The finding of pronounced nuclear pleomorphism associated with prominent myxoid matrix in an otherwise leiomyogenic tumour has not been previously reported at this site and must be distinguished from other neoplasms and non-neoplastic processes. Here, we report such a case, focusing on the importance of its unusual morphological features, and discuss the differential diagnosis with other similar lesions.

Unlike previous reports of sinonasal leiomyogenic tumours, our case had pronounced pleomorphic cells coupled with a prominent myxoid background.

These morphological features may pose some difficulties in defining the malignant potential and in the differential diagnosis. Pathologists would probably recognise the reactive nature of the atypical mesenchymal cells in inflammatory or antrochoanal polyps. These bizarre cells are haphazardly dispersed in an oedematous stroma, typically concentrated in the submucosa or near vascular structures, and stain for fibrohistiocytic markers.⁶

Therefore, the main differential diagnosis included a sarcomatoid carcinoma, a myxoma, and a myxoid leiomyosarcoma. The lack of surface epithelial alterations and of a high mitotic activity, and the negative immunostaining for cytokeratins, excluded a sarcomatoid/spindle cell carcinoma. The large amount of Alcian blue positive myxoid material indicated a loss of the fascicular pattern, giving the impression of a myxoid lesion. However, myxomas usually involve the maxillary sinus and appear as a uniform proliferation of scanty spindle/stellate cells floating in a myxoid background and do not stain for myogenic markers.⁷

In approaching this case, our main difficulty was in defining whether the lesion was benign or malignant. As pointed out by Enzinger et al,⁸ pronounced nuclear pleomorphism and myxoid change are the degenerative aspects of a long standing lesion, and are often seen in so called ancient schwannomas. Furthermore, as stated for uterine smooth muscle tumours, the careful assessment of mitotic activity, infiltrating margins, and coagulative tumour necrosis still remains the main criteria to determine the malignant behaviour of leiomyogenic neoplasms.⁹ In contrast to myxoid leiomyosarcoma, which shows at least one of these histological criteria,¹⁰ our case had none of the above findings, except for the presence of prominent nuclear atypia. In addition, the low labelling index and the indolent clinical course of the disease helped to rule out a diagnosis of malignancy. Finally, we would like to give some suggestions that might help with the diagnosis and characterisation of myofibroblastic lesions, which encompass a large spectrum of pathologies, ranging from reactive proliferations or benign tumours to overtly true sarcomas, and which can be reported at almost any anatomical site, as recently reviewed by Mentzel.¹¹ Myofibroblastic lesions mainly consist of spindle shaped cells with a paler and less fibrillary cytoplasm than that of smooth muscle cells. Moreover, myofibroblasts usually are intersected by hyalinised collagenous, rather than myxoid, stroma and do not stain uniformly for actin or desmin, as is often seen in leiomyogenic tumours. We would like particularly to focus attention on myxoid nodular fasciitis, a lesion also included in the update of myxoid tumours by Graadt van Roggen et al.¹² As in the present case, it is usually well circumscribed, displaying great cellularity in a prominent myxoid background and stains (at least weakly) for leiomyogenic markers. However, unlike the lesion reported here, myxoid nodular fasciitis occurs in the head and neck region of children and shows a rapid growth consisting of fibromyofibroblasts arranged in short fascicles without evidence of cytological atypia.¹²