Midline Destructive Lesions of the Sinonasal Tract: Simplified Terminology Based onHistopathologic Criteria

History

The histopathologic classification of midline destructive lesions has undergone substantial revision over the last two decades. Historically, destructive midfacial lesions, other than those caused by trauma, toxic agents, or infectious or neoplastic processes, were categorized under multiple pathologic labels that did not reflect specific clinicopathologic entities and provided little guidance in disease management.

A confusing variety of terms has been used since 1897, when McBride first described a case of rapid destruction of the face and nose (7). This was followed in 1922, by Stewart's report of 10 cases of a chronic midfacial destructive process, which became known as Stewart's syndrome or Stewart's granuloma. In 1949, Williams popularized the unfortunate term lethal midline granuloma to designate inflammatory midline destructive lesions with no known etiologic factors. Soon after, it became clear that many of the disease processes included under this heading were neither lethal nor granulomatous. This awareness led to the propagation of other nonspecific descriptive terms, such as idiopathic midline granuloma, nonhealing midline granuloma, malignant midline granuloma, and idiopathic midline destructive disease, which were used indiscriminately and which referred to a relatively heterogeneous group of disease processes (Table 2).

Advances in general medical knowledge, and in biochemical and pathologic analysis in particular, led to progressive separation of more specific disease entities from these general nonspecific headings. The first pathologic entity to be clearly distinguished was Wegener's granulomatosis, after Friedmann's histopathologic review of a variety of cases of this disease in 1955 (8, 9). The pathophysiological mechanisms involved in this disease and specific diagnostic criteria were also described (10, 11).

Later, another pathologist, Eichel (12, 13), differentiated another disease process from the idiopathic midline destructive disease group: polymorphic reticulosis, or lymphomatoid granulomatosis. In 1966, he defined this entity as a transitional pathologic process between atypical lymphoid proliferation and frank malignant lymphoma (12–14). The term pseudolymphoma has also been used to refer to this entity.

In the 1970s and 1980s, midline destructive diseases other than Wegener's granulomatosis were subclassified into three different groups that reflected variations in clinical behavior and histology (8). The first group, idiopathic midline destructive disease, was characterized as a locally destructive lesion limited to the upper respiratory tract and with no systemic involvement. The pathologic hallmark was the presence of nonspecific inflammation and necrosis with absence of granulomas and malignant cells (8). Polymorphic reticulosis, or lymphomatoid granulomatosis, the second group, was defined pathologically by the presence of an atypical and polymorphic lymphoid cell population associated with lymphomalike angiocentric and angioinfiltrative growth pattern, causing extensive necrotic changes (10). The last group, extranodal non-Hodgkin's lymphoma, included the cases of frank lymphoma. This was the classification still proposed by Batsakis in 1979 (3, 15).

During the 1980s, further advances in immunohistochemical phenotyping and in molecular genetics allowed better characterization of the cell surface and cell origin. In 1982, Ishi et al (16), using immunofluorescent studies and a variety of antisera directed toward human T- and B-cell surface antigens, definitely linked polymorphic reticulosis with non-Hodgkin's lymphoma. In 1987, Lippman demonstrated activated T-cell phenotype with the typical pattern of peripheral T-cell lymphoma in a case of midline destructive granuloma (17).

In the 1990s, in situ hybridization techniques, gene rearrangement studies, and the discovery of cellular expression of specific cellular membrane molecules allowed for further confirmation of the nature of these lesions (4, 18, 19).

Diagnosis

The diagnosis of destructive diseases of the sinonasal region depends on clinical and pathologic findings, as imaging of these lesions is nonspecific (11, 20–22). Patients usually present with symptoms of nasal obstruction and nasal discharge, which are easily attributed to rhinosinusitis. Epistaxis and facial swelling may be in the clinical spectrum. With disease progression, facial pain and destructive sinonasal lesions may ensue. On physical examination, the most typical finding is the presence of nasal septum perforation, which may or may not be associated with soft-tissue masses and which may progress to autorhinectomy (3, 5, 20).

After assessing the extent of disease with sectional imaging, the first step should be to exclude more common etiologies, such as trauma (accidental or self-induced), cocaine abuse, and infection (Table 3). These may be excluded through clinical history and culture of sinonasal secretions. General systemic symptoms may be present in both sinonasal lymphoma and Wegener's. They include fatigue, night sweats, and weight loss. Although lung and kidney involvement are common in Wegener's granulomatosis, sinonasal disease may be the presenting feature and may antedate involvement of other organs (10, 11). The specific marker of Wegener's granulomatosis, antineutrophil antibodies, may be absent in the serum and confound the diagnosis (11). In this circumstance (negative ANCA and isolated involvement of the sinonasal tract), the diagnosis of Wegener's is based on histologic features, including the presence of noncaseating multinucleated giant cell granulomas and necrotizing vasculitis. Fibrinoid necrosis may or may not be present.

