323
Downloaded from www.ajronline.org by 203.64.11.45 on 02/21/15 from IP address 203.64.11.45. Copyright ARRS. For personal use only; all rights reserved
Masses
of the Hand
and Wrist:
Detection and Characterization with MR Imaging
Larry A. Binkovitz1 Thomas H. Berquist Richard A. McLeod
To assess the value of MR imaging in the detection, delineation, and characterization of mass lesions of the hand and wrist, we reviewed the MR imaging findings of 38 patients referred for evaluation of such lesions. Twenty-five patients had a palpable mass. In an additional 13 patients an occult mass lesion was suspected as the cause of distal ulnar neuropathy Twenty-two mass lesions (16 benign and six malignant) were detected by MR. All were correctly predicted to be benign or malignant. In nine (56%) of the 16 benign mass lesions, the specific diagnosis was suggested. In the remaining seven benign mass lesions and in the six malignant tumors, the MR findings were not
specific distal
enough
to permit
a diagnosis.
Of the 14 patients
referred
for evaluation
of a
ulnar
neuropathy, an occult ganglionic cyst compressing the ulnar nerve was revealed with MR imaging in three. MR imaging of the hand and wrist is accurate in the detection of mass lesions and can correctly distinguish benign from malignant tumors in the majority of cases. Specific diagnoses can be made in certain benign lesions. Occult mass lesions can be confirmed or excluded as the cause of distal ulnar neuropathy with MR imaging. AJR
154:323-326,
February
1990
Excellent soft-tissue contrast and multiple imaging planes make MR imaging an attractive technique for the diagnosis and evaluation of tumors of the hand and wrist. Several small series of such tumors studied with MR imaging have been reported [1 -4]. We describe our results in a large group of patients in whom mass lesions of the hand and wrist were evaluated with MR imaging.
Materials
and
Methods
Between September 1985 and December 1987, i 51 MR examinations of the hand and wrist were performed in 131 patients. The clinical data, imaging studies, and pathologic findings in these patients were reviewed. Twenty-five of the 131 pationts had a mass palpable at physical examination; in 13 others an occult mass was suspected as the cause of a distal ulnar neuropathy. This subgroup comprised 20 females and 18 males 1 1-75 years old. All examinations were obtained at 1 .5 T with a General Electric Signa MR imaging system (Milwaukee, WI). The technique of the examinations varied. Thirty-six patients had double spin-echo imaging, 2000/30, 60 or 2000/40, 80 (TRITE), in the axial plane. High-spatialresolution examinations were done by using a small local coil or extremity volume coil, limited Received July 10, 1989; September 1 1 , 1989.
accepted
after revision
field of view (8 or 12 cm2), 256 x 256 matrix, and one or two acquisitions. The slice thickness usually was 5 mm. A sagittal and/or coronal partial-saturation sequence was obtained as
St., SW., Rochester, MN requests to L. A. Binkovitz. 0361-803X/90/1542-0323
of the clinical findings or the results of other or previous MR images). BCuse the authors were familiar with most cases at the time of review, a truly blinded retrospective evaluation was not possible. Instead, review of the scans was undertaken to assess the ability of MR imaging to detect, delineate, and characterize the mass lesions. Detection refers to recognition of a mass lesion as being present. Delineation refers to defining
C American
the
Presented at theannual meeting ofthe American Roentgen Ray Society, New Orleans, May 1989. I All authors: Department of Radiology, Mayo Foundation, 55905.
200
Address
FIrst
reprint
Roentgen
Ray Society
needed. Scans were interpreted diagnostic
anatomic
studies
extent
(plain
films,
of the
with knowledge
bone
mass
scans,
and
was
arthrograms,
based
on involvement
or lack
of involvement
of
BINKOVITZ
324
important ing
neurovascular
a tumor
as
characteristics,
structures.
being as
benign
well
as
Characterization
or
tumor
malignant
on
rnargination
refers the and
to defin-
basis the
of
signal
presence
or
Downloaded from www.ajronline.org by 203.64.11.45 on 02/21/15 from IP address 203.64.11.45. Copyright ARRS. For personal use only; all rights reserved
absence of surrounding edema, hemorrhage, or invasion. Characterization also refers to the determination of a specific diagnosis. Table 1 lists the mass lesions detected with MR imaging in the 38 patients
in the
study
group.
