 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1997 The Journal of Bone and Joint Surgery, Inc. The Utility of High-Resolution Magnetic Resonance Imaging in the Evaluation of the Triangular Fibrocartilage Complex of the Wrist* , NEW YORK, N.Y.Investigation performed at the Departments of Radiology, Orthopaedic Surgery, and Biostatistics, The Hospital for Special Surgery, New York City
We performed a prospective study in order to assess the utility of high-resolution magnetic resonance imaging in the detection and specific localization of tears of the triangular fibrocartilage complex. Seventy-seven patients who had pain in the wrist were studied with use of a dedicated surface coil and three-dimensional gradient-recalled techniques with a field of view of eight centimeters and a slice thickness of one millimeter. The patients had pain on the ulnar side of the wrist, ligamentous instability, occult ganglia, or a combination of these. Magnetic resonance images were assessed for radial or ulnar avulsion, or both; central defects; degenerative intrasubstance changes; and complex tears of the triangular fibrocartilage complex. Partial tears were differentiated from complete tears. The findings on the magnetic resonance images were then compared with the arthroscopic findings. Fifty-seven of the fifty-nine tears that were suspected on magnetic resonance images were confirmed with arthroscopy; the two suspected tears that were not confirmed had been interpreted as small partial tears on the magnetic resonance images. With use of arthroscopy as the standard, magnetic resonance imaging had a sensitivity of 100 per cent (fifty-seven of fifty-seven), a specificity of 90 per cent (eighteen of twenty), and an accuracy of 97 per cent (seventy-five of seventy-seven) for the detection of a tear (  = 0.93, p < 0.00001). Fifty-three of the fifty-seven tears were localized correctly with use of magnetic resonance imaging. With regard to the location of the tear, magnetic resonance imaging had a sensitivity of 100 per cent (fifty-three of fifty-three), a specificity of 75 per cent (eighteen of twenty-four), and an accuracy of 92 per cent (seventy-one of seventy-seven) ( = 0.9, p < 0.0001).
We concluded that high-resolution magnetic resonance imaging permits accurate depiction and localization of tears of the triangular fibrocartilage complex. When the appropriate pulse sequence is used, magnetic resonance imaging is an accurate and effective method for the non-invasive evaluation of pain in the wrist.
Disorders of the triangular fibrocartilage complex are a common source of pain on the ulnar side of the wrist6,12,19. However, a tear or degeneration of the triangular fibrocartilage complex may not be readily apparent on physical examination because a variety of lesions, including occult fractures, chondral tears, extensor carpi ulnaris tendinitis, osteoarthrosis of the pisiform-triquetral joint, and instability of the distal radio-ulnar joint may produce pain on the ulnar side, making a correct diagnosis difficult. Historically, arthrography of the wrist, particularly the three-compartment injection method, has been shown to be efficacious in the detection of tears or perforations of the triangular fibrocartilage complex, which allow contrast medium to communicate between the radiocarpal and distal radio-ulnar joints14,27. However, the arthrogram often does not permit exact localization of the tear or supply information about the adjacent soft-tissue structures, including articular cartilage and ligaments. Because of its superior soft-tissue contrast and its tomographic nature, magnetic resonance imaging can detect ligamentous injuries of the wrist5,9,18,22,23,25,26,29. The purposes of the present prospective study were to assess the utility of high-resolution magnetic resonance imaging in the detection and specific localization of tears of the triangular fibrocartilage complex in patients who had pain in the wrist and to compare these findings with those of direct arthroscopic inspection.
