| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Musculoskeletal Imaging |
1 Department of Medical Imaging, National Taiwan University Hospital, 7,
Chung-Shan S. Rd. Taipei, Taiwan.
2 Department of Radiology, National Taiwan University, School of Medicine,
Taipei, Taiwan.
Received October 31, 2003; accepted after revision April 5, 2004.
Address correspondence to T. T.-F. Shih
(ttfshih{at}ha.mc.ntu.edu.tw).
OBJECTIVE. The objective of our study was to define the MRI features of tuberculous infection of the wrist.
MATERIALS AND METHODS. We present the MRI findings of eight patients with tuberculous infection of the wrist. Spin-echo T1-weighted, gradient-echo T2*-weighted, and fast spin-echo T2-weighted sequences were performed for all patients. Gadolinium-enhanced MR images were obtained in seven patients. All images were evaluated for the characteristics of tuberculous infection of the wrist, including the presence of synovial thickening around the joints and tendons, signal intensity of the thickened tenosynovium and synovium on the T2-weighted images, synovial fluid collection in the tendon sheath, small low-signal and nonenhanced foci in the synovial fluid, bone erosion, osteomyelitis, and encasement of the median nerve.
RESULTS. The tuberculous infection involved the right (n = 6) and left (n = 2) wrists. All patients had synovial thickening around the flexor and extensor tendons with synovial fluid collection in the tendon sheath. The thickened tenosynovium and synovium revealed low signal intensity on T1-weighted images, intermediate to low signal intensity on T2- and T2*-weighted images, and enhancement on contrast-enhanced MR images. The synovial fluid showed intermediate to low signal on T1-weighted images and homogeneous or heterogeneous high signal intensity on T2- and T2*-weighted images. Multiple small foci of low signal intensity and nonenhancement scattered in the synovial fluid were present in seven patients. Bone erosion occurred in seven patients, osteomyelitis was seen in six patients, and encasement of the median nerve was found in three patients.
CONCLUSION. Characteristic MRI findings of tuberculous infection of the wrist include synovial thickening around the flexor and extensor tendons and synovial fluid collection that contains small low-signal and nonenhanced foci in the tendon sheath. Bone erosion, osteomyelitis, and median nerve encasement are also frequently present. These characteristic manifestations are helpful in diagnosing this disease entity.
Tuberculous infection of the musculoskeletal system comprises 10% of all extrapulmonary cases of tuberculosis [1]. Although musculoskeletal tuberculosis frequently affects the spine (51%), pelvis (12%), hip and femur (10%), knee and tibia (10%), and ribs (7%) [2], tuberculous involvement of the wrist is rare [3–6]. When it occurs, the patient may experience a variety of symptoms including local pain and swelling, limitation of motion, and soft-tissue mass [7–9]. Carpal tunnel syndrome secondary to median nerve compression or encasement may also occur [7, 10, 11].
The preoperative diagnosis of tuberculous involvement of the wrist is often difficult, particularly in the absence of a history of tuberculous infection. MRI has excellent contrast of bone and soft tissue and may aid in differentiation of this disorder from other conditions, such as inflammatory arthritis, pigmented villonodular synovitis, gout, or soft-tissue tumor. We report our experiences with eight cases of tuberculous infection of the wrist and discuss its characteristic MRI features.
Materials and Methods
We retrospectively reviewed the medical records with the diagnosis of tuberculous infection of the wrist during a 4-year period from January 1999 to December 2002. In total, 11 cases were diagnosed as tuberculous infection of the wrist. In eight of the 11 cases, both conventional radiographs and MRI studies of the wrist were obtained before surgery. Six men and two women who ranged in age from 33 to 90 years (mean, 65.5 years) formed the study group (Table 1). None of these patients had a history of tuberculous infection such as pulmonary tuberculosis before the wrist lesions developed. Coexisting diseases were also reviewed from the medical records to investigate the possibility of systemic diseases or other conditions that may deteriorate the immunity. Follow-up conventional radiographs of the chest were obtained after MRI examination in seven patients (cases 1–4 and 6–8).
