signa, GE Healthcare)
were available. In 125 patients, MRI of the pelvis was performed; and in 20
patients, only MRI of the lumbar spine, including the complete sacrum, was
available. The standard pelvis protocol consisted of coronal T1-weighted fast
spinecho (TR/TE, 600 milli seconds/minimum) and STIR (3,000/68; inver sion
time, 150 milliseconds) sequences as well as axial T1-weighted (600
milliseconds/minimum) and fat-saturated T2-weighted fast spin-echo (3,000/68)
sequences. Slice thickness was 4 mm; matrix size, 192 x 192 mm; and
field of view, 32–36 cm. The pelvis protocol covered the proximal femur
at least down to the lesser tro chanter and was used also for suspected hip
fractures. In the 20 patients in whom only lumbar spine MRI was performed,
sagittal and axial T1-weighted (TR range/TE, 500–600
milliseconds/minimum) and fat-saturated T2-weighted fast spin-echo (TR
range/TE range, 3,500–4,000/60–90) sequences as well as a coronal
T1-weighted fast spin-echo (500 milliseconds/minimum) sequences were obtained.
Slice thickness was 4 mm; matrix size, 256 x 192 mm; and field of view,
16–24 cm. Only studies that covered the whole sacrum were included.
CT
In 64 of 145 patients, CT studies performed with one of several MDCT
scanners (8-, 16-, and 64-MDCT scanners, LightSpeed Series, GE Healthcare)
were available in addition to MRI. Imaging included the entire pelvis in all
studies. Thirty-eight of the 64 (59.49%) CT studies were performed with IV
contrast material. The slice thickness ranged from 1.25 to 7 mm; 25 studies
were performed with a slice thickness of < 5 mm and 39 studies, with a
slice thickness of 
5 mm. All studies were performed with 120 kVp and
milliampere values ranging from 150 to 300 mA.
Radiography
In a subset of patients, anteroposterior pelvic and hip radiographs were
also analyzed. These were obtained using a standard radiographic technique,
placing the patient on the Bucky table, and using a grid, whenever possible,
with standard exposure parameters of 20 mAs and 75 kVp.
Image Analysis
The MRI, CT, and radiographic studies were analyzed separately in random
order by two radiologists in consensus. All patients underwent MRI and were
assessed for the absence or presence of a fracture as well as number and
location of fractures. In addition, the absence or presence of a bone marrow
edema pattern and fracture lines was documented. A bone marrow edema pattern
was classified as severe if signal intensity on STIR or fat-saturated
T2-weighted fast spin-echo images was similar to that of spinal fluid or urine
in the bladder, as mild if the bone marrow edema pattern was visualized only
along the fracture line (within 5 mm diameter) but not in the periphery, and
as moderate if the bone marrow edema pattern was visualized > 5 mm around
the fracture line or had a diameter of > 10 mm if no fracture line was
present and signal intensity was less than that of spinal fluid and urine. A
fracture line was defined as a linear structure visualized on either STIR,
T1-weighted, or fat-saturated T2-weighted fast spin-echo sequences. Location
of soft-tissue abnor malities was documented, and severity was graded as
severe (signal abnormality affecting the whole muscle compartment and fluid
signal was noted in the muscles on STIR or fat-saturated T2-weighted fast
spin-echo images), moderate (signal abnormality affecting muscles but signal
intensity less than that of fluid), or mild (signal abnormality around muscles
in adjacent soft tissues and only minimal muscle signal abnormality). Subacute
hematoma was documented if signal abnormality was high on T1-weighted images
(indicating the presence of methemoglobin).
CT examinations were studied for absence or presence of a fracture as well as number and location of fractures. Additional documented findings included presence or absence of a fracture line, focal area of sclerosis, adjacent radioluc ency consistent with resorptive bone changes, and vacuum phenomenon. Presence or absence of soft-tissue lesions was also documented. Signs of a soft-tissue lesion included muscle swelling, fat-tissue stranding, and active extravasation of contrast material on contrast-enhanced studies. Because of the limited visualization of soft-tissue abnormalities on CT, we did not differentiate grades.