Sinonasal lymphoma is one of the rarest forms of extranodal lymphoma in Western populations, representing less than 0.5% of cases (3). This contrasts with the prevalence in some Asian countries, in which sinonasal lymphoma is the second most common type of extranodal lymphoma. In this geographic group, over 90% of cases have T-cell markers, and Epstein-Barr virus has been consistently demonstrated in the cell genome (3).

The diagnosis of sinonasal non-Hodgkin's T-cell lymphoma is based on the presence of a monoclonal lymphocytic proliferation and specific immunohistochemical markers (CD45 and CD45RO). Angio- and bony invasion is a typical feature of T-cell lymphomas in this region, thus accounting for the term angiocentric lymphoma (2, 3).

Recently, Barker et al (23), reported a single case of idiopathic destructive sinonasal disease that did not conform to the diagnostic criteria of either Wegener's granulomatosis or T-cell lymphoma. It is not clear from this case report what the specific characteristics were of the centrally necrotic granuloma found in one of the nasal biopsy specimens. The authors attributed it to a reaction to previous surgery or topical medication. However, as discussed above, granulomas may be part of the definite diagnostic criteria of early or atypical Wegener's (with negative ANCA and without involvement of other organs). Also, the control of the disease with cyclophosphamide and prednisolone, the recommended therapy for Wegener's granulomatosis, may argue in favor of this disease. Whether cases such as this one represent diagnostic insufficiency or a separate pathologic entity remains elusive and stands in the face of a growing body of literature to suggest otherwise (3, 4, 18, 19).

In the absence of a definite diagnosis, biopsies should be performed and special stainings used as needed. Both T-cell lymphoma and Wegener's granulomatosis may exhibit foci of necrosis and dense mononuclear infiltrates. Specific immunohistochemical markers are useful in establishing a definite diagnosis, particularly when secondary inflammatory changes are extensive. One important concern in tissue sampling is to perform deep biopsies in order to avoid the necrotic and inflammatory components of the lesion. These two factors most likely contributed to the nonspecificity of pathologic findings in the past. Biopsies performed under imaging guidance may be required to improve the diagnostic yield.

Imaging

Imaging of early midline destructive lesions (Figs 1A and 2A and B) is likely to reveal little more than nonspecific findings of mucosal thickening, suggestive of chronic sinonasal inflammation. Bony erosion and destruction are the hallmarks of aggressive lesions but are still nonspecific, eliciting a long differential diagnosis (Table 2). Such destruction is typically seen to first involve the nasal septum and occasionally to spread to the paranasal sinuses, more commonly to the medial wall of the maxillary sinus and ethmoidal trabecula. Advanced disease may lead to destructive lesions of the hard palate, sinonasal-oral fistulas, or complete nasal destruction (autorhinectomy), which is more specific for angioinvasive nasal lesions, such as lymphoma and Wegener's granulomatosis, but can also be seen in angioinvasive fungal infections, such as aspergillosis and mucormycosis. After transgressing bony landmarks, these pathologic processes may extend to adjacent structures. The premaxillary soft tissues, retroantral fat, pterygopalatine fossa, infratemporal fossa, and orbit are the most commonly involved. Intracranial involvement may also result, usually involving the anterior cranial fossa, owing to spread of disease through the cribriform plate.

It is apparent from the radiologic literature that several attempts have been made to distinguish the various types of midline destructive processes on the basis of imaging findings. In 1989, Drake-Lee and Milford (21), after reviewing the plain films and CT scans of 20 cases of Wegener's granulomatosis and seven cases of lethal midline granuloma, concluded that these diseases have no specific radiologic features and that the differences between the two pathologic process were quantitative, with more extensive destructive changes seen in midline destructive disease than in Wegener's.

In 1990, Teng et al (14) described the CT findings in a series of 11 cases of polymorphic reticulosis and concluded that the imaging features were nonspecific. In 1991, Marsot-Dupuch et al (22) described and tried to quantify imaging findings in 13 cases of lethal midline granuloma evaluated with CT and MR imaging. These authors concluded that there are no specific imaging findings and that the main role of imaging this disease is to evaluate the extent of the disease, monitor its progression over time, and ascertain the effect of treatment. They also concluded that CT with high-resolution bone algorithms is the best method to evaluate bony changes, such as remodeling and erosion, and that MR imaging should be used to determine the extent of soft tissue, orbital, and intracranial involvement. They also concluded that MR imaging is useful for distinguishing fluid retention within the sinuses from mucosal thickening and intranasal masses.

The present series was too small to attempt to compare and quantitate the imaging findings; these are highly dependent on the progression of the disease before treatment. In both Wegener's granulomatosis and sinonasal lymphoma, extensive destructive changes, including autorhinectomy, were seen.