Acral myxoinflammatory fibroblastic sarcoma (AMIFS) was first described in 1998 by Meis-Kindblom and Kindblom,5 Montgomery et al,7 and Michal.6 These authors described a multinodular, poorly circumscribed tumor that occurs in all age groups as a painless mass, most commonly in the distal extremities. It behaves as a low-grade malignant tumor with a high rate of local recurrence and a low rate of metastasis. Histologically, the lesions are composed of a myxoid matrix containing numerous types of inflammatory cells, fibrosis, and scattered, large, bizarre tumor cells. The neoplastic cells vary and include spindle or epithelioid cells, Reed Sternberg-like cells, ganglion-like cells, and multivacuolated cells simulating lipoblasts.
The histologic differential diagnosis for this lesion is broad, as it contains myxoid, inflammatory, and atypical features, and includes conditions such as tenosynovitis, giant cell tumor of tendon sheath, inflammatory myofibroblastic tumor, liposarcoma, epithelioid sarcoma, and myxoid malignant fibrous histiocytoma (MFH).
We present five patients with AMIFS and review the relevant clinical, radiologic and histologic features.
We identified five patients (Table 1), three men and two women, ranging in age from 22 to 66 years (mean, 43.6 years). Each presented to the referring orthopaedist with a history of a mass in his or her upper extremity for several months to years (range, 3 months-6 years). The location of the lesion included the dorsal hand overlying the base of the second metacarpal, the proximal phalanx of the middle finger, the web space between the ring and small fingers, the dorsoulnar border of the hand at the carpometacarpal junction, and the superior deltoid. Three of the patients, including the patient with the shoulder lesion, described the lesion as a firm, painless mass that was noted incidentally. The fourth patient noted her lesion secondary to tenderness with direct pressure. The fifth patient described the lesion as being painless for 4 years, then becoming increasingly painful for 2 years before seeking medical care. None of the patients complained of any constitutional symptoms at the time of presentation such as fever, weight loss, night sweats, or anorexia. Only one of the five patients had any past medical conditions for which he was taking medication (Synthroid for a goiter [Synthroid, Abbott Park, IL]). None of the patients currently used tobacco, and only one had smoked previously. Only one patient had any type of occupational exposure to potential carcinogens (industrial strength cleaning solvents).
Four of the patients were thought to have a benign condition of the upper extremity at the time of initial evaluation by the referring orthopaedist. Two patients were diagnosed with a ganglion cyst, the third was diagnosed with a giant cell tumor of the tendon sheath, and the fourth was thought to have a benign tumor of the hand. No preliminary diagnosis was given to the patient with the shoulder lesion. None of the patients had a preoperative biopsy. Four of the five patients had preoperative magnetic resonance imaging (MRI) to evaluate the mass. The preoperative MRI scans were interpreted as showing an atypical ganglion, a giant cell tumor of the tendon sheath, and a benign tumor of the hand. All five lesions were marginally or intralesionally excised. Histologic examinations were consistent with AMIFS (although the occurrence in the shoulder originally was considered a benign nodular tenosynovitis), and the patients subsequently were referred to an orthopaedic oncologist for evaluation and treatment.
Each patient, except for the patient with AMIFS in the shoulder who did not keep his followup appointment, was staged with a computed tomography (CT) scan of the chest. All CT scans were negative for metastatic disease. Repeat MRIs also were done, each showing nondiagnostic postoperative changes. A wide tumor bed resection was performed, achieving clear margins in all five patients. A groin flap was required for soft tissue coverage in one of the patients. In another patient, a third-ray resection was required to achieve a negative margin. No adjuvant therapy was administered to any of the patients.
Histologic slides from the first excision in Patient 1 showed irregular nodules of tumor embedded in a fibrous and adipose tissue background (Fig 1). The nodules were separated by dense, sclerotic, fibrous bands containing small to medium-sized blood vessels. At the periphery of the larger nodules there was a moderate inflammatory infiltrate. The infiltrate was comprised of acute and chronic inflammatory cells (Fig 2). The center of each nodule was characterized by chronic inflammatory-type cells floating in a myxoid matrix material. The individual cells, plasma cells, and lymphocytic or histiocytic cells formed a loose reticular array in the extracellular myxoid substance. Interspersed were numerous polymorphonuclear cells and many of the ganglion-like cells or virocytes (Fig 3) that have been described in association with this lesion.2
Specimens from Patient 2 also showed an overall nodular configuration of the lesion. In contrast to Patient 1, this material was more densely cellular and contained less of a myxoid component. The lesion from Patient 2 showed a preponderance of histiocytic cells with an acute and chronic inflammatory cell component interspersed throughout. Occasional histiocytes had large irregular nuclei and abundant eosinophilic cytoplasm characteristic of the virocytes associated with AMIFS.