For each of the two imaging techniques, confidence of fracture diagnosis was graded on a scale of 1–3, where 1 represented a definite fracture; 2, a probable fracture; and 3, no definite fracture.
The definition of a standard of reference for a study in which no surgical procedures or patho logic examinations are performed is challenging. In our study, we used all available imaging data and imaging follow-up, together with the clinical history, as a standard of reference. The diagnosis of an insufficiency fracture was supported by CT follow-up studies that showed fracture lines more clearly and areas of increased sclerosis or periosteal bone formation; and by MRI follow-up studies that showed persistent or decreasing areas of bone marrow edema or better visualization of fracture lines.
Data Analysis
All statistical analyses were performed using statistical software S-plus
(Insightful) and SAS software (SAS). Basic descriptive statistics were used.
Means and SDs were calculated for continuous variables, and frequency and
proportion for categoric variables. Fracture detection rates for CT and MRI
were calculated using the previously described standard of reference.
Sensitivities were calculated as the percentage of patients identified by each
technique based on this standard of reference, and confidence intervals were
provided. We used binomial exact confidence intervals, also known as Clopper
and Pearson confidence intervals. The significance of dif ferences between
individual groups was calculated using the Student's two-tailed t
tests for continuous variables and the chi-square test for categoric
variables, with a 5% level of significance. To show that no significant
selection bias existed between the subgroups of patients who underwent both CT
and MRI versus the subgroup who underwent only MRI, we compared the covariates
such as age, sex, fracture location, and number of fractures between those in
each subgroup. No statistically significant differences were seen between the
two groups in any of these covariates. Student's two-tailed t tests
and Fisher's exact tests with a 5% level of significance were used to compare
known fracture risk factors and patient characteristics between patients in
the two subgroups. The sensitivities of CT and MRI were compared using a
paired proportion test (the McNemar test).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
The most common fractures were located at the sacrum (n = 67) and the pubic ramus (n = 29) (Figs. 1A, 1B, 1C, 1D, 2A, 2B, 2C, 3A, 3B, 3C). Table 1 shows the number of fractures at each fracture site as well as the percentage of fractures detected at the location for each imaging technique. CT showed all ischial fractures. However, at all other sites, 25–57% of fractures were missed with CT, with the highest percentage of missed fractures at the femoral head. In addition to determining which technique performed better in visualizing fractures, we also compared the two methods for depiction of fracture morphology. Of the 88 fractures diagnosed with both imaging techniques, 32 of 88 (36.4%) fractures were better depicted with MRI, 26 (29.5%) better with CT, and 30 (34.1%) fractures were visualized equally well with both techniques. Figure 2A, 2B, 2C shows an example of a pubic fracture better shown on CT, whereas the sacral fracture in Figure 3A, 3B, 3C is better visualized on MRI.
|
|
|
|
|
|
|
The ability of both imaging techniques to detect fracture lines was compared. Of the 128 fractures detected on MRI, fracture lines were present in 122 (95.3%) cases, and no fracture lines were most often found at the sacrum (n = 4). With CT, a similar percentage of fracture lines was shown (78/89, 89.7%). Again, cases with no fractures lines were most often located at the sacrum (n = 6); and in all of these cases, increased sclerosis was present. Overall in the CT studies, sclerosis was present in 69 of 89 (79.3%) fractures.
Comparisons were also drawn between MRI and CT in the ability to detect soft-tissue abnormalities. Fifty-seven of the 64 patients had at least one soft-tissue lesion, for a total of 103 lesions. MRI detected 102 of 103 lesions (99%), whereas CT identified only 13 of 103 lesions (12.6%). MRI missed one soft-tissue lesion at the erector spinae muscles because of incomplete fat saturation.
Additional comparisons were made between thin-section (slice thickness <
5 mm) and thick-section (slice thickness 5 mm) CT studies. Twenty-five CT
studies were obtained with a slice thickness of < 5 mm, and of the 56 total
fractures diagnosed, 76.8% (43/56) were detected. In the 39 thick-section CT
studies, 73 total fractures were diagnosed and only 65.7% (48/73) were
detected. However, statistical analysis proved this difference to be
insignificant (p = 0.243).
MRI and CT Versus Radiography
Forty-eight patients had radiographs available as well as MR images. Of the
108 fractures diagnosed with MRI, radiography visualized only 16 (14.8%). Most
notably, only 3.8% (2/53) of sacral fractures and only 11% (1/9) of acetabular
fractures were detected. Thirty-one patients had CT scans and radiographs
available, and of the 63 fractures visualized with CT, radiographs visualized
only 13 (20.6%). Again, sacral fractures were shown infrequently (6.1%,
2/33).
Overall Characteristics of Insufficiency Fractures Based on MRI Findings
Table 2 lists locations of
all 307 fractures found in the 145 patients. Fractures averaged 2.1 per
patient (range, 1–6 fractures per patient). In 102 of 145 patients
(70.3%) more than one fracture was found. Figure
4A,
4B,
4C,
4D shows an example of a
patient with multiple insufficiency fractures at both femoral necks and the
ischia. Many fractures were frequently associated with concomitant fractures
in typical locations (Table 3).
In the case of pubic fractures, 89.2% (33/37) of patients with pubic fractures
had one or more fractures at another site. In total, 57 fractures occurred
with pubic fractures. The sites most commonly associated with pubic fractures
were the sacrum and the acetabulum, with 40 (70.1%) and nine (15.8%)
concomitant fractures, respectively. Other frequent combinations of
concomitant fractures were noted. Twenty-five patients were diagnosed with
acetabular fractures, and in 19 (76%) of these patients concomitant fractures
were present. The sacrum (n = 13), pubis (n = 12), and femur
(n = 11) were sites of 36 of the 37 fractures that occurred with
acetabular fractures. Forty percent of the 25 patients with femoral fractures
had concomitant fractures. In total, 31 fractures occurring with femoral
fractures were found, which occurred mostly at the sacrum (13/31, 42%) and the
acetabulum (10/31, 32.3%).
|
|
|
|
|
|
Most frequently, insufficiency fractures were characterized by the presence of a bone marrow edema pattern and fracture lines (267/307, 87%). In 21 of 307 cases (6.8%), only fracture lines were shown; and in 19 of 307 (6.2%) cases, only a bone marrow edema pattern was found. Fracture lines were better shown on T1-weighted sequences, and proximal femur fracture lines were best shown on sequences obtained in a coronal orientation. While a bone marrow edema pattern alone was more frequently found at the sacrum (10/162), only one fracture at the proximal femur showed no fracture line (1/49). Bone marrow edema pattern was also graded in all fractures in which it was present. A bone marrow edema pattern was diagnosed in 286 of 307 fractures (93.2%). Of note, severe bone marrow edema patterns were found to be frequently associated with pubic fractures (15/54, 27.8%), whereas a moderate bone marrow edema pattern was most frequently associated with sacral fractures (79/172, 48.8%). A mild bone marrow edema pattern was more frequently associated with femoral fractures (18/49, 36.7%).
Femoral, acetabular, and pubic fractures had the highest proportion of associated soft-tissue abnormalities. Figure 5 illustrates the total number of fractures and locations as well as the number of fractures associated with soft-tissue abnormalities at each location and displays the proportion of fractures with soft-tissue abnormalities at a certain location. Note that, although a high percentage of soft-tissue edema was found with pubic and acetabular bone fractures, it was less frequently seen with sacral fractures. The amount and intensity of soft-tissue edema were also graded higher in patients with pubic, acetabular, and femur fractures versus sacral fractures. Subacute hematoma was documented in only two of 145 patients.
|
Associated Clinical Findings
Interestingly, insufficiency fractures were associated with prior
malignancy in 63 of 145 (34.4%) patients. Previous radiation therapy to the
pelvis was noted in 29 of 145 (20%) patients, and previous chemotherapy, in 24
of 145 (16.6%) patients (Fig.
