Idiopathic Midline Destructive Disease: Fact Or Fiction

Home , Lethal midline granuloma, Rhinoscleroma

Oral Oncology (2005) 41, 340–348

http://intl.elsevierhealth.com/journals/oron/

REVIEW Idiopathic midline destructive disease: fact or ﬁction

Juan Pablo Rodrigo a,b, Carlos Sua´rez a,b, Alessandra Rinaldo c, Kenneth O. Devaney d, Antonino Carbone e, Leon Barnes f, Dennis K. Heffner g, Alﬁo Ferlito c,* a Department of Otolaryngology, Hospital Universitario Central de Asturias, Oviedo, Spain b Instituto Universitario de Oncologı´a del Principado de Asturias, Oviedo, Spain c Department of Surgical Sciences, ENT Clinic, Policlinico Universitario, University of Udine, Piazzale S. Maria della Misericordia, I-33100 Udine, Italy d Department of Pathology, Foote Hospital, Jackson, MI, USA e Division of Pathology, National Cancer Institute, IRCCS, Aviano, Italy f Department of Pathology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA g Department of Endocrine and Otorhinolaryngic/Head & Neck Pathology, Armed Forces Institute of Pathology, Washington, DC, USA

Received 4 October 2004; accepted 15 October 2004

KEYWORDS Summary The differential diagnosis of a progressive destructive lesion of the mid- Idiopathic midline face and upper airway region includes both neoplastic and non-neoplastic entities; destructive disease; of these, the majority of cases prove to be either Wegener’s granulomatosis or lym- Midfacial granuloma phoma. Historically, these sorts of necrotizing midfacial lesions were diagnosed clin- syndrome; ically, and as a consequence a variety of overlapping categories of disease sprang up. Wegener’s As pathologic examination of biopsy material became both more widespread and granulomatosis; (particularly in the last several years) more sophisticated, many lesions previously Malignant lymphoma thought to be of mysterious origins have proven to be examples of lymphoma (in particular, sinonasal natural killer cell or T cell [NK/T] lymphomas). At present, the evaluation of a patient with a progressive destructive process involving the midface region should include imaging studies (to delineate the extent of disease) as well as biopsy (with sampling of lesional tissue for application of sophisticated testing—including immunohistochemical studies, ﬂow cytometry, or molecular studies as necessary—to exclude the possibility of a NK/T cell lymphoma). There

* Corresponding author. Tel.: +39 0432 559302; fax +39 0432 559339. E-mail address: [email protected] (A. Ferlito).

remain occasional patients whose necrotizing midfacial lesions continue to be dif- ﬁcult to classify despite the application of extensive testing; such patients are sometimes described as suffering from the nebulous entity of ‘‘idiopathic midline destructive disease’’. While it remains to be seen whether such patients will ulti- mately be assigned to other diagnostic groups (as, for example, occult toxic inju- ries—as in the case of cocaine abusers who are not forthcoming with regard to their drug usage), it seems likely that ‘‘idiopathic midline destructive disease’’ is a diagnostic term of questionable validity which should be used only with extreme reticence in modern practice. c 2004 Elsevier Ltd. All rights reserved.