The resection specimen from Patient 3 consisted of fragments of soft tissue. These had a vaguely nodular configuration of fibroblastic and inflammatory elements. The inflammatory infiltrate was sparse in this lesion and confined to scattered aggregates resembling lymphoid follicles. The histiocytic component was especially prominent. Numerous histiocytes had atypical features including nuclear convolutions and large prominent nucleoli. Immunohistochemical staining was performed on this material, and the lesion stained diffusely positive for vimentin antigen and focally positive for S100 antigen. There was negative staining for muscle and epithelial markers.
The outside slides from Patient 4 showed a neoplasm with a predominant storiform pattern. The cellularity ranged from scant in the myxoid areas to quite dense. In many areas these cells had abundant pale eosinophilic cytoplasm. The nuclei were generally vesicular, but without prominent nucleoli. Their shapes were quite irregular. Large hyperchromatic nuclei also were present. Mitosis is uncommon, generally one per 10 high-power fields. The periphery of the lesion showed dense connective tissue and lymphoplasmacytic infiltrate, which also was present in the neoplasm.
Review of the material from the reexcision in Patient 5 revealed a circumscribed lesion in the subcutaneous tissue. The overall cellularity was low and it had a nodular configuration. The tumor was comprised of enlarged-appearing fibroblastic cells. These had large nuclei, prominent nucleoli, and indistinct cytoplasmic borders. The fibroblastic cells were embedded in a myxoid background. Rare binucleate virocyte cells were identified.
Because the clinical appearance of AMIFS can be similar to that of benign lesions of the hand, patients do not always have imaging studies much more involved than plain preoperative radiographs. Four of our patients had preoperative plain radiographs and MRIs available for review. The fifth patient was operated on by an outside physician based on a clinical examination in which the AMIFS was thought to be consistent with a ganglion cyst.
Plain radiographs did not show abnormalities in two cases. A noncalcified soft tissue mass was seen on radiographs in two cases, one of them with subtle bony erosion.
On MRI, three lesions were located primarily in the subcutaneous tissue of the fingers or shoulder. One lesion was located in the fourth interdigital space of the hand, one in the middle finger, extending into its volar and dorsal aspects and infiltrating the adjacent phalanx, and the last abutting the acromioclavicular joint. In the fourth case, the lesion was located deep to the fifth extensor tendon compartment of the hand. Lesional borders were well defined in two cases, partially well defined in one case, and ill defined and infiltrative in one case.
On T1-weighted sequences, three lesions were hypointense to skeletal muscle, one of them with an isointense peripheral thin rim (Fig 4). In one patient, the tumor was slightly hyperintense to skeletal muscle. In all four patients, the tumor was of lower signal than that of subcutaneous fat on T1-weighted images. Three lesions were homogeneous, and the other was heterogeneous on T1-weighted sequences.
On T2-weighted sequences, the tumor was hyperintense to fat in two cases (Fig 5). In one case, the lesion was predominantly hyperintense to the fat, but contained areas of intermediate and low signal intensity. In the remaining patient, the lesion presented a central hyperintense zone surrounded by a peripheral thick rim of intermediate signal intensity. Gradient-echo recalled T2-weighted sequences were available for this last patient and showed the lesion to be homogeneously hyperintense, with loss of the peripheral, intermediate signal intensity rim seen in the corresponding T2-weighted sequence. Intense enhancement after intravenous contrast administration was homogeneous in two patients, one of them with a peripheral rim of decreased enhancement, and mildly heterogeneous in another.
Most soft tissue tumors of the distal extremities are clinically benign, such as giant cell tumors of the tendon sheath, fibromatosis, ganglion cysts, and tendon sheath fibromas. Sarcomas rarely present in the distal extremities; however, they are often high-grade lesions when present (eg, epithelioid, clear cell, and synovial sarcomas). Acral myxoinflammatory fibroblastic sarcoma is a recently described, relatively rare, slow growing, low-grade soft tissue sarcoma reported exclusively in the extremities, but not restricted to acral sites.2 Previous descriptions of this entity have appeared exclusively in the pathology literature.1,2,5-7,10
We performed a search of the medical literature using the Medline database. The search found 10 articles published since 1998 that examined cases of AMIFS (Table 2). There were 128 cases identified. Fifty-four percent of the cases were in male patients, and 46% were in female patients. The ages of the patients ranged from 4 to 87 years (mean, 38-53 years). All tumors were localized exclusively in the extremities. Sixty-eight percent of the lesions were found in the upper extremity; most commonly (80%) in the fingers or hand. Recurrences after surgical excision were common, with 36 patients having at least one recurrence. However, there have been only four reported cases of metastasis (three confirmed by pathologic diagnosis).
Pathologic characteristics for AMIFS were described by three separate groups in 1998.5-7 These groups each described a poorly circumscribed, multinodular, tumor that often invaded adjacent synovial lining, tendon sheaths, subcutaneous fat, dermis, surrounding sweat glands, and skeletal muscle. The tumors did not involve the epidermis or the bone in any of the patients.