Results
Benign
Masses
All 1 6 benign masses were detected retrospectively with MR imaging. The correct diagnosis was suggested on MR images in nine (56%) of these 1 6 tumors, including six ganglionic cysts. All of the ganglionic cysts were well manginated and had no surrounding edema, hemorrhage, on invasion. The signal characteristics were isointense to slightly hypenintense when compared with muscle on short TR/short TE scans (Ti weighted) and markedly hypenintense on long TR/Iong TE scans (T2-weighted). The signal intensity was homogeneous. Two ganglionic cysts were not detected before surgery: one TABLE 1: Mass MR Imaging
Lesions
of the Hand and Wrist
Detected
No.
Lesion
Benign
6a 1 2
Ganglionic cyst Lipoma Giant cell tumor, tendon origin Hypertrophic
palmaris
longus
1
muscle
Bone lesion Other Malignant Epithelial
3a
3
Surgically
Fig.
arising
1a 1a 1a 1a
confirmed.
1.-MR image (2000/60). Ganglionic from scapholunate ligament dorsally
rows) was mistaken as joint effusion. ured 1.5 x 1.3 x 0.5 cm at surgery.
AJR:154, February 1990
was mistaken for a joint effusion (Fig. 1); the other was obscured by subcutaneous fat when imaged with a local coil (Fig. 2A) but was well demonstrated 5 weeks later when imaged with a volume coil (Fig. 2B). The MR appearances of a lipoma (Fig. 3) and a hypertnophic palmaris longus muscle simulating a mass lesion were thought to be characteristic enough that biopsies were unnecessary. In a single case of giant cell tumor of the tendon, the diagnosis was suggested on the basis of MR findings. Areas of relative T2 shortening (Fig. 4) were believed to be suggestive of the paramagnetic effects of hemosidenin. A fibroma or hemorrhagic cyst was considered in the differential diagnosis of this tumor as well. Pathologic findings confirmed a giant cell tumor of the tendon with hemosiderin present. In no case was a benign tumor predicted to be malignant on the basis of the MR findings.
Malignant
Tumors
All six malignant tumors were detected with MR imaging. Five of the six malignant tumors had features typical for malignancy including inhomogeneous signal intensity, which increased with T2 weighting. Because ofthe superior contrast resolution of MR imaging, these malignant tumors were quite conspicuous. However, the margins of these lesions were less well defined than were those of benign masses. There was increased signal in adjacent soft tissue on T2-weighted scans in two (33%) of six malignant tumors. There was marked distortion of tissue planes and encasement of tendons in one (1 7%) of six (Fig. 5). Only the synovial sarcoma (Fig. 6) did not exhibit at least one of these malignant features. It was characterized as malignant in light of clinical findings and size.
2a
sarcoma
Clear cell sarcoma Leiornyosarcoma Synovial sarcoma Rhabdomyosarcoma a
with
ET AL.
cyst (ar-
Cyst meas-
Distal
Ulnar Neuropathy
In 14 patients MR scans were obtained for the evaluation of distal ulnar neuropathy. In 1 3 patients, no mass was palpable. In four patients MR revealed a mass lesion com-
Fig. 2-A, MR image (2000/60). Ganglionic cyst in subcutaneous superficial vein. Local coil technique results In high signal Intensity of adjacent to volar-placed coil. This obscures ganglionic cyst. B, 5 weeks later, lesion was clinically palpable. MR image (2000/60) (arrOws). Volume coil technique resuits in more uniform intensity and
fat (arrows) subcutaneous shows
tumor
was mistaken for fat (fat burnout) to have enlarged
lesion is easily detected
AJR:154,
February
MR
1990
OF
HAND
AND
WRIST
MASSES
325
Downloaded from www.ajronline.org by 203.64.11.45 on 02/21/15 from IP address 203.64.11.45. Copyright ARRS. For personal use only; all rights reserved
Fig. 3.-MR image shows mass with smooth lobulated margins, internal separations, and high signal intensity on Ti-weighted sequence (700/ 20). Appearance is characteristic of lipoma.
Fig. 4.-MR image (2000/60) shows globular soft-tissue mass at level of distal ulna with Inhomogeneous signal intensity and lower than cxpected signal intensity with T2 weighting (arrows). Findings are suggestive of pigmented villonodular synovitis. At surgery a giant cell tumor of tendon was excised.
Fig. 5.-MR image (2000/80) shows large mass with relatively poorly defined margins engulfing tendon of flexor pollicis longus (arrow). Inhomogeneous signal intensity increased with T2 weighting Note biopsy defect inferiorly with surroundIng hemorrhage. Features are typical of malignant tumor. Surgical resection showed alveolar rhabdomyosarcoma.
