| Sign In to gain access to subscriptions and/or personal tools. |
DOI: 10.1016/j.jhsb.2005.06.009
Magnetic Resonance Imaging in the Diagnosis of Glomus Tumours of the HandFrom the Division of Plastic Surgery at King Saud University and King Fahad National Guard Hospital Riyadh, Saudi Arabia Correspondence: Prof M.M. Al Qattan, Department of Surgery, King Saud University, PO. Box 18097, Riyadh 11415, Saudi Arabia. Tel.: +966 505 274 885; fax: +966 146 79493. E-mail: moqattan{at}hotmail.com
This paper studies the sensitivity, specificity and predictive values of MRI in the diagnosis of glomus tumours of the hand and investigates the final diagnosis and outcome in cases with false positive or negative imaging tests. A total of 42 cases with the clinical diagnosis of a glomus tumour were included in the study. All patients underwent MRI and the results of MRI were correlated with the final histological diagnosis of the excised lesion. MRI had a sensitivity of 90%, a specificity of 50%, a positive predictive value of 97% and a negative predictive value of 20%. The four cases in which the MRI was negative all proved histologically to be glomus tumours. All four tumours were small (2–3 mm in diameter) and the lack of delineation of the lesions by MRI was attributed to their small size. Despite negative MRIs, surgical exploration identified the glomus tumours. Based on the results of the current study and the cost of MRI, the senior author has stopped ordering pre-operative MRIs in patients clinically diagnosed with a glomus tumour.
Key Words: magnetic resonance imaging • glomus tumour • hand
The glomus tumour is a benign growth which arises from specialized cells normally found within the glomus apparatus of the skin. The most common site for the tumour is the distal phalanx, particularly the subungual location, but it can occur at any site of the upper limb including bone and nerve (Johnson et al., 1993; Kline et al., 1990; Smith et al., 1992; Takei and Nalebuff, 1995). Several authors have advocated magnetic resonance imaging (MRI) as part of the pre-operative work up for patients with a clinical diagnosis of a glomus tumour (Hou et al., 1992; Jablon et al., 1990; Takata et al., 2001). The current paper studies the sensitivity and specifi-city and predictive values of MRI in the diagnosis of glomus tumours of the hand and investigates the final diagnosis and outcome in cases with false positive or negative imaging tests.
A total of 42 cases with the clinical diagnosis of a glomus tumour were included in this retrospective study. All patients underwent MRI as part of their pre-operative work up. Both T1- and T2-weighted images were performed for all patients. If the lesion was not delineated well using standard images, better delineation was achieved using the fat suppression technique or with magnetic resonance angiography. Evaluation was done by a radiologist according to standard evaluation techniques (Theumann et al., 2002). The following data were documented: age, sex, site, clinical symptoms and signs and the results of MRI. The sensitivity, specificity and predictive values of MRI were calculated. The final diagnosis and outcome were documented in cases with false positive or negative imaging tests.
The mean age of the 42 patients studied was 40 (range 16–70) years. There were 26 females and 16 males. The duration of symptoms prior to presentation to our Hand Clinic ranged from 3 months to 7 years. The lesion was located in the subungual area in 28 patients and in the pulp in 8 patients and, hence, the distal phalanx was involved in the majority of cases (n = 36). The remaining 6 patients had lesions in the middle phalanx (n = 2), proximal phalanx (n = 2) and palm (n = 2). Pre-operatively, all patients had the triad of pain, cold sensitivity and point tenderness (positive Love test) and were clinically diagnosed as having a glomus tumour. Among the group of 28 patients with subungual lesions, nail ridging was documented in 6 patients and bony indentation was evident on plain X-ray in 5 patients. All 42 patients underwent pre-operative MRI. The pre-operative clinical diagnosis of a glomus tumour was confirmed histologically in 40 patients and, of these, MRI was positive in 36 and negative in 4. The final diagnosis in the remaining two patients was a neuroma (n = 1) and a skin appendage cystic tumour (n = 1). The MRI was negative in the former and positive in the latter patient. A positive MRI showed the localization of a hypointense (dark) lesion on T1 weighted-images which became brighter (higher signal intensity) on T2 weighted images (Fig 1). Five patients required fat suppression or magnetic resonance angiography to delineate the lesion more effectively (Fig 2).
The results of the MRI were correlated with the final histological diagnosis of the excised lesion (Table 1). When we analysed the four cases in which the MRI was negative and the final histology was a glomus tumour, we found that all four tumours were small (2–3 mm in diameter) and located on the palmar aspect of the distal phalanx. The failure of MRI to delineate these lesions was attributed to the small size of the tumours. The patient with a small neuroma in the pulp and a negative MRI admitted after surgery that he had a puncture wound in the same area a few years earlier. The failure of MRI to delineate the neuroma was also attributed to its small size.
In our study, MRI had a sensitivity of 90%, specificity of 50%, positive predictive value of 97% and a negative predictive value of 20%. Two out of the four patients with a false negative MRI were covered by insurance. The Insurance Company, initially, refused to cover the exploration procedure because of the negative MRI. Both patients paid personally for the surgical procedure and were, later, reimbursed after the histological diagnosis of a glomus tumour had been established. After a mean follow-up of 5 (range 2–10) years, 2 patients experienced recurrences of their glomus tumours and both recurrences were seen 2–3 years after surgery.
The triad of pain, cold sensitivity and point tenderness is characteristic of glomus tumours, although other hand tumours may mimic this clinical presentation (Al-Qattan, 1996; Al-Qattan and Clarke, 1994). Several clinical tests have been described for the diagnosis of glomus tumours (Bhaskaranad and Navadgi, 2002; Ekin et al., 1997; Giele, 2002) and these are summarized in Table 2. In our series, only the Love test (point tenderness) was done and it was positive in all patients. Several diagnostic tests have also been described for glomus tumours, viz. plain radiography, angiography, thermography, ultrasonography, scintigraphy and MRI (Chen et al., 2003; Geertruyden et al., 1996; Ogino and Ohnishi, 1993) and their advantages and disadvantages are summarized in Table 3. It is important to realize that many Hand Surgeons still believe that a glomus tumour of the hand is a clinical diagnosis and, hence, they do not perform any pre-operative radiological investigations (Tomak et al., 2003; Vasisht et al., 2004). Despite that, many patients with glomus tumours of the hand are initially misdiagnosed and treated as other lesions, such as chronic nail infections and trigger finger (Figs 3 and 4).
Most glomus tumours (Fig 1) are seen as a dark lesion on T1 and as a bright, high signal intensity lesion on T2-weighted MRI (Drape et al., 1995; Jablon et al., 1990; Matloub et al., 1992). However, some glomus tumours have low signal intensity or isointensity on T2-weighted images and these lesions are better delineated with T2 fat suppression images, with the use of contrast MRI or with magnetic resonance angiography as seen in Fig 2 (Theumann et al., 2002). It is also important to realize that the classic MRI findings of a dark lesion on T1 and a bright lesion on T2-weighted images are not specific for glomus tumours. Similar MRI findings are also seen with cysts and other solid hand tumours (Jablon et al., 1990; Peterson et al., 2004). Hou et al. (1992) showed that the MRI appearance of a high signal central dot surrounded by a zone of less signal intensity (nidus appearance) is characteristic of glomus tumours (see Fig 2A). However, not all glomus tumours will show this nidus appearance. In our study, the sensitivity and the positive predictive vale of MRI were high (90% and 97% respectively) indicating that a positive MRI in a patient with a suspected glomus tumour is highly suggestive of a glomus tumour. In contrast, the specificity and negative predictive value of MRI were low indicating that a negative image does not rule out a glomus tumour. When we analysed our 4 cases with false negative imaging results, we found that all cases had small (2–3 mm) tumours and concluded that they were probably missed by the MRI because of their small size. Both the treating surgeon and patient decided to proceed with surgical exploration in all cases with a false negative MRI. This attitude of exploration, despite a negative MRI, is practiced by other hand surgeons (Geertruyden et al., 1996; Takei and Nalebuff, 1995). This brings into question why we need to do MRI for glomus tumours, bearing in mind the cost of this investigation, as we are going to explore the area regardless of the result of MRI. Some authors who advocated pre-operative MRI for glomus tumours stated that MRI is justified for patients with obscure signs and symptoms, patients with recurrent tumours and for the detection of multiple tumours (Matloub et al., 1992; Theumann et al., 2002). Although multiple glomus tumours is a true entity and several cases have been reported (Dailiana et al., 1999; Graham and Wolff, 1992; Looi et al., 1999; Noor and Masbah, 1997), this is actually rare and, hence, it is hard to justify using MRI in every case in order to detect multiple tumours. Based on the results of the current study and the cost of MRI, one of the authors (M.M.A) has stopped ordering a pre-operative MRI in patients clinically diagnosed with a glomus tumour. However, it is likely that many hand surgeons will continue to order a pre-operative MRI for glomus tumours because surgeons are generally happier operating on a known target and opening a definite or near-definite score.
Received for publication April 21, 2005.
  CiteULike  Connotea  Del.icio.us  Digg  Reddit  Technorati What's this?
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
 |
Sign In to gain access to subscriptions and/or personal tools.
TOP
Abstract
INTRODUCTION