|
MRI was performed on a 1.5-T magnetic unit (Signa, GE Healthcare; or Sonata, Siemens Medical Solutions) with a dedicated extremity coil. Our imaging protocol included spin-echo T1-weighted coronal and axial images (TR range/TE range, 450–650/14–18), gradient-echo T2*-weighted coronal images (450–650/18–24; flip angle, 30°), and fast spin-echo T2-weighted sagittal and axial images (TR range/effective TE range, 3,000–4,000/80–128; echo-train length, 7–8). Seven patients (cases 1, 2, and 4–8) had IV administration of gadopentetate dimeglumine (Magnevist, Schering) at a dose of 0.1 mmol/kg, and then fat-saturated spin-echo T1-weighted coronal images (450–650/14–18) and spin-echo T1-weighted axial images (450–650/14–18) without fat saturation were obtained.
The MR images were evaluated for the presence of synovial thickening around the joints and tendons, signal intensity of the thickened tenosynovium and synovium, synovial fluid collection in the tendon sheath, small low-signal and nonenhanced foci in the synovial fluid, bone erosion, osteomyelitis, and encasement of the median nerve. To evaluate the synovial enhancement, we chose the unenhanced T1-weighted sequence in the axial plane without fat saturation and the contrast-enhanced T1-weighted sequence in the axial plane without fat saturation for comparison. Bone erosion was defined as marginated bone lesion with juxtacortical localization and loss of normal dark signal intensity of cortical bone on all sequences. Osteomyelitis was defined as bone inflammation with evidence of decreased signal intensity of the bone marrow on the T1-weighted image; high signal intensity on T2-weighted images; and enhancement after IV gadopentetate dimeglumine administration, which was related to the inflammatory and hyperemic reaction of the bone. Median nerve encasement was defined as the median nerve encased by surrounding tissue in the carpal tunnel with poor identification of the nerve margin and palmar bulging of the flexor retinaculum.
Results
Medical records were reviewed for these patients to investigate coexisting diseases (Table 1), and we found diabetes mellitus in three patients, rheumatoid arthritis in one patient receiving long-term steroid therapy, and nasopharyngeal carcinoma with concurrent chemo- and radiotherapy in one patient. Tuberculous infection involved the right (n = 6) and left (n = 2) wrists. All these patients presented with local pain and swelling, limitation of motion, or soft-tissue mass for a durations of 3–24 months (mean, 10.5 months) (Table 1). Carpal tunnel syndrome was seen in three patients. Tenosynovectomy or synovectomy was performed in five patients (cases 1, 3, 4, 6, and 8), synovectomy with additional regional tumor excision was performed in two patients (cases 2 and 5), and local needle aspiration for diagnosis was performed in one patient (case 7). The diagnosis of tuberculous infection was confirmed by identification of granulomatous inflammation with Langhans' giant cells or acid-fast bacilli in the histopathologic specimen analysis or by the presence of Mycobacterium tuberculosis in the specimen culture (Table 1). "Rice bodies" were identified in the pathologic specimen in two patients (Table 1).
Conventional radiographs of the eight wrists revealed localized osteoporosis in all eight patients and bone erosion or destruction in seven patients (Fig. 1A). Densities for soft-tissue nodules or masses were seen in five patients (cases 2 and 5–8).
|
The MRI findings are summarized in Table 2. All patients had synovial thickening around the flexor and extensor tendons with synovial fluid collection in the tendon sheath (Figs. 1D, 1E, 1F, 1G, 2A, 2B, 2C and 3A, 3B, 3C). The thickened tenosynovium and synovium revealed low signal intensity on T1-weighted images and heterogeneous intermediate signal (n = 3), intermediate to low signal (n = 2), and low signal intensity (n = 3) on T2- and T2*-weighted images (Figs. 1B, 1D, 1F, 2A, 2B and 3A, 3B, 3C). The synovial fluid showed intermediate to low signal intensity on T1-weighted images (Fig. 2A). On T2- and T2*-weighted images, the synovial fluid revealed homogeneous or heterogeneous high signal intensity (Figs. 1B, 1D, 1F, 2B and 3A, 3B, 3C). Multiple small foci of low signal intensity scattered in the synovial fluid were present on T2- and T2*-weighted images in seven patients (Figs. 1B, 1D, 1F, 2B and 3A, 3B, 3C).