Fig. 6.-MR image (2000/60) shows large, rather homogeneous mass with well-defined margins and hypointensity with T2 weighting. Findings are not typical of malignant neoplasm. Surgical specimen revealed synovial sarcoma.
pressing the ulnan nerve. There were three nonpalpable ganglionic cysts at the level of Guyon canal (Fig. 7) and one palpable giant cell tumor of the tendon sheath at the distal nadioulnan joint (Fig. 4). No mass was identified in the remaining 1 0 patients. Surgical exploration in five of these patients confirmed no mass lesion involving the distal ulnan nerve. In three other patients symptoms resolved. In two patients symptoms remain unexplained.
Discussion
The earliest musculoskeletal images obtained with MR were of the hand and wrist [5]. The potential for MR imaging in the detection, delineation, and characterization of mass lesions was recognized early [1 6, 7]. High-resolution images were made achievable with the development of local coils and improved software. By 1 986, the highly detailed anatomy of the wrist could be imaged, and several groups presented the normal anatomy of the wrist using high-spatial-resolution techniques [3, 4, 8, 9]. MR imaging has been shown to be accurate in the detection of soft-tissue lesions [1 0, 1 1]. In our series, all 22 tumors of the hand and wrist were detected on the basis of the MR appearance. Only two lesions were not identified prospectively. Both were small ganglionic cysts. One 4-mm cyst was ,
initially overlooked; the second cyst was mistaken for joint effusion. No malignant tumor was missed prospectively. All lesions resulting in ulnan neuropathy at the wrist were detected; none were missed on the basis of surgical (n = 5) on clinical (n = 5) follow-up. MR imaging has been shown to provide accurate information regarding the extent of the soft-tissue lesions [1 0-1 5]. In no case in our series was the tumor extent at surgery found to be beyond that shown by MR imaging. Accurate assessment of neunovascular involvement, tendon encasement, and tumor extent is useful before surgical excision. This has been particularly helpful in the planning of surgical treatment of soft-tissue malignant neoplasms and vascular malformations. Characterization of benign masses was good with MR imaging. Benign masses were well marginated, did not invade or encase neunovasculan or tendinous structures, and had rather uniform signal intensity. There was little or no surrounding tissue reaction. Specific diagnoses were suggested in cases of ganglionic cysts, lipoma, hypertrophic palmaris Iongus muscle, and giant cell tumor of tendon. Giant cell tumors of tendon represent localized extnaarticular pigmented villonodular synovitis. Kransdorf et al. [1 6] found the MR imaging appearance of pigmented villonodulan synovitis to be characteristic enough to allow the correct preoperative diagnosis in six of eight cases. Similar results were reported by Jelinek et
Downloaded from www.ajronline.org by 203.64.11.45 on 02/21/15 from IP address 203.64.11.45. Copyright ARRS. For personal use only; all rights reserved
Fig. 7.-A, Anatomic drawing shows ulnar artory and nerve within Guyon canal (arrow). B, MR image (2000/80) shows ganglionic cyst (arrow) displacing ulnar nerve and artery (arrowhead). Patient had severe ulnar neuropathy for 18 months with muscle wasting, but no mass was palpable. Surgical exploration confirmed ganglionic cyst arising off pisohamate ligament and compressing motor branch of ulnar nerve. (Reprinted with permission from Binkovitz et al. [4].)
al. [1 7]. Areas of decreased or absent signal on Ti - and T2weighted images caused by hemosiderin deposition are key findings in the preoperative diagnosis. The three bone tumors (fibrous dysplasia, osteoid osteoma, and giant cell tumor of the distal radius) were better characterized with plain films than with MR images. The ability to distinguish benign from malignant lesions with MR imaging has been investigated [1 0, 1 1 1 6, 1 8]. Analysis of signal homogeneity and intensity and tumor mangination have been the standard features used for histologic charactenization. Neunovascular invasion, extension beyond one muscle group, and bone destruction have also been used to identify malignant tumors. No single parameter or set of parameters has been found to accurately distinguish benign from malignant lesions in all cases. Recently, Knansdorf et al. [1 6] analyzed 1 1 2 soft-tissue tumors. On the basis of the parameters discussed above, they found “no reliable criteria to distinguish the MR images of malignant from benign softtissue tumors on tumorlike masses.” In our study, all 22 tumors on tumorlike masses were correctly identified as benign on malignant at the time of the initial interpretation. Our study differed from that of Kransdorf et al. in several ways. Our series was smaller and limited to masses of the hand and wrist. We based our final diagnostic impression on the parameters of signal homogeneity and intensity, margination, and soft-tissue invasion or reaction, as well as knowledge of clinical findings. In addition, our series had a relatively lange number of benign lesions that had distinctive MR features (ganglionic cyst, lipoma, giant cell tumor of tendon) on plainfilm findings (fibrous dysplasia, osteoid osteoma, and giant cell tumor of bone). The favorable distribution of lesions may be related to the fact that our series was limited to the hand and wrist. Of the six malignant tumors in our series, only the ,
synovial sarcoma failed to demonstrate any MR features that would suggest malignancy. Histologic examination of this tumor did not reveal fibrosis, hypocellularity, or hemosidenin deposition, which might have accounted for T2 shortening [1 9]. The cause of its atypical MR imaging appearance is unclean, but has been described in other synovial sarcomas [16]. Despite the limitations of our study, we are encouraged with the usefulness of MR imaging of soft-tissue masses. It is accurate in the detection and delineation of mass lesions. The distinction between benign and malignant lesions often can be predicted with MR imaging.