These investigators described AMIFS as having a significant inflammatory infiltrate, including lymphocytes, plasma cell, granulocytes, and eosinophils. Germinal centers often were observed in this infiltrate. Hemosiderin deposition was observed focally, giving a histologic appearance of pigmented villonodular synovitis. Varying degrees of fibrosis often were seen accompanying this inflammatory infiltrate. The neoplastic nature of AMIFS often was obscured by this inflammatory response.
Myxoid zones with scant vascularization and cellularity often were seen demarcated from solid tumor areas by an abrupt transition. These zones were present in varying amounts across their series.
Various atypical cells were observed throughout the specimens. The cells ranged from bizarre ganglion-like cells to multivacuolated lipoblast-like cells and were seen in all specimens. The ganglion-like cells were scattered through the various specimens or clustered in tumor nodules. The multivacuolated cells were found most commonly in the myxoid regions, with their vacuoles containing samples of the surrounding myxomatous tissue. Inflammatory cells often were present in the cytoplasm of these atypical cells. The cytoplasm of the atypical tumor cells was positive for vimentin staining in all cases.
Local recurrences did show apparent neoplastic progression when compared with initial resections. An increase in the number of atypical cells and the extent of tumor cell clustering was observed.5
Meis-Kindblom and Kindblom5 presented a series of 44 patients with AMIFS whose clinical behavior supports its pathologic classification as low grade. Their series of patients had a median followup of 5 years and a 67% rate of local recurrence. Fifty percent of these recurrences were in multiple locations, and 30% required amputation for definitive treatment. In patients who were followed up for longer than 10 years, the local recurrence rate increased to 78%, and the multiple local recurrence rates increased to 56%. There was a metastasis rate of 6% (two of 36 patients), with one case metastasizing to the inguinal lymph nodes and the other to the lungs. The metastases were detected 1.5 and 5 years, respectively, after the primary surgery.5 No metastases were reported by Michal7 or Montgomery et al.7 Sakaki et al reported one of five patients with AMIFS had recurrence and metastasis (regional lymph node) 3 months after primary excision.10
The histologic differential diagnosis of AMIFS is broad and reflects the predominant component seen in the histologic specimen: inflammatory, myxoid, or bizarre/atypical. A high inflammatory component may suggest an inflammatory or reactive process including tenosynovitis, pigmented villonodular synovitis, proliferative or nodular fasciitis, or giant cell tumor of the tendon sheath. Inflammatory myofibroblastic tumor or inflammatory fibrosarcoma may be considered; however, the distal location and higher number of bizarre cells differentiates AMIFS. Lesions with high myxoid components may be confused with ganglion cysts, myxomas, neurofibromas, liposarcomas, extraskeletal myxoid chondrosarcoma, MFH, or inflammatory MFH. A high proportion of atypical or bizarre cells can resemble lesions such as proliferative fasciitis, inflammatory myofibroblastic tumor, inflammatory fibrosarcoma, and some MFHs.5 Lambert et al4 performed cyto-genetic analysis on a patient with AMIFS who had clonal chromosomal changes. Their findings supported classifying AMIFS as a separate neoplastic entity, and potentially could be used to help diagnostically distinguish it from other conditions.
Radiologic features from our series of patients did not indicate a common pattern of MRI findings. The variable MRI appearance of AMIFS mirrors the variable histologic appearance of this tumor. It may have a signal intensity similar to that of a cyst on unenhanced sequences (ie, low signal intensity on T1-weighted images and high signal on T2-weighted images). Contrast-enhanced T1-weighted images more accurately reflect the solid nature of the lesion. Acral myxoinflammatory fibroblastic sarcoma also can present with heterogeneous hyperintensity on T2-weighted images, with focal hypointense areas and wide contact with tendons. This appearance may lead to misdiagnosis of giant cell tumor of tendon sheath.
Acral myxoinflammatory fibroblastic sarcoma is a rare sarcoma that predominantly affects the distal extremities. It is well described in the pathology literature, but not the in the surgical literature. Its clinical presentation can be similar to several benign processes. Patients diagnosed with AMIFS are treated with wide resection because recurrence is frequent (greater than 50%). Postoperative tumor surveillance should include physical examination for the presence of local or regional metastatic disease and a chest radiograph to evaluate for distant metastatic disease. Because AMIFS behaves like a low-grade sarcoma, we perform these evaluations quarterly for the first 2 years, then every 6 months up to 5 years. It is important for physicians to include AMIFS in their differential diagnosis for painless, slow-growing masses in the distal extremities to appropriately treat the lesion and reduce the chance of metastasis.
We thank Dr. Marco Rizzo for providing clinical data and mentorship during this project.