|
|
|
|
|
|
|
|
|
|
|
|
Gadolinium-enhanced MRI studies revealed enhancement of the thickened tenosynovium and synovium when compared with the unenhanced and contrast-enhanced T1-weighted axial image without fat saturation (Figs. 1E, 1G and 2C). The low-signal foci or "dots" in the synovial fluid on T2- and T2*-weighted images were not enhanced on the gadolinium-enhanced images. Bone erosion occurred in seven of our patients, six of whom had osteomyelitis (Figs. 1B, 1C and 2A, 2B, 2C); only one patient was found to have cortical erosion without osteomyelitis. Encasement of the median nerve with blurring of its margin was found in three patients (Figs. 1F, 1G and 3C), and these patients were the same ones who clinically had carpal tunnel syndrome (Table 1). Only one patient had miliary tuberculosis on the follow-up chest radiograph 6 months later and Mycobacterium tuberculosis present in a sputum culture (case 4).
|
Discussion
Tuberculous infection of the wrist is a rare but well-documented disorder [3–5]. Because of the increased number of immunocompromised individuals and drug-resistant bacterial strains, the incidence of extrapulmonary mycobacterial infection has gradually increased over the past decade [1, 2]. In our research, more than 50% (5/8) of the patients had coexisting medical problems that may compromise the immunity. Approximately one third of the patients with musculoskeletal tuberculosis have involvement of other viscera such as the lung [2]. However, pulmonary tuberculosis was found in only one (12.5%) of our eight patients on the follow-up study. In addition, clinical manifestations of tuberculous infection of the wrist are usually chronic and have an insidious onset; hence, the diagnosis can be difficult and the treatment is often delayed, resulting in irreversible osteoarticular damage [12, 13]. The delay from the onset of symptoms to an accurate diagnosis in our study ranged from 3 to 24 months (mean, 10.5 months), which is consistent with those previously reported [3, 14].
Radiographic features of tuberculous infection of the wrist may include localized osteoporosis, bone erosion or cavitation, gross destruction, and eventually joint ankylosis [15, 16]. These findings are not specific and may deceptively mimic inflammatory arthritis, pigmented villonodular synovitis, gout, or soft-tissue tumor.
MRI findings of tuberculous tenosynovitis of the wrist have been reported in a few cases [7–9, 17]. MR images have revealed prominent synovial thickening around the flexor tendons with fluid collection in the tendon sheath [8, 9, 17]. The thickened tenosynovium and synovium usually present as low signal intensity on T1-weighted images. It has been reported that the characteristic hypointense synovium on T2-weighted images that is suggestive of granuloma is seen in approximately 40% of cases [18]. Moreover, heterogeneous intermediate to low signal intensity of the thickened tenosynovium and synovium was seen on T2- and T2*-weighted images in our patients: intermediate signal intensity (n = 3), intermediate to low signal intensity (n = 2), and low signal intensity (n = 3). Synovial fluid has been reported to manifest as intermediate to low signal intensity on T1-weighted images and predominantly high signal intensity on T2- and T2*-weighted images [8, 9, 17]. The small low-signal foci or dots scattered in the synovial fluid on T2- and T2*-weighted images may refer to regions of tissue debris, caseous material, or rice bodies, as described in the literature [8, 17, 19–21]. Rice bodies were also identified in the pathologic specimen in two (25%) of our eight cases. In addition to tuberculous infection, rheumatoid arthritis and seronegative inflammatory arthritis may have a similar appearance and be associated with rice-body formation [8, 22]. The unenhanced MRI findings in our study are consistent with those previously reported.
After IV administration of gadopentetate dimeglumine, the thickened tenosynovium and synovium enhanced in our cases. The enhancement of the thickened tenosynovium and synovium may relate to increased vascularity when tuberculous granulation tissue involves the synovium at the granulomatous stage. Because of the enhanced thickened tenosynovium, contrast-enhanced MRI has been described to be particularly useful in differentiating the thickened tenosynovium from the fluid and the surrounding structures and in determining the extent of the lesion [9]. In addition, the nonenhancement of the small low-signal foci in the synovial fluid on T2- and T2*-weighted images in our study has not, to our knowledge, been previously reported. These features are important MRI findings in diagnosing the tuberculous infection.