Seventy-seven patients who had pain in the wrist were evaluated prospectively, from January 1993 to April 1996, with a standardized magnetic resonance imaging technique before being examined arthroscopically. The patients ranged in age from thirteen to seventy years (mean and standard deviation, 35.7 ± 13.5 years). Fifty-two patients were male and twenty-five were female. Fifty-eight patients had had a previous traumatic injury of the wrist that had been caused by hyperextension (thirty-nine patients), twisting (fourteen patients), or lifting (five patients); the remaining nineteen had no history of trauma. The study was reviewed and approved by the Institutional Review Board. All of the magnetic resonance images were made with use of a 1.5-tesla superconducting magnet (Signa; General Electric, Milwaukee, Wisconsin) and either a five-inch (12.7-centimeter) curved receive-only surface coil or a quadrature design phased-array wrist coil (both from Medical Advances, Milwaukee, Wisconsin), the latter of which became available in November 1994. The patient was placed in the supine position with the affected hand at the side and the forearm in full pronation. Coronal images were made with use of a volumetric gradient-recalled acquisition technique, with a repetition time of forty to fifty-six milliseconds, an echo time of twenty milliseconds, a flip angle of 20 degrees, and a field of view of eight centimeters. Images were made from the palmar extrinsic ligaments to the dorsal extrinsic ligaments. The coronal volumetric acquisition was centered over the proximal aspect of the lunate, just distal to the radiocarpal joint space. The slice thickness was one millimeter with no interslice gap. The images were made with either a 256-by-192 matrix (fifty patients) or a 256-by-256 matrix (twenty-seven patients) at two excitations. (The matrix was increased at the time of acquisition of the quadrature coil.) The field of view was centered over the ulnar side of the proximal aspect of the lunate. Axial images were made with a long repetition time (3000 to 4000 milliseconds) and a long effective-echo time (eighty-eight to 102 milliseconds). Fast-spin-echo images were made with an eight echo train and a 256-by-192 matrix at two excitations. The field of view was eight to nine centimeters, and the slice thickness was four millimeters with no interslice gap. No contrast medium was used in any patient. The magnetic resonance images were studied for abnormalities of the triangular fibrocartilage complex, including radial or ulnar avulsion or detachment; complete or partial perforation of the mid-substance of the articular disc; degenerative intrasubstance changes, as manifested by increased signal intensity and loss of the normal contour of the articular disc; and joint fluid coursing through a defect in the articular disc. Complete tears were defined as full-thickness foci of discontinuity that extended from the radiocarpal side of the articular disc to the distal radio-ulnar side. Partial tears were defined as foci of abnormal morphology and increased signal intensity that did not extend all the way from one side of the articular disc to the other. If a partial tear was detected, a note was made as to whether it was on the radiocarpal side or the distal radio-ulnar side. A tear needed to be identified on only one magnetic resonance image in order to be considered present, but the criteria regarding abnormal morphology and increased signal intensity had to be met. To simplify the statistical analysis, the location of the tears was classified as ulnar, radial, central, or complex (two or more components). An articular disc was considered to be normal if images revealed uniformly low signal intensity at both the radial and the ulnar attachment (Fig. 1), no central defect, and no foci of thinning or increased signal intensity.
In addition, the images were studied for concomitant tears of the scapholunate and lunotriquetral ligaments, defects of the articular cartilage, tears of the dorsal and volar distal radio-ulnar ligaments, and occult ganglia of the wrist. The images were evaluated by one radiologist, who did not have access to the clinical findings or the plain radiographs.
Arthroscopic Procedure Patients who had evidence of a partial-thickness tear of the triangular fibrocartilage complex on the magnetic resonance images but who did not have arthroscopic examination of both the radiocarpal and the distal radio-ulnar joint were not included in the study. In patients who had a normal-appearing triangular fibrocartilage complex on the preoperative magnetic resonance images but who were examined arthroscopically because of evidence of another injury or condition (such as a chondral injury, an injury of the scapholunate or lunotriquetral ligament, a soft-tissue ganglion, or a combination of these) on the images or at the clinical examination, the triangular fibrocartilage complex was carefully inspected and probed for subtle abnormalities such as attenuation and partial tears. The findings on the preoperative magnetic resonance images were compared with the arthroscopic findings with regard to the presence and location of a tear of the triangular fibrocartilage complex or a ligamentous injury, or both. The time from the magnetic resonance imaging to the arthroscopic examination ranged from one day to forty-one weeks (mean, 6.5 ± 7.6 weeks).
Statistical Analysis
A tear of the triangular fibrocartilage complex was identified prospectively on the magnetic resonance images of fifty-nine of the seventy-seven patients (Fig. 2). Fifty-seven of the fifty-nine suspected tears were confirmed at the arthroscopic examination; the two tears that were not confirmed arthroscopically had been interpreted as small partial tears on the magnetic resonance images. Twenty articular discs were found to be intact on direct visual examination and probing at the time of arthroscopy. With use of arthroscopy as the standard, magnetic resonance imaging had a sensitivity of 100 per cent (fifty-seven of fifty-seven) (95 per cent confidence interval, 92 to 100 per cent), a specificity of 90 per cent (eighteen of twenty) (95 per cent confidence interval, 83 to 97 per cent), and an accuracy of 97 per cent (seventy-five of seventy-seven) (95 per cent confidence interval, 93 to 100 per cent) for the detection of a tear of the triangular fibrocartilage complex ( = 0.93; p < 0.00001).