Introduction History

The term ‘‘midline granuloma syndrome’’ or ‘‘mid- At the meeting of the Laryngological Society of facial granuloma syndrome’’ are clinical terms that London on 9 December 1896, McBride4 showed pho- encompass a broad spectrum of diseases character- tographs of a patient with progressive destruction ized by a progressive (and, often fatal) ulceration of the face and nose. Under the heading of ‘‘malig- and destruction of the upper airway, with involve- nant granuloma’’, Woods5 described two patients ment of the nose, the paranasal sinuses, oral cavity in 1921 with a progressive ulcerated lesion of the and the soft tissues of the face. Over the years, oronasal region; lacking, however, was a clear there has been confusion over the various terms description of the pathological appearances. In used to refer to this clinical condition: malignant 1929, Kraus6 reported three patients with granulo- granuloma, granuloma gangraenescens, progres- mas and extensive destruction of the nose, oral sive lethal granulomatous ulceration, lethal mid- cavity and pharynx and suggested the diagnostic line granuloma, non-healing midline granuloma, term ‘‘granuloma gangraenescens’’. In 1933, Stew- polymorphic reticulosis, midline malignant reticu- art7 extensively reviewed the disease, which be- losis, idiopathic midline granuloma, lymphomatoid came known as Stewart’s syndrome. The best reticulosis, and others have all been employed to known term, ‘‘lethal midline granuloma’’ was describe progressive ulceration and destruction of introduced by Williams8 in 1949, to describe the midfacial region.1 inflammatory destructive lesions with no known The histopathologic classification of these etiologic factors. destructive lesions has undergone substantial revi- The histopathologic features of some of these sion in the last years. Now it is known that, after lesions were studied by Friedmann9 who, after exclusion of granulomatous infections and neopla- eliminating infectious and neoplastic diseases, pos- sias, the differential diagnosis of midfacial destruc- tulated the existence of two entities, the Wege- tive processes includes the following entities: ner’s and the Stewart’s type of midline Wegener’s granulomatosis (WG), malignant lym- granuloma. In the first one, a giant cell inflamma- phoma, Churg-Strauss syndrome, microscopic poly- tory granulomatous infiltration was present, angitis and drug abuse. Currently there is doubt whereas in the second the infiltrate was more pleo- about the existence of idiopathic midline destruc- morphic, with or without histiocytic cell predomi- tive disease (IMDD), and some authors think that nance. Later, Eichel et al.10 recognized a possible after exclusion of WG nearly all remaining cases relationship between these latter lesions to pri- presenting as midfacial granuloma syndrome are mary non-Hodgkin’s lymphoma of the nose, and non-Hodgkin’s lymphomas2,3—in particular, the coined the term ‘‘polymorphic reticulosis’’, refer- subtype of angiocentric NK/T-cell lymphoma. Each ring to a mixed infiltrate of lymphoreticular cells. of these diseases has its own treatment and as a In 1977, Michaels and Gregory11 reexamined histo- consequence, older descriptive terms should be re- logic sections from 30 patients who had previously placed by the pathologic diagnosis that accurately been clinically diagnosed as malignant granuloma; describes the disease entity to avoid inappropriate of these, 10 showed features of malignant lym- therapy. phoma. Therefore, at this point, the histopathol- In this review, the clinico-pathological charac- ogy of midline granuloma syndrome included: teristics, diagnosis and therapy of ‘‘midfacial gran- WG; polymorphic reticulosis; and conventional uloma syndrome’’ are revisited. lymphoma of the nose.12,13 In 1982, Tsokos 342 J.P. Rodrigo et al. et al.14 identified an additional group of patients fied as nasal-type extranodal NK/T-cell with midline granuloma syndrome who had distinc- lymphoma.18 tive clinico-pathologic features and which were aggregated under the term ‘‘idiopathic midline destructive disease’’. This was characterized as a locally destructive lesion limited to the upper Clinical features respiratory tract and with no systemic involvement. The pathologic hallmark was the presence Cases of midline destructive disease (Figs. 1–5) of non-specific inflammation and necrosis with ab- present over a wide age range, peaking in the sixth sence of granulomas and malignant cells. decade, with males predominating. Non-Hodgkin’s During the 1980s, the advances in immunohisto- lymphoma of the midface region is one of the rar- chemical phenotyping and in molecular genetics fi- est forms of extranodal lymphoma in the Western nally permitted a more detailed characterization world, representing less than 0.5% of cases. The of the cell origin in these cases of midfacial incidence of these tumors is much higher among destructive diseases. It was now possible to demon- Asian and Southern and Central American popula- strate the monoclonal (neoplastic) lymphoid origin tions, where most of the cases are of the T-cell of the atypical cells in polymorphic reticulosis.15 or NK-cell phenotype.17,19 The term ‘‘angiocentric’’ lymphoma was used to The initial symptoms are usually those of a non- describe these nasal lymphomas, with the term specific rhinitis or sinusitis (with nasal obstruction ‘‘angiocentric’’ referring to the invasion by lym- and nasal discharge). Epistaxis and facial swelling phoma cells of blood vessels with destruction of may enter into the clinical spectrum as well. With blood vessel wall, leading to the widespread disease progression, ulcerations spread, destroying necrosis.16 the soft tissues, cartilage and bone. Subsequently, All of these lymphomas were originally believed facial pain and facial deformities develop. Patients to have a T-cell origin; however, more recently, it may also demonstrate signs which stem from infil- has become apparent that many of them originate tration of local structures, such as cranial nerve from natural killer (NK) cells.17 These NK cells palsies or diplopia and proptosis due to intra-orbi- share many phenotypic features with T-cells and tal or skull base extension. Similar destruction so are grouped with them—resulting in the term may extend to the oral cavity, the larynx and the NK/T-cell lymphoma. Therefore, in the current hypopharynx. Death after a long duration may be proposed World Health Organization classification the result of cachexia, hemorrhage, meningitis, of lymphoid malignancies, these tumors are classi- or inter-current infection.1

Figure 1 CT showing the absence of the nasal septum. Idiopathic midline destructive disease: fact or ﬁction 343

Figure 2 Patient with a wide destruction of the palate Figure 3 Lateral view of a boy showing facial deformity and nasal septum. due to destruction of nasal bones and septum.

The patient’s general condition typically re- on the nasal septum or midline palate. Large mains surprisingly good throughout much of the amount of necrotic tissues are often present in course of the disease in many cases. Systemic the affected structures. symptoms such as fever, chills, night sweats, and Several authors have attempted to distinguish weight loss are sometimes reported in patients the various types of midline destructive processes with non-Hodgkin’s lymphomas but are not typical. on the basis of imaging findings. They have con- In the case of WG systemic symptoms due to dis- cluded that there are no specific imaging findings seminated vasculitis generally develop early in and that the main role of imaging in this disease the course of the disease, although there are sev- is to evaluate the extent of the disease, monitor its progression over time, and ascertain the effect eral cases reported in which the disease remains 3,20–22 confined to the upper airways for several years, be- of treatment. Computed tomography (CT) fore the full systemic manifestation of generalized with high-resolution bone algorithms is the best WG develops.14 method to evaluate bone changes, such as remod- eling and erosion, whereas magnetic resonance imaging (MRI) should be used to determine the extent of soft tissue, orbital and intra-cranial Diagnosis involvement. After assessing the extent of disease via radio- The diagnosis of destructive diseases of the sinona- graphic imaging studies, the next step should be sal region depends on clinical and pathologic find- to exclude more common etiologies, such as trau- ings, as imaging of these lesions is non-specific.3 ma, chemical exposure (cocaine abuse), infection, On physical examination, the most typical find- and epithelial neoplasms. The entities that should ing is the presence of a nasal septal perforation. be included in the differential diagnosis of the sino- Rarely, disease may present in the nasopharynx or nasal tract are listed in Table 1. Once these lateral wall of nasal cavity. The lesions are gray common causes have been eliminated through or yellow with a friable granular surface and arise clinical history and bacteriologic cultures, the 344 J.P. Rodrigo et al.

the identiﬁcation of the presence of pulmonary and/or renal disease. However, as previously mentioned, sinonasal disease may be the presenting symptom and may antedate involvement of other organs. In such cases, the presence of antineutro- phil cytoplasmic antibodies (ANCA) in the serum of the patients may facilitate the diagnosis. These antibodies are relatively speciﬁc markers of WG and are present in 89% of patients with active disease.23 In the rare cases of isolated sinonasal disease and negative ANCA, the diagnosis is based on histological features.

