| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Thomas Jefferson University Hospital, 111 S. 11th
St., Rm. 3390 Gibbon, Philadelphia, PA 19107.
2
Present address: FMH Diagnostische Radiologie, Oberer Batterieweg 57, 4059
Basel, Switzerland.
Received April 25, 2001;
accepted after revision August 1, 2001.
Address correspondence to H. P. Ledermann.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. One hundred ten contrast-enhanced 1.5-T MR imaging foot examinations in 102 patients (28 women, 74 men; mean age, 59 years), 82% of whom had diabetes mellitus, were reviewed by two musculoskeletal radiologists for the presence of areas without recognizable enhancement. The number, size, location, signal characteristics, and enhancement ratio of nonenhancing regions were noted. MR imaging findings were compared with surgical and histology reports.
RESULTS. Nonenhancing regions were found in 27 feet (24.5%, 96.3% in diabetic patients, p = 0.032) at the forefoot (n = 16), toes (n = 8), and heel (n = 3). The mean size of the nonenhancing regions was 4.1 x 2.7 x 1.4 cm. Signal characteristics on T1-weighted images were isointense to muscle (n = 21, 77.8%), hypointense to muscle (n = 3, 11.1%), heterogeneous (n = 2, 7.4%), and isointense to fat (n = 1, 3.7%). On T2-weighted images, the signal was hyperintense to muscle (n = 12, 44.4%), heterogeneous to muscle (n = 9, 33.3%), equal to fluid (n = 3, 11.1%), and hypointense to muscle (n = 3, 11.1%). The mean signal increase after contrast administration was 3.57% for observer 1 and 2.68% for observer 2. Necrotic tissue was surgically confirmed in the nonenhancing areas in 26 feet (96.3%). Five abscesses and three cases of osteomyelitis were misdiagnosed on MR images because of lack of enhancement.
CONCLUSION. Nonenhancing areas are seen in one fourth of pedal infections, occur almost exclusively in diabetic patients, and represent necrotic tissue. Only contrast-enhanced images allow reliable recognition of these regions. Lack of enhancement in these areas can mask the presence of abscess and osteomyelitis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Pedal infection has been extensively investigated using MR imaging [5,6,7]. The exact delineation of soft-tissue infection and osteomyelitis on MR images was shown to influence the choice of further therapy [8, 9] and to help in planning targeted minimal surgical intervention [6, 8,9,10]. However, to our knowledge no prior reports address the potential of MR imaging to identify necrotic tissue.
Reviewing MR images of pedal infections in daily practice, we have anecdotally observed the presence of relatively sharply defined areas of soft tissue that did not show visible enhancement. The goal of this study was to analyze the frequency, size, and location of nonenhancing soft-tissue areas on MR images in pedal infection and to investigate their clinical correlation. Because osteomyelitis may be masked at bone scintigraphy as a result of severely compromised vascular perfusion [11,12,13], we also evaluated signal characteristics and enhancement patterns of abscesses and osteomyelitis occurring in regions of nonenhancement to determine if similar findings are seen on MR images.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MR imaging was performed with a 1.5-T scanner (Signa; General Electric Medical Systems, Milwaukee, WI). An extremity coil was used in the evaluation of 106 feet in 98 patients (field of view, 14-20 cm), and a head coil was used for the examination of both feet in four patients (field of view, 16-20 cm). Images of all feet were obtained in at least two orthogonal planes.