Twenty-one partial tears were identified with magnetic resonance imaging, and fourteen of these were confirmed as partial tears on arthroscopic examination (Fig. 3). Thirty-eight complete tears were identified with magnetic resonance imaging, and thirty-seven of these were confirmed as complete tears on arthroscopic examination. (Two articular discs that were thought to have a partial tear on magnetic resonance images were found to be intact at the time of arthroscopy, as mentioned previously.) Thus, the extent of fifty-one of the fifty-seven confirmed tears had been correctly interpreted on magnetic resonance images. The extent of the remaining six confirmed tears had been misinterpreted: five tears that had been interpreted as partial on the magnetic resonance images were found to be complete on arthroscopic examination (Fig. 4), and one tear that had been interpreted as complete on the magnetic resonance images was found to be partial at the time of arthroscopic examination. When misinterpretation of the extent of the tear was considered a false-positive result, magnetic resonance imaging had a sensitivity of 100 per cent (fifty-one of fifty-one) (95 per cent confidence interval, 91 to 100 per cent), a specificity of 69 per cent (eighteen of twenty-six) (95 per cent confidence interval, 59 to 79 per cent), and an accuracy of 90 per cent (sixty-nine of seventy-seven) (95 per cent confidence interval, 83 to 97 per cent) ( = 0.7; p = 0.0001).
Twenty-nine tears were ulnar, eight were radial (Figs. 5 and 6), nine were central, and eleven were complex (Figs. 7-A and 7-B) as determined on arthroscopic examination compared with thirty, eight, nine, and twelve, respectively, as determined on the preoperative magnetic resonance images. Fifty-three of the fifty-seven confirmed tears had been localized correctly on the magnetic resonance images. Statistical analysis for location was performed with use of a strict criterion as the findings on the magnetic resonance images had to agree absolutely with the findings of the arthroscopic examination in order to yield a so-called true-positive result. According to this criterion, magnetic resonance imaging had a sensitivity of 100 per cent (fifty-three of fifty-three) (95 per cent confidence interval, 91 to 100 per cent), a specificity of 75 per cent (eighteen of twenty-four) (95 per cent confidence interval, 65 to 85 per cent), and an accuracy of 92 per cent (seventy-one of seventy-seven) (95 per cent confidence interval, 86 to 98 per cent) ( = 0.9; p < 0.0001).
To determine if age had a predictive role in the determination of the site of the tear, the locations were recategorized from ulnar, radial, central, and complex to the broader groups of peripheral (ulnar and radial), central, and complex. The mean age was 35.5 ± 14.4 years (range, thirteen to sixty-six years) for the thirty-seven patients who had a peripheral tear, 40.7 ± 14.4 years (range, twenty-one to seventy years) for the nine patients who had a central tear, and 39.3 ± 12.3 years (range, sixteen to fifty-seven years) for the eleven patients who had a complex tear. In our relatively young study group (mean age, 35.7 years), the two-tailed t test and one-way analysis of variance showed that there was no significant age difference among the patients in these categories. A two-tailed t test and one-way analysis of variance were used to compare the age of the patients and the time from the magnetic resonance imaging to the arthroscopic examination with regard to gender. The mean age was 36.2 ± 15.1 years (range, thirteen to seventy years) for the male patients and 34.8 ± 9.4 years (range, twenty-one to fifty-six years) for the female patients (p = 0.7). The mean time from the magnetic resonance imaging to the arthroscopic examination was 7.0 ± 8.2 weeks (range, one day to forty-one weeks) for the male patients and 5.4 ± 6.1 weeks (range, one to twenty-four weeks) for the female patients (p = 0.5). In addition to the triangular fibrocartilage complex, the scapholunate and lunotriquetral ligaments were consistently identified and surveyed for abnormalities on the magnetic resonance images. Arthroscopic inspection, however, was not conducted for every patient; fifty-three of the seventy-seven patients had arthroscopic examination of the scapholunate ligament, and thirty-six had arthroscopic examination of the lunotriquetral ligament. Fifteen patients had a tear of the scapholunate ligament on both magnetic resonance imaging and arthroscopic examination, thirty-six patients had an intact scapholunate ligament on both magnetic resonance imaging and arthroscopic examination, and two patients were thought to have an intact scapholunate ligament on magnetic resonance imaging but were found to have a tear on arthroscopic examination. Two patients had a tear of the lunotriquetral ligament on both magnetic resonance imaging and arthroscopic examination, and thirty patients had an intact lunotriquetral ligament on both magnetic resonance imaging and arthroscopic examination. In the remaining four patients, the status of the ligament was misinterpreted on the magnetic resonance images; there were one false-positive and three false-negative findings. Additional incidental ligamentous injuries that were noted on both magnetic resonance imaging and arthroscopic examination included four tears of the palmar distal radio-ulnar ligament and two tears of the radioscaphocapitate ligament.
