Cellular angiofibroma is a benign mesenchymal lesion that was first described in 1997,1 and which chiefly involves the vulval region. The original report described four cases of this distinctive lesion, all occurring in middle aged women, and the authors considered this to represent a benign neoplasm with little or no potential for local recurrence if excised with a rim of uninvolved normal tissue. Since then, an identical lesion has been described in a woman involving the subcutaneous tissue of the chest wall.2 Similar lesions have also been reported in the inguinoscrotal region of men.1,3 Here, we describe a vulval cellular angiofibroma that, although initially excised with a rim of normal tissue, exhibited tumour recurrence in a relatively short time period.
A 49 year old woman presented with a mass in the posterior aspect of the left labia majora. A well circumscribed lesion measuring 4 cm in diameter was excised with a rim of normal tissue. Six months later she developed a recurrent swelling at the site of the previous excision. This lay in the angle between the posterior wall of the vagina and the anterior aspect of the external anal sphincter. A magnetic resonance imaging scan confirmed the presence of a recurrent lesion and the mass was excised. The mass was well circumscribed and was dissected free without complication. Ten months after excision of the recurrent lesion the patient is well with no further evidence of local recurrence.
The original surgical specimen consisted of a well circumscribed 4 cm firm white lesion, which was completely surrounded by a rim of normal tissue. The recurrent lesion consisted of a well circumscribed 6.5 cm diameter firm white lesion.
Histology of the original biopsy showed a well circumscribed but unencapsulated lesion, which was completely surrounded by a rim of compressed uninvolved tissue. The lesion was composed of short interlacing bundles and fasicles of spindle shaped cells with bland vesicular nuclei and abundant eosinophilic cytoplasm (fig 1A1A).). There was no necrosis and few or no mitotic figures. A notable feature was the prominent vascularity of the lesion and many of the blood vessels were characterised by thick hyalinised walls (fig 1B1B).). Small numbers of mature adipocytes were present within the lesion, especially around the periphery (fig 1C1C)) and there were scattered stromal mast cells and occasional small collections of mature lymphocytes. Histology of the recurrent lesion showed similar features. Again the lesion was well circumscribed. There were foci of mildly increased cellularity and decreased vascularity compared with the original lesion and scattered mitotic figures were identified, the mitotic count being < 1/10 high power fields. An additional feature was the presence of many small lymphoid aggregates throughout the lesion. These were mostly composed of mature lymphocytes with occasional germinal centres (fig 22).
Immunohistochemistry of both the original and recurrent lesion showed positivity of the spindle shaped cells for vimentin (Dako, Copenhagen, Denmark) but no staining for desmin (Dako), α smooth muscle actin (Sigma, Poole, Dorset, UK), S-100 protein (Diagnostic Products Limited, Abingdon, UK), CD34 (Serotec, Oxford, UK), or AE1/AE3 (Dako). There was weak staining for epithelial membrane antigen (EMA; Dako). There was diffuse strong nuclear positivity for the oestrogen receptor (ER; clone ID5; Dako; fig 33)) and the progesterone receptor (PR; clone 1Ab; Dako).
Electron microscopy showed spindle shaped tumour cells embedded in a collagen rich matrix containing vascular channels. Most of the tumour cells had a fibroblastic appearance with prominent cisterni of rough endoplasmic reticulum. Some of the tumour cells contained small numbers of cytoplasmic actin filaments with focal deposits of external lamina and occasional subplasmalemmal densities, in keeping with myofibroblastic differentiation.
The lesion we describe, which exhibits identical histological features to the cases of cellular angiofibroma reported by Nucci et al,1 was characterised by local recurrence in a relatively short period of six months. This was in spite of the fact that the lesion was adequately excised originally with a rim of uninvolved tissue. This is the first report of recurrence of a cellular angiofibroma and illustrates the potential for local recurrence even with adequate excision. The recurrent lesion contained foci of increased cellularity but there were no histological features to suggest malignancy. We do not feel that the fact that the neoplasm recurred is indicative of malignancy.
It is not our purpose here to reiterate in detail the morphological differential diagnosis of cellular angiofibroma, which has been adequately dealt with previously.1,4 However, in the vulval region this differential diagnosis may include neoplasms that are relatively specific to this site, such as aggressive angiomyxoma and angiomyofibroblastoma. Other neoplasms that are not specific to the vulva, such as solitary fibrous tumour, spindle cell lipoma, superficial angiomyxoma, smooth muscle tumours, nerve sheath tumours, and perineurioma, also enter into the differential diagnosis. The distinction of cellular angiofibroma from these lesions is predominantly by morphology, although ancillary immunohistochemical studies may contribute. Cellular angiofibroma is characterised by vimentin positivity together with negative staining for desmin, α smooth muscle actin, and S-100 protein. Negative staining for desmin and α smooth muscle actin helps to exclude aggressive angiomyxoma, angiomyofibroblastoma, and smooth muscle neoplasms, whereas the lack of staining for S-100 protein excludes a nerve sheath tumour. However, it should be stressed that in many of these lesions there is immunohistochemical overlap and close morphological examination remains the mainstay of diagnosis.
The present tumour was negative for CD34, excluding a solitary fibrous tumour, which has been described in this region.5 Although cellular angiofibroma was initially thought to be CD34 negative, two additional cases described in addendum to the original publication were found to be positive, with cellular angiofibroma thus joining the ever increasing list of mesenchymal lesions that may express the CD34 antigen.
The case we describe exhibited weak positive staining for EMA, which has not been described previously. EMA positivity raises the possibility of a perineurioma, a benign lesion of perineural cells.6 However, perineurioma characteristically exhibits a prominent storiform growth pattern and moreover ultrastructural examination in our patient revealed no evidence of perineural differentiation. Rather, electron microscopy showed fibroblastic and myofibroblastic features. The predominant fibroblastic differentiation is in keeping with a vimentin positive but desmin and α smooth muscle actin negative immunophenotype; positivity with the last two antibodies being more characteristic of myofibroblastic than fibroblastic differentiation.
The present lesion exhibited diffuse strong positivity for ER and PR. A recent study investigating the hormone status of a variety of vulvovaginal mesenchymal lesions found most of these to be positive for ER and/or PR,7 and positivity may simply be a reflection of the presence of these receptors normally in subepithelial mesenchymal cells of the lower female genital tract. Nucci et al did not perform immunostaining for ER or PR on their patients but, interestingly, the cellular angiofibroma involving the chest wall exhibited no staining for ER or PR.2