Histopathology

As mentioned, in most cases the diagnosis of destructive diseases of midfacial region hinges on the light microscopic examination of biopsy tissue from the lesions (often supplemented by ancillary studies, such as immunohistochemistry or flow cytometry). Because the abnormal tissue is largely necrotic, multiple biopsy specimens are often required before a diagnosis is made. Moreover, since the incidence of nasal lymphomas is low, the biopsy specimen is usually immediately placed in Figure 4 Frontal view of the same patient demonstrat- formalin, whereas the best approach would be to ing signs of infiltration of facial structures, with intra- submit the specimen fresh and unfixed to the 17 orbital extension. pathologist for a lymphoma work-up. Specific immunohistochemical markers are often useful in establishing a definite diagnosis, particularly main differential diagnostic considerations rest be- when secondary inflammatory changes are tween Wegener’s disease and malignant lym- extensive. phoma. Some authors still consider the diagnosis When all of the characteristic features of WG are of IMDD when the destructive disease is restricted present in a single biopsy specimen, the diagnosis of to the upper aerodigestive system and did not con- WG is relatively easy; the disease is characterized form to the diagnostic criteria of either WG or lym- by the presence of non-caseating multinucleated phoma.23 The diagnosis of WG is usually aided by giant cell granulomas and necrotizing vasculitis,

Figure 5 Necrotic tissues in the midline palate and premaxilla of the same patient. Idiopathic midline destructive disease: fact or ﬁction 345

Table 1 Conditions which may present clinically with the ‘‘midline granuloma syndrome’’ Cocaine abuse Trauma Infectious diseases Bacterial: brucellosis, syphilis, rhinoscleroma, leprosy, actinomycosis, tuberculosis Fungal: histoplasmosis, candida, mucormycosis, blastomycosis, rhinosporidiosis, coccidiomycosis Parasitic: leishmaniasis, myiasis Inﬂammatory diseases Sarcoidosis Wegener’s granulomatosis Systemic lupus Polyarteritis nodosa Hypersensitivity angiitis Idiopathic midline destructive disease Neoplastic diseases Squamous cell carcinoma Basal cell carcinoma Esthesioneuroblastoma Adenoid cystic carcinoma Sinonasal lymphoma

associated with a mononucleate inflammatory infil- Immunohistochemically, many of the NK/T-cell trate composed mainly of histiocytes and eosinoph- lymphomas (up to 65% of them) have an NK cell ils. Fibrin deposition in small arteries, capillaries phenotype.25,26 These cells are positive for CD2, and venules may also be present.24 Such a classic CD3e, and CD56, which is a NK cell marker, but lack appearance does not, typically, appear in a small other T-cell antigens such as surface CD3, CD4, biopsy sampling and therefore several biopsies CD5, and also lack staining for other NK cell mark- are often required before a confident diagnosis of ers such as CD16 and CD57.17,19,25–27 Occasional WG can be made. cases of the NK/T-cell lymphomas are positive for By contrast, the diagnosis of NK/T-cell lympho- CD7 and CD30.18 Conversely, in cases of T-cell lym- mas can be extremely difficult. In addition to the phomas, the atypical cells are positive for T-cell histological appearance of the lesion, it usually re- markers such as CD2, CD3, CD5, and CD45RO. These quires an analysis of the immunophenotype of the cells lack TdT, CD34, and CD1a, which are markers atypical cells, ideally supplemented by genomic of immature, thymic cells.17,19 studies seeking the presence of rearrangements Genotypically, an important distinction between of T-cell receptor genes.16 Histologically, these le- T-cells and NK cells is that NK cells do not exhibit sions tend to be characterized by a prominent poly- rearrangements of the T-cell receptor genes.28 morphic inflammatory cell infiltrate (eosinophils, Thus, NK cells show the T-cell receptor gene in neutrophils, histiocytes, and benign lymphocytes), germline configuration, while T-cells show a mono- containing variable numbers of recognizable atypi- clonal rearrangement in their T-cell receptor cal malignant cells. They are characterized by genes. angiocentricity or angiotropism in which tumor Although the mechanisms leading to malignant cells infiltrate and destroy blood vessel walls, with change in NK/T-cell lymphomas have not been elu- resultant widespread ischemic necrosis.16,17,19 cidated, the presence of Epstein-Barr Virus (EBV) in Unfortunately, however, the recognizable malig- tumor cells has been demonstrated in nearly all nant cells can be very scarce; this fact, coupled cases.16,17,19 Therefore, EBV may very well play a with the prominence of the reactive infiltrate and role in the development of these tumors. the widespread necrosis, is likely to account for Because of the modern immunohistochemical the fact that many of these lesions have previously studies and gene rearrangement analyses, we know been designated as non-specific inflammatory pro- now that nearly all patients suffering from ‘‘idio- cesses.19 In contrast, B-cell lymphoma cells are pathic midline destructive lesion’’ have nasal NK/ not characterized by angiotropism. T-cell lymphomas. However, there remain recent 346 J.P. Rodrigo et al. reports23,29 of cases that could be included in the nervous system can also be seen.25 The best treat- entity described by Tsokos et al.14 as ‘‘idiopathic ment for sinonasal NK/T-cell lymphomas consist of midline destructive disease’’. As mentioned, the in the combination of an anthracycline-based che- histopathology of this lesion was characterized by motherapy regimen (e.g. CHOP—cyclophospha- non-specific acute and chronic inflammation, with- mide, doxorubicin hydrochloride, vincristine out any evidence of malignancy. Whether these sulfate, and prednisone-) along with locoregional cases represent diagnostic insufficiency or a sepa- radiotherapy.17,35 This approach is found to be rate pathologic entity remains elusive, but stands superior to chemotherapy alone in patients with in face of a growing body of literature that suggest intermediate and high-stage disease (stages IIE to otherwise.2,3 IVE), with overall five-year survival rates of 82% Indeed, some authors have raised the possibility in patients receiving combined treatment, as com- that a significant number of cases of putative IMDD pared to 72% in patients receiving chemotherapy may actually represent unsuspected cocaine-in- alone.35 Patients treated with radiotherapy alone duced disease.30 The magnitude of destruction were reported to have a recurrence rate of 49%.36 associated with cocaine-induced disease may be However, for patients with low-stage tumors truly impressive, even rivaling or exceeding that (Stage IE) radiotherapy alone is felt to be ade- seen in WG in some instances; the natural reti- quate.17 There is some suggestion that patients cence of many patients to confess to the use of with nasal NK/T-cell lymphoma have a worse prog- an illegal substance may well play a role in keeping nosis, compared with those with B- and T-cell nasal the concept of IMDD alive as a diagnostic entity. lymphoma.17,25,35,36 However, immunophenotype is not an independent prognostic factor, the only factors independently related to overall survival Treatment and prognosis being the stage of disease and patient age (young patients have better prognosis).17 If, instead of repeated biopsies and immunohis- In the past, most cases designated as ‘‘lethal mid- tochemical and genetic studies, the diagnosis re- line granuloma’’ were treated with variable doses mains that of IMDD, the treatment of choice is of local radiotherapy in an attempt to halt or slow 14 15,31 radiation therapy. This ‘‘disease’’ remains al- the progression of the disease. Many patients ways localized to the upper aerodigestive tract, treated in this way were free of disease after and the prognosis after radiation therapy is good.14 lengthy follow-up, while some were observed to evolve into non-Hodgkin’s lymphoma, probably reflecting the fact that the initial lesion was al- ready a lymphoma. Harrison32 in his analysis of 36 Conclusion patients with midline destructive granuloma treated personally over many years found that despite The advances in immunocytochemical phenotyping the effective local control some patients died from and molecular genetics have revealed that, with disseminated lymphoma, possibly T-cell. the exception of WG, most cases of midline Today, each disease has its own treatment based destructive disease syndrome are now thought to on histopathologic diagnosis. WG is best treated represent extranodal non-Hodgkin’s lympho- with immunosuppressive cytotoxic therapy, partic- mas,2,11,32,37–43 especially sinonasal NK/T-cell lym- ularly cyclophosphamide. Corticosteroids, which phomas.17,19,44 Therefore, although from a clinical reduce the vasculitis, are given concurrently. point of view it may still be useful to retain Long-term complete remission can be achieved descriptive clinical terms, the use of the exact hist- with therapy, even with advanced disease.33 opathologic diagnosis should be relied upon to Sinonasal NK/T-cell lymphoma is an aggressive avoid inappropriate therapy. As a histologic diagno- disease. Patients with this type of lymphoma follow sis, IMDD should disappear.2,43 It would be best to an aggressive clinical course and are significantly think of the acronym IMDD as representing an ‘‘in- associated with an extremely poor prognosis com- ane, moronic, dangerous diagnosis’’.43 pared with B-cell and T-cell lymphoma.34 Left un- In conclusion, it is evident that molecular diag- treated, the disease is uniformly fatal. Some nostic techniques, including a search for the pres- cases remain isolated in the nose for prolonged ence of viral infection, molecular analysis of periods, resulting in local tissue destruction. Other clonality and immunohistological studies, are of cases show early dissemination with frequent pul- primary importance in the definition of this distinct monary and skin involvement, although dissemina- type of lymphoma. In the diagnostics practice, tion to the liver, gastrointestinal tract, and central early diagnosis of sinonasal NK/T-cell lymphomas Idiopathic midline destructive disease: fact or fiction 347 with the possibility of inhibiting tumor growth be- 21. Teng MMH, Chang CY, Guo WY, Li WY, Chang T. CT fore widespread dissemination occurs might suc- evaluation of polymorphic reticulosis. Neuroradiology ceed in shifting the prognosis of this aggressive 1990;31:498–501. 22. Ooi GC, Chim CS, Liang R, Tsang KWT, Kwong YL. Nasal T- process in a more favorable direction. cell/natural killer cell lymphoma: CT and MR imaging features of a new clinicopathologic entity. AJR Am J Roentgenol 2000;174:1141–5. References 23. Barker THW, Hosni AA. Idiopathic midline destructive disease—does it exist? J. Laryngol. Otol. 1998;112:307–9. 24. Batsakis JG. Midfacial necrotizing disease. Ann Otol Rhinol 1. Friedmann I. McBride and the midfacial granuloma syn- Laryngol 1982;91:541–2. drome. J Laryngol Otol 1982;96:1–23. 