3A,
3B,
3C). In 29 of 63 patients
(46.0%), the primary tumor was located in the pelvis; 24 of these patients had
gynecologic malignancies. In nine of 63 patients (14.3%), hematologic
malignancies were recorded, and in six (9.5%) breast cancer was diagnosed. In
six of 145 patients (4.1%) organ transplantation had been performed (two heart
and two liver transplants), and in eight of 145 patients (5.5%) chronic renal
failure was recorded. Rheumatologic diseases, such as systemic lupus
erythematosus and rheumatoid arthritis, were found in six of 145 patients
(4.1%). A history of corticosteroid therapy within the previous 6 months was
recorded in 18 of 145 (12.4%) patients. In only 93 of 145 (64.1%) patients,
the chief symptom leading to the MRI examination was pain.
In addition to underlying clinical conditions, age was also a major contributor to fracture risk. The number of fractures increased with age; most of the fractures were found in patients 70–89 years old (146/307) compared with 50 of 307 and 83 of 307 at age groups 30–49 and 50–69 years, respectively. Also, the number of fractures in an individual patient increased with age. On average, patients with sacral fractures were older (67.6 ± 17.5 years) than those with femoral fractures (61.2 ± 19.8 years), yet differences were not significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Only a small number of studies have been performed comparing MRI and CT in diagnosing stress fractures. Grangier et al. [3] compared MRI and CT findings in nine patients with insufficiency fractures of the sacrum and acetabular roof and found that three fractures were missed on CT. Those authors also recommended MRI as the best diagnostic test, comparing it not only with CT but also with bone scintigraphy and radiography. The limitation of that study was that thin-slice MDCT with multiplanar reformations was not available at the time the study was performed; only 4- to 8-mm axial images with sagittal reconstructions were available. Gaeta et al. [19] examined stress fractures at the tibia in 42 athletes and found a higher sensitivity for MRI than for CT (88% vs 42%) in diagnosing stress injuries. However, comparison with our study is limited because those investigators examined fatigue fractures, which have different bone mass and quality. Staatz et al. [20] reported six patients with osteoporotic sacral insufficiency fractures; although all six were visualized with MRI, only five were shown with CT. Those authors recommended gadolinium-enhanced MRI to better differentiate insufficiency fracture from malignant disease and to exclude intraosseous tumor formation. In our study, we used both STIR and fat-saturated T2-weighted sequences, which are sensitive to fractures and the bone marrow edema pattern, but we did not use gadolinium-based contrast agents. Because a number of patients with insufficiency fractures have limited renal function—such as older patients and cancer patients undergoing chemotherapy—the risk is increased that gadolinium will induce nephrogenic systemic fibrosis, so standard contrast administration is not recommended at our institution.
In addition to the better visualization of fractures, MRI was also more sensitive than CT in detecting soft-tissue abnormalities, as has also been shown by Feydy et al. [21] in 15 patients with longitudinal stress fractures at the tibia. However, those authors concluded that MRI, although more sensitive than CT, may be misleading; they controversially recommended CT as the better imaging technique. However, this recommendation may apply for the tibia in fatigue fractures, but not for pelvic insufficiency fractures if the radiologist is familiar with typical imaging findings. Interestingly, in our study the presence and severity of soft-tissue edema varied according to the location of the fractures: Although a high percentage of soft-tissue edema was found with pubic and acetabular fractures, it was less frequently seen with sacral fractures. The amount and intensity of soft-tissue edema give additional information on the extent of injury and may have clinical implications relating to the degree of pain and disability as well as indicating the prognosis of an insufficiency fracture.
In 70.3% of cases, multiple fracture sites were present. Caution should be exercised when one site with an insufficiency fracture is discovered because other fracture sites likely exist. This is especially true of pubic fractures; 90% of the pubic fractures in this study had concomitant fractures present. Also, 76% of acetabular fractures had concomitant fractures.
Interestingly, fractures at the femoral head and acetabulum were better detected with MRI, which may be because these fractures are frequently horizontally aligned in the direction of the axial CT plane, and even thin-section reformations will be subject to partial volume effects and will not necessarily reveal fractures.
Our results clearly show that MRI is the superior imaging technique for visualizing insufficiency fractures of the pelvis and associated soft-tissue abnormalities when the clinical status suggests a fracture in the setting of negative radiographs. Chen et al. [22] argued that CT was better because of cost, diagnostic rate, and ease of fracture line visualization. However, consideration should be given to the implications of a missed diagnosis that may misdirect or delay treatment. Treatment in pelvic insufficiency fractures varies according to the affected site; although sacral insufficiency fractures are usually treated with analgesia and bed rest, acetabular and femur fractures may require surgical therapy. In particular, proximal femur neck and intertro-chanteric fractures are prone to substantial complications and displacement if they are not surgically revised [23–25], which has implications for clinical management and recommendations from radiologists to referring physicians, particularly when radiographs are negative and there is a strong clinical suspicion of a fracture. If proximal femur fractures are suspected, mobility of the patient should not be permitted before cross-sectional imaging has been obtained.
Limitations of our study include its retrospective study design and that a relatively long time elapsed between the CT and MRI studies in some patients. However, because of the limited metabolic activity of bone in these patients, imaging features of fractures are expected to persist over a period of several months, as has been shown by previous studies [17, 18]. Also, we did not have a true reference standard for insufficiency fractures because to do so would have required either surgery or autopsy, neither of which was feasible in our patients. The standard of reference we used—longitudinal imaging, clinical follow-up, and all other available imaging studies over several years—is the best available alternative. That only insufficiency fractures were included from the beginning, and that the sensitivities of CT and MRI can only be estimated, are other limitations, yet these are limitations of the retrospective study design, and the sensitivity of these techniques in a typical clinical scenario seems most important.
In addition, bone scintigraphy and dual x-ray absorptiometry (DXA) were not included in our analysis, which may be considered a limitation of the study. However, only a small subset of patients underwent radionuclide and DXA studies, which prevented a statistically representative analysis. According to a previously published study [8], MRI is more sensitive than bone scintigraphy in the early stages of insufficiency fractures. Because most of the patients in this study were elderly and a large number underwent chemotherapy or radiation therapy, thus reducing bone metabolism, bone scintigraphy may have had limited sensitivity. In addition, MRI and CT usually can be obtained faster, particularly in an emergency setting, and cross-sectional techniques are superior in differentiating insufficiency fractures from tumor invasion or metastatic disease.
In conclusion, because insufficiency fractures are increasingly frequent, important yet underdiagnosed lesions, the radiologist must be familiar with these findings and their location, morphology, and associated clinical features. MRI is the superior technique compared with CT and should be the imaging technique of choice in patients with suspected insufficiency fractures. This is currently the established standard at our institution and is requested by our orthopedic surgeons in patients with suspected fractures and negative radiographic findings. Radiologists also must be aware that multiple pelvic insufficiency fractures are frequently found, particularly in the presence of a pubic or an acetabular fracture. Consequently, the entire pelvis must be imaged in these patients, and a careful search for concomitant fractures is always warranted.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. A. Sankey, J. Turner, J. Lee, J. Healy, and C. E. R. Gibbons The use of MRI to detect occult fractures of the proximal femur: A STUDY OF 102 CONSECUTIVE CASES OVER A TEN-YEAR PERIOD J Bone Joint Surg Br, August 1, 2009; 91-B(8): 1064 - 1068. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