With regard to the distribution of tuberculous infection of the wrist, the dominant hand has been reported to have a higher prevalence of involvement [17, 23], and our series upholds this finding. The flexor tendon sheath and radioulnar bursae have been reported to be the most common sites of tenosynovitis, with the dorsal wrist compartment affected less often [17, 23]. However, all patients in our study had extensive synovial lesions around both the flexor and extensor tendons. Moreover, tuberculous infection of the wrist has been described to usually originate from the tenosynovium and then extend to the joint or bone [24, 25]. Our results support this thesis because extensive tenosynovitis without evidence of adjacent bone erosion or joint involvement was seen in one (12.5%) of our eight patients.
Because the radioulnar bursae extend through the carpal tunnel, carpal tunnel syndrome has been reported as a result of tuberculous tenosynovitis [10, 23]. In our series, encasement of the median nerve shown on MR images was seen in three patients (37.5%), and these patients were the ones who clinically had carpal tunnel syndrome (Table 1). MRI has the advantage of great image resolution in evaluating the extent of the lesion and where the median nerve is encased.
The differential diagnosis of the tuberculous joints and tenosynovitis of the wrist should include rheumatoid arthritis, juvenile rheumatoid arthritis, seronegative inflammatory arthritis, pigmented villonodular synovitis, hemophilia, gout, tumors originating from the synovium such as synovial chondromatosis, amyloidosis, and nontuberculous mycobacterial tenosynovitis such as Mycobacterium marinum infection [7, 8, 10, 11, 18, 22, 26]. The aforementioned lesions may exhibit similar MRI findings and need to be differentiated from tuberculous infection. Hypointense synovium on T2-weighted images along with central erosions, bone chips, and rim-enhancing abscesses may differentiate tuberculous arthritis from other types of monoarticular inflammatory arthritis (including rheumatoid arthritis) [18]. In addition, we found that the nonenhancement of the small low-signal foci or dots in the synovial fluid on T2- and T2*-weighted images was a unique finding that may be helpful in the differential diagnosis. Thus, the characteristic MRI appearance, such as T2 low-signal synovial thickening around the flexor and extensor tendons and synovial fluid collection that contains low-signal and nonenhanced foci in the tendon sheath, may aid us in reaching the proper diagnosis. However, surgical débridement of the thickened tenosynovium and synovium and antituberculous medication may be necessary for management in the some patients.
Our study has two limitations. First, the small number of patients selected may make it difficult to include all MRI features of tuberculous infection of the wrist in this study. Second, we have limited some MRI parameters, such as the lack of fat-saturated unenhanced T1-weighted images (coronal view) available to evaluate the synovial enhancement from the fat-saturated contrast-enhanced T1-weighted images (coronal view). There is no doubt that fat saturation increases the conspicuity of image enhancement. Furthermore, because of the phenomenon of pseudoenhancement, fat-saturated contrast-enhanced images should only be compared with fat-saturated unenhanced images of the same sequence. Thus, we can compare only the degree of enhancement from the axial section of the unenhanced and contrast-enhanced T1-weighted images without fat saturation because we obtained only the T1-weighted image with fat saturation (coronal view) in the contrast-enhanced series but non-fat-saturated T1-weighted image (coronal view) in the unenhanced series. If we compared the degree of enhancement from the coronal view, the phenomenon of pseudoenhancement could occur. However, comparison of the degrees of enhancement was made only from the axial section images and would not have the conspicuity of enhancement.
In conclusion, the diagnosis of tuberculous infection of the wrist should be considered when T2 low-signal synovial thickening around the flexor and extensor tendons is present and the synovial fluid collection contains low-signal and nonenhanced foci in the tendon sheath. In addition, bone erosion, osteomyelitis, and median nerve encasement are also frequently present.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  C. C. Roberts, P. T. Liu, and F. S. Chew Imaging Evaluation of Tendon Sheath Disease: Self-Assessment Module Am. J. Roentgenol., March 1, 2007; 188(3_Supplement): S10 - S12. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J TEH and G Whiteley MRI of soft tissue masses of the hand and wrist Br. J. Radiol., January 1, 2007; 80(949): 47 - 63. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