T1-weighted spin-echo images were obtained with one to two signals averaged, a TR range/TE range of 400-750/10-20, and a matrix size of 256 x 192 or 256 x 256. T2-weighted images were obtained using a fast spin-echo technique with two signals averaged, an echo-train length of 8, TR range/TE range of 2000-7800/75-108, and a matrix size of 256 x 128 or 256 x 192. Contrast-enhanced fat-suppressed T1-weighted images were obtained in all 110 feet using a fast multiplanar spoiled gradient-recalled echo technique. Fast multiplanar spoiled gradient-recalled echo sequences with fat suppression were obtained before and after contrast administration using the same parameters (TR/TE, 250/2.1; flip angle, 90°; matrix, 256 x 128 or 256 x 192). Imaging began approximately 2 min after the IV administration of gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) at a dose of 0.1 mmol/kg of body weight. Gadolinium-enhanced images were available for all feet. For all T2-weighted and gadolinium-enhanced T1-weighted sequences, fat suppression was accomplished using selective presaturation of lipid-resonant frequency. Fast spin-echo short tau inversion recovery (STIR) images were available for 98 feet. Fast spin-echo STIR images were obtained with an echo-train length of 8, TR range/TE range of 2000-3000/20-78, inversion time of 150-160 msec, and a matrix of 256 x 128 or 256 x 192.
MR imaging criteria for the diagnosis of osteomyelitis were based on those described in previous studies [6, 14, 15]: focally decreased marrow signal intensity on T1-weighted images, focally increased signal intensity on fat-suppressed T2-weighted and fast spin-echo STIR images, and focal marrow enhancement on gadolinium-enhanced fat-suppressed T1-weighted images. To augment diagnostic confidence, secondary signs of pedal osteomyelitis [16], including cutaneous ulcer, cellulitis, soft-tissue mass, soft-tissue abscess, sinus tract, and cortical interruption, were also applied. MR imaging criteria for the diagnosis of abscess were focally isointense or hypointense signal on unenhanced T1-weighted images, focal fluid signal on T2-weighted images, and rim enhancement on contrast enhanced fat-suppressed T1-weighted images [17,18,19].
Two experienced musculoskeletal radiologists, working in consensus, reviewed all studies. Both reviewers knew that the studies were performed in patients with foot infections but they were unaware of any other clinical data. All sequences in a single patient were reviewed at the same time on a PACS (picture archiving and communication system) workstation (Canon Medical Systems, Irvine, CA). Incidence, size, location, and signal characteristics of nonenhancing regions were noted. For purposes of describing the location of nonenhancing tissue, the foot was divided into the following four compartments: toes, forefoot (from the metatarsophalangeal joint to Lisfranc's joint), midfoot (from Lisfranc's joint to Chopart's joint), and hindfoot (from Chopart's joint to the ankle joint). In the forefoot and in the toes, the exact location of nonenhancing tissue was additionally defined by the involved toes and rays and by the dorsal, plantar, medial, or lateral location. Each observer measured the signal increase of every soft-tissue region identified as nonenhancing on the workstation. Two to three rectangular regions of interest (mean diameter, 9 mm; minimum, 6 mm; maximum, 14 mm) were placed in muscle tissue in these nonenhancing areas by each reviewer on fast multiplanar spoiled gradient-recalled echo images obtained before and after contrast administration (Figs. 1A and 1B). A standard of reference more proximally in the plantar muscle tissue was also measured with a rectangular region of interest (mean diameter, 10 mm; minimum, 7 mm; maximum, 16 mm) before and after contrast administration in all patients. For each pair of unenhanced and contrast-enhanced regions of interest, the percentage of signal intensity increase was calculated, and the mean percentage of enhancement was determined using the following formula: [(SIpost — SIpre) / SIpre] x 100%, where SI is signal intensity and "pre" and "post" are, respectively, before and after the injection of gadopentetate dimeglumine.
|
|
Statistical comparison of the relative signal increase after contrast administration of the nonenhancing areas with the reference areas more proximally was achieved using a two-tailed Student's t test for both observers. Statistical analysis of the prevalence of diabetes mellitus in patients with nonenhancing areas and the prevalence of diabetes mellitus in the entire study group was performed using a chi-square test.
After the MR examination, bone samples were collected in all feet either by biopsy or by surgery to evaluate for osteomyelitis. The following procedures were performed after the MR examination: toe amputation (n = 29), débridement (n = 29), ray amputation (n = 22), below-knee amputation (n = 13), transmetatarsal amputation (n = 7), Chopart's amputation (n = 5), percutaneous bone biopsy (n = 3), foot amputation (n = 1), and above-knee amputation (n = 1). Pathology results from bone samples were available in all patients, and microbiology results were available for 94 examinations. The locations of the three bone biopsies were chosen by a musculoskeletal radiologist after MR evaluation. All other patients had intraoperative bone samples as a gold standard for the presence of osteomyelitis. In the 107 surgical procedures, the areas of biopsy were chosen by the surgeon. A diagnosis of osteomyelitis was defined with culture growth or characteristic histologic findings of osteomyelitis. These histologic findings were aggregates of inflammatory cells (including neutrophils, lymphocytes, histiocytes, and plasma cells), erosions of trabecular bone, and marrow signal changes that ranged from loss of normal marrow fat with acute osteomyelitis to fibrosis and reactive bone formation with chronic disease.
Review of the charts and surgical reports was performed for all patients by a musculoskeletal radiologist (research fellow) who was not involved in image interpretation. We noted for each patient the location of, type of, and indication for surgery; results from pathology and microbiology reports; and associated conditions such as diabetes mellitus and peripheral vascular disease. We also reviewed whether patients had antibiotic treatment before bone biopsy, the duration of the antibiotic treatment, and whether antibiotic therapy was stopped before bone biopsy.
In patients with nonenhancing tissue, we noted what kind of surgical procedure was performed after MR imaging, the location of the surgical intervention, and whether resection of necrotic tissue was described in a location corresponding to nonenhancing tissue on MR imaging. The results of bone biopsies (histology and cultures) from bones that were located in nonenhancing tissue were noted. We also reviewed the patients' charts, indications for surgery, and pathology reports (macropathologic description of amputation samples) to evaluate whether the presence of established gangrene was described in a nonenhancing region. We furthermore noted in patients with nonenhancing areas whether abscesses were encountered during surgery in the areas of nonenhancement. If an abscess or osteomyelitis was not diagnosed on initial review of MR images but was seen at surgery (abscesses) or histologic or microbiologic examination (osteomyelitis), we reevaluated the corresponding MR images to analyze the signal characteristics of the misinterpreted areas.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The size of nonenhancing areas ranged from 9 x 5 x 3 cm to 1
x 0.8 x 0.5 cm with a mean size of 4.1 x 2.7 x 1.4 cm.
Sixteen areas were seen in the forefoot (59%), eight were located in the toes
(30%), and three in the hindfoot (11%). T1-weighted images displayed the
following signal characteristics in nonenhancing regions: isointense to muscle
(n = 21, 77.8%), hypointense to muscle (n = 3, 11.1%),
heterogeneous signal (n = 2, 7.4%), and isointense to fat (n
= 1, 3.7%). On fat-suppressed T2-weighted images, the following signal
characteristics were observed: hyperintense to muscle but less than fluid
(n = 12, 44.4%), mixed hyper- and hypointense (n = 9,
33.3%), equal to fluid (n = 3, 11.1%), and hypointense to muscle
(n = 3, 11.1%). Comparing contrast-enhanced and unenhanced
fat-suppressed images, the mean signal increase in these areas was 24.20
± 47.74 (3.57% ± 5.59%) for observer 1 and 22.45 ± 41.53
(2.68% ± 3.96%) for observer 2. The nonenhancing areas displayed a
homogeneous low signal intensity after contrast enhancement without visible
differences among the signals of muscle tissue, soft tissue, or bone marrow.
The reference areas in muscle tissue more proximally had a signal increase of
287.5 ± 197 (28.1% ± 11.2%) for observer 1 and 389.4 ±
397.5 (27.4% ± 16.8%) for observer 2. The difference in enhancement
between the nonenhancing areas and the standard of reference more proximally
was statistically significant (p < 0.0001) for both observers. All
27 areas in which no contrast enhancement was visually perceived had a
measured relative signal increase of less than 11.4% after contrast
administration, and 90% of areas enhanced less than 7%. Most of these areas
displayed an abrupt cutoff of enhancement at the border of enhancing tissue,
which often resulted in a geographic demarcation of nonenhancing tissue (Figs.
1C,2C,3D,4C
and 4D). This clearly defined
border between enhancing and nonenhancing tissue was often accentuated on
contrast-enhanced images by rim enhancement of surrounding viable tissue
(Figs.
1C,2C,3D,4C,
and 4D). Because most areas
were clearly defined by their abrupt cutoff to enhancing tissue and the
surrounding rim enhancement, no disagreement existed between the reviewers
concerning the presence of nonenhancing
regions.
,
|
|
|
|
|
|
|
Twenty-six feet (96.3%) with devitalized (nonenhancing) areas on MR images had a surgical intervention within 45 days of MR imaging. The following procedures were performed as primary surgical treatment: toe amputation (n = 8), débridement (n = 6), ray amputation (n = 6), below-knee amputation (n = 3), transmetatarsal amputation (n = 2), and above-knee amputation (n = 1). Eleven patients needed revascularization procedures. Ten patients underwent additional revision operations after primary surgical treatment including débridement (n = 8) and further amputation (n = 4), resulting in a mean of two procedures per patient within 45 days of MR imaging. Resection of necrotic tissue in areas of nonenhancement was confirmed in the surgical reports of 18 patients. Correct local correlation between the areas of nonenhancement and the location of necrotic tissue at surgery was ensured by the fact that toe amputations, ray amputations, and all débridement were targeted, limited procedures at a clearly defined area of the foot that could be easily compared with the MR images. The relatively large mean size of the nonenhancing areas (4.1 x 2.7 x 1.4 cm) also helped to localize and to match locations on MR images and at surgery. In the other eight patients, either a procedure was performed without exploration of the necrotic area or intraoperative findings were not described in the surgical report. However, the indication for surgery in these eight patients was clinically confirmed gangrene in the area of nonenhancement. Four of these eight patients had an amputation of the lower leg, with a precise macropathologic description of the presence and location of necrotic tissue in the amputation samples. One patient did not have surgery in the nonenhancing area on the heel but had a toe amputation instead.
Review of the patients' charts revealed that 82 patients had antibiotic treatment before surgery or bone biopsy. Cessation of antibiotic treatment for at least 3 days before surgery or bone biopsy to avoid interference with bone cultures was documented in 47 (55%) of 85 procedures. Including only the 84 bone samples of patients without antibiotic treatment (n = 25) and the samples of the patients with a documented cessation of antibiotic therapy before intervention (n = 47) yielded the following results for the diagnosis of osteomyelitis on MR images: true-positive (n = 48), true-negative (n = 26), false-positive (n = 5), and false-negative (n = 5). These findings resulted in a sensitivity of 91% and a specificity of 84%.
In the entire study group, seven MR diagnoses of osteomyelitis were proven to be false-negative by intraoperative biopsy. The interval between MR imaging and biopsy or surgery in all seven patients was less than 21 days. Four of these seven bones had no enhancement after gadolinium administration. Three of these four bones were in a nonenhancing area of soft tissue (Fig. 3A,3B,3C,3D). The diagnosis of osteomyelitis was confirmed in two patients by positive findings on bone cultures, and all had positive findings on the histology report. In the other three bones with a false-negative diagnosis of osteomyelitis, contrast enhancement was present but was initially misinterpreted. The three nonenhancing infected bones in devitalized (nonenhancing) areas all had normal fat marrow signal on T1-weighted images, and only two displayed a hyperintense signal on T2-weighted images.