In the past, the diagnosis of a tear of the triangular fibrocartilage complex was made by either direct operative visualization or arthrographic evaluation. Arthrographic detection of such tears requires the communication of contrast medium between the radiocarpal and distal radio-ulnar joints. Previous reports have described the use of a multiple-compartment injection technique for arthrographic evaluation of the wrist. Levinsohn et al. noted that only sixteen (67 per cent) of twenty-four complete tears of the triangular fibrocartilage complex that were identified at the time of an operation had been seen with preoperative triple-compartment arthrography. Metz et al. found no association between the localization of symptoms and the specific site of an arthrographic abnormality. More recently, Chung et al. reported that the results of triple-injection arthrography agreed with those of arthroscopic examination for only sixty-three (42 per cent) of 150 patients. We believe that proper magnetic resonance imaging technique is crucial for the detection of subtle morphological changes in the triangular fibrocartilage complex. Previous reports have suggested that magnetic resonance imaging is inferior to arthrography10, but those studies were limited by low-contrast, thick-section (three-millimeter) image acquisition resulting from the limitations of previous versions of software. Pederzini et al. stated that magnetic resonance imaging was unable to localize the precise site of injury of an articular disc. Cerofolini et al. reached a similar conclusion with use of a relatively large field of view and conventional spin-echo low-contrast pulse sequencing. Sugimoto et al. reported abnormally high signal intensity within the substance of the triangular fibrocartilage complex on the magnetic resonance images of thirty-five (50 per cent) of seventy asymptomatic volunteers. That study, however, was limited by a relatively large field of view (fifteen centimeters), which yielded poor spatial resolution. Previous studies on the use of magnetic resonance imaging for the detection of tears of the triangular fibrocartilage complex have relied largely on a so-called arthrogram effect, which occurs when fluid (characterized by high signal intensity) courses through the triangular fibrocartilage complex (characterized by low signal intensity)9,29. Because magnetic resonance images are digitized, spatial resolution is limited by the size of the picture element, or pixel. A smaller field of view and a higher matrix yield small pixels and impart superior spatial resolution. In addition, thin (one-millimeter) slices promote less signal average. Thus, magnetic resonance images made with thin slices, a small field of view, and a high matrix may obviate the diagnostic reliance on the communication of high-signal-intensity fluid between the radiocarpal and distal radio-ulnar joints. The use of such thin-section, high-resolution techniques may allow acceptable diagnostic accuracy to be achieved without the use of intra-articular injections of contrast media such as gadolinium, thereby preserving the non-invasive advantage of magnetic resonance imaging. It should be recognized that gradient-echo techniques are limited by increased sensitivity to inhomogeneities in the magnetic field, including paramagnetic agents (such as hemoglobin degradation products) and ferromagnetic instrumentation (such as suture anchors). These techniques, therefore, may be of limited use in the postoperative setting, when thin-section fast-spin-echo techniques may prove superior to gradient-echo sequences. It is established that fast-spin-echo techniques are more effective for the evaluation of soft tissues surrounding metallic implants11,21. In addition, short echo-time sequences, such as gradient-echo techniques, are subject to an artefact known as the so-called magic angle phenomenon8. This artefact is an increase in signal intensity within the substance of type-I collagen constructs due to a decay in T2 relaxation time, which occurs when the long axis of fibrocartilage, tendon, or ligament is oriented at 55 degrees relative to the long axis of the magnetic field. However, with regard to the triangular fibrocartilage complex, the magic angle phenomenon can only account for abnormal signal intensity within the articular disc, and thus it should not be confused with morphological abnormalities such as retraction, fraying, and detachment. An awareness of this artefact is important for proper positioning of the wrist relative to the long axis of the magnetic field. Inappropriate positioning may result in artefactually increased signal intensity, possibly leading to misdiagnosis of intrasubstance degeneration. Arthroscopy is an effective diagnostic tool for the detection and localization of lesions of the