REFERENCES 1 . Brady TJ, Gebhardt MC, Pykett IL. NMA imaging of forearms in healthy volunteers and patients with giant-cell tumor of bone. Radiology 1982;
144:549-552 2. Weiss KL, Beltran J, Lubbers LM. High-field MA surface-coil imaging of the hand and wrist. Part II. Pathologic correlations and clinical relevance. Radiology i986;160: 147-1 52 3. Koenig H, Lucas D, Meissner A. The wrist: a preliminary report on highresolution MR imaging. Radiology i986;1 60:463-467 4. Binkovitz LA, Ehman AL, Cahill DA, Berquist TH. Magnetic resonance imaging of the wrist: normal cross sectional imaging and selected abnormal cases. RadioGraphics 1988;8(6): 1 171 -1 201 5. Hinshaw WS, Bottomley PA, Holland GN. Radiographic thin-section imaging of the human wrist by nuclear magnetic imaging. Nature i977;270: 722-723 6. Mansfield P, Maudsley AA. Medical imaging by NMA. Br J Radio! i977;50:
188-1 94 7. Hinshaw WS, Andrew ER, Bottomley PA, Holland GN, Moore WS, Worthington BS. An in-vivo study of the forearm and hand by thin section NMA imaging. Br J Radio! 1979;52:36-43 8. Weiss KL, Beltran J, Sharnarn OM, Stilla AF, Lefey M. High-field MA surface-coil imaging of the hand and wrist. Part I. Normal anatomy. Radiology i986;1 60: 143-1 46 9. Middleton WD, Kneeland JB, Kellrnan GM, et al. MA imaging of the carnal
tunnel: normal anatomy and preliminary syndrome. AJR 1987;148:307-316
findings
in the carnal tunnel
1 0. Petasnick JP, Turner DA, Charters AJ, Citelis 5, Zacharias CE. Soft tissue masses of the locomotor system: comparison of MA imaging with CT. Radiology 1986;160: 125-133 1 1 . Sundaram M, McGuire MH, Herbold DR. Magnetic resonance imaging of soft tissue masses: an evaluation of fifty-three histologically proven tumors. Magn Reson Imaging i988;6:237-248 1 2. Zimmer WE, Berquist TH, McLeod RA, et al. Bone tumors: magnetic resonance imaging versus computed tomography. Radiology i985;155: 709-718 13. Hudson TM, Hamlin DJ, Enneking WF, Pettersson H. Magnetic resonance imaging of bone and soft tissue tumors: early experience in 31 patients compared with computed tomography. Skeletal Radio! 1985;1 3:134-146 14. Demas BE, Heelan AT, Lane J, Marcove A, Hajdu 5, Brennan MF. Softtissue sarcomas of the extremities: comparison of MA and CT in determining the extent of disease. AJR i988;150:615-620 1 5. Aisen AM, Martel W, Braunstein EM, McMillin KI, Phillips WA, King TF. MRI and CT evaluation of primary bone and soft-tissue tumors. AJR 1986;146:749-756 1 6. Kransdorf MJ, Jelinek JS, Moser AP, et al. Soft-tissue masses: diagnosis using MR imaging. AJR 1989;153:541-547 17. Jelinek JS, Kransdorf MJ, Utz JA, et al. Imaging of pigmented villonodular synovitis with emphasis on MA imaging. AJR 1989;152:337-342 18. Totty WG, Murphy WA, Lee JKT. Soft-tissue tumors: MA imaging. Radio!ogy 1986;160:135-141 1 9. Sundaram M, McGuire MH, Schajowicz F. Soft-tissue masses: histologic basis for decreased signal (short T2) on T2-weighted MR images. AJR 1987;148: 1247-1 250