|
|
|
A total of 12 abscesses were found on the 110 MR examinations. Three (11%) of these 12 abscesses were seen in the group of 27 feet in nonenhancing areas. However, review of the surgical reports of these 27 feet revealed that the surgeons encountered five additional abscesses in the nonenhancing regions that were not initially seen on MR imaging. Even retrospective analysis of these MR examinations did not show typical rim enhancement (Fig. 4A,4B,4C,4D). On unenhanced T1-weighted images, all areas were isointense to muscle. On T2-weighted images, two areas were hyperintense to muscle, two were heterogeneous, and only one area had a signal equivalent to that of fluid.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MR imaging is increasingly used preoperatively to define the extent of deep soft-tissue pedal infection because it enables a targeted surgical approach, thereby limiting damage to surrounding tissue [8,9,10]. However, the potential diagnosis of necrotic tissue on MR imaging has so far not been reported. In fact, evaluation of the extent of necrosis is presently still performed intraoperatively, and definitive treatment depends on intraoperative findings.
All nonenhancing regions but one in our patient group were operated on, and necrotic tissue was confirmed in all these areas either by the intraoperative report or by the indication for surgery. Therefore, visual identification of a relatively sharply demarcated area of nonenhancement was a reliable marker for established gangrene or necrosis in our population. Measurements of relative signal increase on contrast-enhanced versus unenhanced MR images on a workstation are, in our opinion, not needed to confirm necrosis. The mean percentage of signal increase in necrotic areas was less than 5% for both our observers, although values up to 11.4% were recorded.
The lack of recognition of necrosis by earlier reports on MR imaging of pedal infections can be explained by the fact that most patients in those studies were examined without gadolinium administration. As has been previously reported, the diagnosis of tumor necrosis cannot be reliably accomplished by unenhanced T1- and T2-weighted spin-echo sequences [28]. The signal characteristics of necrotic tissue in our patient collective on unenhanced T1- and T2-weighted sequences were quite variable. On T2-weighted images, the signals in nearly half these areas were hyperintense to muscle, one third had a mixed signal intensity, and approximately one tenth were fluid-equivalent. On T1-weighted images, most necrotic tissues were isointense to muscle, but other areas had mixed signal or were isointense to fat or hypointense to muscle. Only on gadolinium-enhanced images was it possible to reliably diagnose necrotic tissue, which usually showed an abrupt cutoff of enhancement at the border of enhancing tissue (Figs. 1C,2C,3D,4C, and 4D). This finding resulted in most cases in a confined and clearly defined geographic area of nonenhancement. We often observed a rim of enhancement around the necrotic tissue that may represent surrounding granulation tissue. The proximal extent of necrotic tissue was usually better displayed on sagittal or axial images (Figs. 1C and 4D).
Review of surgery and pathology reports in our study group revealed that osteomyelitis and abscesses cannot be reliably diagnosed in necrotic tissue on MR imaging. It is important that the radiologist recognize these limitations in diagnosing osteomyelitis and abscesses in nonenhancing areas. However, contrast-enhanced imaging is necessary to reliably identify these necrotic areas. Unenhanced T1-weighted images did not help to identify osteomyelitis in our three patients with nonenhancing osteomyelitis because the marrow signal was normal (Fig. 3A). T2-weighted images, on the other hand, may indicate infection when marrow hyperintensity is seen, as in two of our three patients. The scintigraphic literature indicates that severe ischemia may result in false-negative studies for osteomyelitis [11, 13]. However, this causative relationship has not yet been reported in MR imaging examinations. Because most feet with nonenhancing areas have severe peripheral artery disease and potentially delayed enhancement, it may be valuable in this group of patients to add delayed contrast-enhanced images to the protocol to diagnose osteomyelitis.
The surgeons in our series encountered five abscesses in necrotic tissue that were not diagnosed on preoperative MR images. We were not able, even retrospectively, to locate these abscesses because none of the areas contained a typical focal hyperintensity on T2-weighted images or a detectable mass-effect on T1-weighted images. Focal rimlike contrast enhancement suggesting an abscess was likewise absent in all five areas of abscesses in regions of necrosis (Fig. 4C). Therefore, it may be difficult or impossible to recognize established pus collections in necrotic tissue. Because most necrotic areas in our study presented with a surrounding rim of enhancement and had a hyperintense or even a fluid-equivalent signal on T2-weighted sequences similar to abscesses, in some cases distinguishing areas of necrosis from abscesses may be difficult.
All but one patient (96.3%) in our study group had surgical treatment after MR evaluation. The relatively high frequency of surgical procedures within 1 month of MR imaging in our patients with necrosis reflects the surgeons' attempts to salvage ischemic limbs with limited resection and revascularization and is comparable to other reports [1, 29]. Preoperative delineation of the extent of necrotic tissue on contrast-enhanced MR images may be helpful in such a therapeutic regimen to plan a targeted, limited resection. Ten patients in our study group required additional surgical intervention after primary treatment, including further débridement, amputation, and revascularization. It would be interesting to investigate in a prospective study whether preoperative mapping of the extent of necrosis on contrast-enhanced MR imaging can reduce the need for revision procedures.
Some limitations of our study need to be acknowledged. First, correlation with intraoperative findings was accomplished by review of surgical reports, which did not allow direct anatomic correlation of size and extent of necrotic tissue and abscesses. Eight cases had no surgical correlation, because intraoperative findings were not described in four surgical reports and four amputation specimens with distal necrosis were not dissected. Second, because we evaluated only MR images of patients who had surgery or biopsy after MR imaging for suspected osteomyelitis, a selection bias was introduced that may have resulted in a higher percentage of missed cases of abscesses and osteomyelitis than in a randomized population. Third, the goal of this study was to describe the frequency and distribution of nonenhancing tissue in pedal infection and to determine its clinical correlation and potential pitfalls in diagnosing abscesses and osteomyelitis. To determine the sensitivity and specificity of nonenhancing tissue in diagnosing established necrosis would require a prospective study.
In conclusion, approximately one quarter of patients with suspected pedal osteomyelitis have areas of nonenhancing tissue on MR images that correspond to regions of necrotic tissue. Such nonenhancing areas occur almost exclusively in diabetic patients (96.3%), and only contrast-enhanced images allow a reliable diagnosis of necrotic tissue. The presence of osteomyelitis and abscess can be masked in areas of nonenhancement.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  C. C. Roberts, W. B. Morrison, and P. T. Liu Imaging evaluation of foot and ankle pathology: self-assessment module. Am. J. Roentgenol., March 1, 2008; 190(3 Suppl): S18 - S22. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P L Tan and J Teh MRI of the diabetic foot: differentiation of infection from neuropathic change Br. J. Radiol., November 1, 2007; 80(959): 939 - 948. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kapoor, S. Page, M. LaValley, D. R. Gale, and D. T. Felson Magnetic Resonance Imaging for Diagnosing Foot Osteomyelitis: A Meta-analysis Arch Intern Med, January 22, 2007; 167(2): 125 - 132. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. V. Salgami, F. L. Bowling, R. W. Whitehouse, and A. J. M. Boulton Charcot Neuroarthropathy: An Unusual Case and a Review of the Literature International Journal of Lower Extremity Wounds, September 1, 2006; 5(3): 207 - 212. [Abstract] [PDF]
|
|
|
|
 |
|
 B. A. Lipsky, A. R. Berendt, H. G. Deery, J. M. Embil, W. S. Joseph, A. W. Karchmer, J. L. LeFrock, D. P. Lew, J. T. Mader, C. Norden, et al. Diagnosis and Treatment of Diabetic Foot Infections J Am Podiatr Med Assoc, March 1, 2005; 95(2): 183 - 210. [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Bloem, T. J. Mosher, M. R. Schmid, M. Zanetti, H. P. Ledermann, W. B. Morrison, and M. E. Schweitzer Dynamic Gadolinium-enhanced MR Imaging in Bone Marrow Disorders [letter] * Dr Mosher responds: * Drs Schmid and Zanetti respond: * Dr Ledermann and colleagues respond: Radiology, April 1, 2003; 227(1): 303 - 305. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