triangular fibrocartilage complex; however, at our institution it is considerably more expensive than magnetic resonance imaging. Therefore, while arthroscopy may confirm an intact articular disc, it may not be cost-effective unless therapeutic operative intervention is indicated. In the present series, the level of agreement between magnetic resonance imaging and arthroscopy with regard to the localization of tears of the articular disc was high, as indicated by an accuracy of 92 per cent (seventy-one of seventy-seven). The ability of magnetic resonance imaging to distinguish between partial and complete tears, however, was not quite as good, as indicated by an accuracy of 90 per cent (sixty-nine of seventy-seven). Full-thickness disruptions that were misinterpreted as partial disruptions on the magnetic resonance images may have reflected fragments of fibrocartilage that were disrupted but closely apposed, creating the illusion of partial continuity (Fig. 4). Despite this limitation, the superior soft-tissue contrast of magnetic resonance imaging made it possible to visualize the morphology of the triangular fibrocartilage complex and enabled distinction between degenerative intrasubstance tears, which demonstrate increased signal intensity within the articular disc (Fig. 6), and traumatic, more discrete peripheral detachments (Fig. 2). Most degenerative tears occur in older patients as a result of age-related attritional changes within the relatively avascular portion of the disc, as previously shown in anatomical studies2,16. Age alone, however, does not necessarily impart predictive value with regard to the localization of tears, as older patients may also sustain a traumatic peripheral avulsion. Several patients in the present series sustained a complex tear with traumatic avulsion superimposed on a central attritional defect (Figs. 7-A and 7-B). We could not detect a significant difference in age when the tears were divided according to location (peripheral, central, or complex). Thus, in our group of relatively young patients (mean age, 35.7 years), the site of the tear of the triangular fibrocartilage complex could not be predicted on the basis of age alone. The high prevalence of peripheral tears (thirty-seven of fifty-seven tears) is most likely a function of the relatively young age of our patients as well as the frequency of an antecedent traumatic injury involving hyperextension of the wrist with forced rotation of the forearm. An interesting finding in the present study was that traumatic injuries were more commonly localized to the ulnar aspect of the disc than to the radial aspect. This is in contrast with the findings of previous reports in which a high rate of radial tears was noted6. Our results are similar to those reported by Totterman et al.26, who used a similar volumetric gradient-recalled sequence. In that series, in which thirty-one patients were evaluated with magnetic resonance imaging and subsequent arthroscopy of the wrist, eleven of twelve full-thickness tears of the triangular fibrocartilage complex were described accurately. The findings of the present series confirm the conclusion of Totterman et al. that high-resolution magnetic resonance imaging is a reliable technique for the detection of tears of the triangular fibrocartilage complex. A valid criticism of our study is that the rate of false-negative findings could not be determined accurately because hand surgeons at our institution are hesitant to perform an arthroscopic examination for patients who have normal findings on preoperative high-resolution magnetic resonance images. We believe, however, that the high prevalence of tears of the triangular fibrocartilage complex in asymptomatic patients, as noted by Kirschenbaum et al. and Cantor et al. with arthrography and by Sugimoto et al. with magnetic resonance imaging, does not invalidate the utility of such imaging in the evaluation of symptomatic patients. This position was emphasized in a recent editorial comment by Miller and Totterman, and it is not inconsistent with the findings of previous magnetic resonance imaging studies of the spine in asymptomatic patients3,28. We concluded that high-resolution magnetic resonance imaging is an effective tool for the assessment of patients who have pain on the ulnar side of the wrist in whom an abnormality of the triangular fibrocartilage complex is suspected. In addition, we conclude that our imaging technique is accurate for the detection and specific localization of tears of the triangular fibrocartilage complex and that it can be used to direct appropriate operative intervention or non-operative treatment as needed.
NOTE: The authors thank Lewis Lane, M.D., and Michelle Gerwin, M.D., for providing additional patients.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |