Fine Needle Aspiration Cytology for Soft Tissue Tumours of the Hand

INTRODUCTION

The preoperative diagnosis of soft tissue tumours of the hand is usually based only on clinical findings. This is because of the low frequency of malignant tumours and the limited number of common soft tissue tumours in the hand (Savage and Mustafa, 1984). However, the preoperative diagnosis is not always correct. Localized giant cell tumour of tendon sheath, which is one of the most common soft tissue tumours of the hand, was only correctly diagnosed preoperatively in 20% to 42% of cases in some studies (Rao and Vigorita, 1984; Savage and Mustafa, 1984), and we have encountered malignant lesions which were misdiagnosed as ganglia or only recognized as such postoperatively, following a simple excision. The hand is a special anatomic region in which many important structures such as nerves, vessels, tendons, and bones are confined in a small space, which makes reoperation after recurrence and further operations for malignancy difficult (McFarland, 1982). It is thus desirable that the preoperative diagnosis of soft tissue tumours of the hand is precise (Savage and Mustafa, 1984).

The use of fine needle aspiration (FNA) cytology for the preoperative diagnosis of soft tissue tumours is well documented. It has many advantages, such as the ease of performance in the out-patient clinics, negligible risk of tumour spread, and allowing the preoperative planning of appropriate treatment (Åkerman, 1997). However, there are few reports available on the use of FNA cytology for hand tumours (García-Solano et al., 2000). The purpose of this study was to evaluate the usefulness of FNA cytology for the preoperative diagnosis of soft tissue tumours of the hand.

PATIENTS AND METHODS

The study included 93 soft tissue tumours and tumour-like lesions at the wrist and distal to the wrist, including two recurrent cases, which presented to the Nippon Medical School Hospital between 1988 and 2001. There were 41 men and 52 women with an average age of 45 (range, 3–83) years. Fifty-three lesions were in the fingers, 12 in the wrist, and 28 in other parts of the hand. The diameters of the lesions ranged from 0.5 to 8.0 cm, with a mean diameter of 2.3 cm.

FNA was performed using a standard technique (Frable, 1976). A 20-ml syringe with a 21-gauge needle and a holder was used for the aspiration. Papanicolaou staining following alcohol fixation and May–Grunewald–Giemsa staining following air-drying was performed. One of the authors (YK), an orthopaedic surgeon, performed almost all the aspirations and the other three authors (MM, a cytotechnologist; ZN and MY, pathologists) assessed almost all the smears. The lesions in all the patients were classified according to the cytological findings as malignant, benign, or unclassified. We also attempted to make a specific diagnosis by cytology. We compared the cytological with the postoperative histopathological diagnoses.

RESULTS

Of the 65 lesions (excluding 11 ganglion cysts) from which sufficient material for cytological evaluation was obtained, 62 were cytologically diagnosed as benign, two as malignant, and one as unclassified (Table 1). Of those 65 lesions, 47 were assessed histopathologically postoperatively. The cytological differentiation between benign and malignant tumours did not produce any false-positive nor false-negative results (sensitivity and specificity were 100%). Cytologically benign lesions that were not diagnosed histopathologically were followed up from 1 to 99 months (mean 29 months) and none showed malignant behaviour. The unclassified lesion was a giant cell tumour of tendon sheath. This was not cytologically diagnosed as benign because of the presence of cell-rich material and a locally destructive growth on the X-rays. The diagnosis of all ganglion cysts was made at the time of aspiration, and the cytology of these lesions revealed a few histiocytes.

Of the 47 lesions with sufficient material for cytology which were postoperatively diagnosed histopathologically, 35 (including one recurrent lesion) were specifically diagnosed by FNA cytology (Table 2). For giant cell tumours of tendon sheath (Fig 1), epidermoid cysts, and lipomas, the specific FNA cytology diagnosis were correct in most instances (90–100%). For the schwannomas and fibrous tumours (fibroma of tendon sheath, fibroma, fibrous histiocytoma, and fibrolipoma), the specific cytological diagnoses were less accurate (0–25%). This was because the aspirated material was hypocellular (Table 2). In four chondromas of soft tissue origin, one lesion was not specifically diagnosed by FNA cytology because of the presence of immature cells on the smear, which were also present at the centre of the histopathological sections.

Seventeen of the 93 smears were considered insufficient for cytological evaluation, containing only blood or a small amount of fat (Table 3). A large proportion of lesions with insufficient material were considered to be haemangioma, based on their histological or clinical features (Table 3).

No complications were encountered.

DISCUSSION

This study has demonstrated the usefulness of FNA cytology for the preoperative diagnosis of a soft tissue tumour of the hand, and its ability to distinguish between malignant and benign lesions of the hand. The high sensitivity and high specificity of this procedure found in this study concurs with the results of another study on hand lesions in which the sensitivity and specificity were 91% and 100%, respectively (García-Solano et al., 2000). For the entire body, Wakely and Kneisl (2000) reported that the cumulative sensitivity and specificity of FNA cytology in the literature are 98% and 86%, respectively. The rate of smears considered insufficient for cytological evaluation in our study (18%) is not high despite the small size of lesions in the hand. The reported rates for the entire body are between 7% and 36% (Barth et al., 1992; Bennert and Abdul-Karim, 1994; Costa et al., 1996; García-Solano et al., 2000; Layfield et al., 1986; Wakely and Kneisl, 2000), and those for benign lesions are higher than for malignant lesions (Barth et al., 1992; Layfield et al., 1986). Most lesions which had insufficient material for cytology in our study were haemangioma, and these aspirates consisted mainly of blood (Åkerman, 1997).

Results of this study also indicate that FNA cytology can provide the specific diagnosis of soft tissue tumours of the hand with a relatively high diagnostic rate (75%) (García-Solano et al., 2000). However, the usefulness of FNA cytology for the specific diagnosis of soft tissue tumours in the entire body is controversial (Costa et al., 1996). The rate of specific diagnoses for soft tissue sarcoma has been reported at between 21% and 70%, but there are few reports regarding the rate of specific diagnoses for benign soft tissue tumours (Costa et al., 1996; Kilpatrick et al., 1999; Layfield et al., 1986). Hand lesions, however, may have an advantage for specific diagnosis by cytology over those at other sites such as the thigh, which is the site of predilection for various sarcomas. Some common soft tissue tumours of the hand, such as haemangioma, lipoma, schwannoma and glomus tumours have characteristic clinical features and, in general, the cytological diagnosis of these is made not only according to the cytological features but also according to clinical information, as is the case with the histological diagnosis (Liu et al., 1999). The common features of subcutaneous haemangioma are blue colour, pain and a change in size; that of lipoma include soft consistency upon palpation, that of schwannoma is radiating pain on tapping, and those of glomus tumour are pain, tenderness, and cold sensitivity (McFarland, 1982).

Many common soft tissue tumours of the hand also exhibit characteristic cytological features. Giant cell tumour of the tendon sheath, epidermoid cyst, and lipoma have been reported to exhibit unique cytological features and have been specifically diagnosed by FNA cytology (Cardillo et al., 2001; Eimani and Kumar, 1999; Layfield et al., 1997; Wakely and Frable, 1994; Walaas and Kindblom, 1985). Giant cell tumour of tendon sheath, which is the most common soft tissue tumour of the hand aspirated in our study, has been shown to harbour a mixture of mononuclear cells and multinucleated giant cells on smears (Layfield et al., 1997; Wakely and Frable, 1994). Epithelioid sarcoma, which was the only malignant tumour in our study, has the relatively unique cytological findings of epithelioid features and spindle cells on a background of necrosis and inflammation; they sometimes require ancillary studies such as immunohistochemistry for cytological diagnosis (Cardillo et al., 2001; Kilpatrick et al., 1999). We noted that there are difficulties in the cytological-specific diagnosis of extraskeletal chondromas, one of which was misdiagnosed as a schwannoma in our study because of the presence of immature cells on the smear. In rare cases, extraskeletal chondroma shows immature cells histologically (Enzinger and Weiss, 2001).

The results of this study highlight the main limitation of FNA cytology for diagnosing hand tumours, which is the difficulty in diagnosing schwannoma and fibrous tumours, which are common. Smears from these tumours exhibit the hypocellular and nonspecific features of spindle cells in many cases (García-Solano et al., 2000; Resnick et al., 1997; Yu et al., 1999). Yu et al. (1999) reported that of nine schwannomas, only one was diagnosed by FNA cytology and that schwannoma smears tended to show hypocellular or nondiagnostic features. Although the diagnostic accuracy of FNA cytology for these tumours is poor, the overall success rate for specific diagnosis by FNA cytology in our institution was relatively high (72.3%). We thus believe that FNA cytology is useful for the preoperative diagnosis of soft tissue tumours of the hand, but that an additional diagnostic biopsy may be needed for malignant lesions. An accurate preoperative diagnosis by FNA cytology should be useful in management of patients in an out-patient clinic and in appropriate surgical planning.