25. Cheung MMC, Chan JKC, Lau WH, Foo W, Chan PT, Ng CS, 2. Cleary KR, Batsakis JG. Sinonasal lymphomas. Ann Otol et al. Primary non-Hodgkin’s lymphoma of the nose and Rhinol Laryngol 1994;103:911–4. nasopharynx: clinical features, tumor immunophenotype, 3. Borges A, Fink J, Villablanca P, Eversole R, Lufkin R. Midline and treatment outcome in 113 patients. J Clin Oncol destructive lesions of the sinonasal tract: simplified termi- 1998;16:70–7. nology based on histopathologic criteria. AJNR Am J 26. Kwong YL, Chan ACL, Liang R, Chiang AKS, Chim CS, Chan Neuroradiol 2000;21:331–6. TK, et al. CD56+ NK lymphomas: clinicopathological fea- 4. McBride P. Photographs of a case of rapid destruction of the tures and prognosis. Br J Haematol 1997;97:821–9. nose and face. J Laryngol Otol 1897;12:64–6. 27. Wang J, Han D, Yin H. Analysis of clinical manifestations in 5. Woods R. Malignant granuloma of the nose. Br Med J nasal NK/T-cell lymphomas. Lin Chuang Er Bi Yan Hou Ke Za 1921;2:65–6. Zhi 2004;18:82–3., [in Chinese]. 6. Kraus EJ. U¨ber drei Fa¨lle einer eigenartigen Neubildung der 28. Chiang AKS, Srivastava G, Lau PWF, Ho FCS. Differences in Nasen-Rachen-und Mundho¨hle. Klin Wochenschr 1929;8: T-cell-receptor gene rearrangement and transcription in 932–4. nasal lymphomas of natural killer and T-cell types: impli- 7. Stewart JP. Progressive lethal granulomatous ulceration of cations on cellular origin. Hum Pathol 1996;27:701–7. the nose. J Laryngol Otol 1933;48:657–701. 29. Batra P, Shah N, Mathur S. Midline lethal granuloma—a 8. Williams HL. Lethal granulomatous ulceration involving the clinical enigma. Indian J Dent Res 2003;14:174–83. midline facial tissues. Ann Otol 1949;58:1013–54. 30. Trimarchi M, Gregorini G, Facchetti F, Morassi ML, Manfre- 9. Friedmann I. The pathology of malignant granuloma of the dini C, Maroldi R, et al. Cocaine-induced midline destruc- nose. J Laryngol Otol 1955;69:331–41. tive lesions: clinical, radiographic, histopathologic, and 10. Eichel BS, Harrison EG Jr, Devine KD, Scanlon PW, Brown serologic features and their differentiation from Wegener HA. Primary lymphoma of the nose including relationship to granulomatosis. Medicine (Baltimore) 2001;80:391–404. lethal midline granuloma. Am J Surg 1966;112:597–605. 31. Fauci AS, Johnson RE, Wolff SM. Radiation therapy of 11. Michaels L, Gregory MM. Pathology of ‘non-healing (midline) midline granuloma. Ann Intern Med 1976;84:140–7. granuloma’. J Clin Pathol 1977;30:317–27. 32. Harrison DF. Midline destructive granuloma: fact or fiction. 12. Eichel BS, Mabery TE. The enigma of the lethal midline Laryngoscope 1987;97:1049–53. granuloma. Laryngoscope 1968;78:1367–86. 33. Fauci AS, Haynes BF, Katz P, Wolff SM. Wegener’s granu- 13. Fechner RE, Lamppin DW. Midline malignant reticulosis. A lomatosis: prospective clinical and therapeutic experience clinicopathologic entity. Arch Otolaryngol 1972;95: with 85 patients for 21 years. Ann Intern Med 1983;98: 467–76. 76–85. 14. Tsokos M, Fauci AS, Costa J. Idiopathic midline destructive 34. Kim GE, Koom WS, Yang WI, Lee SW, Keum KC, Lee CG, disease (IMDD): a subgroup of patients with the ‘‘midline et al. Clinical relevance of three subtypes of primary granuloma syndrome’’. Am J Clin Pathol 1982;77:162–8. sinonasal lymphoma characterized by immunophenotypic 15. Ishii Y, Yamanaka N, Ogawa K, Yoshida Y, Takami T, analysis. Head Neck 2004;26:584–93. Matsuura A, et al. Nasal T-cell lymphoma as a type of so- 35. Miller TP, Dahlberg S, Cassady JR, Adelstein DJ, Spier CM, called ‘‘lethal midline granuloma’’. Cancer 1982;50: Grogan TM, et al. Chemotherapy alone compared with 2336–44. chemotherapy plus radiotherapy for localized intermediate- 16. Harris NL, Jaffe ES, Stein H, Banks PM, Chan JK, Cleary ML, and high-grade non-Hodgkin’s lymphoma. New Engl J Med et al. A revised European-American classification of lym- 1998;339:21–6. phoid neoplasms: a proposal from the International Lym- 36. Burton GV, Atwater S, Borowitz MJ, Huang AT. Extranodal phoma Study Group. Blood 1994;84:1361–92. head and neck lymphoma. Prognosis and patterns of 17. Vidal RW, Devaney K, Ferlito A, Rinaldo A, Carbone A. recurrence. Arch Otolaryngol Head Neck Surg 1990;116: Sinonasal malignant lymphomas: a distinct clinicopatholog- 69–73. ical category. Ann Otol Rhinol Laryngol 1999;108:411–9. 37. Friedmann I. Ulcerative/necrotizing of the nose and para- 18. Chan JKC, Jaffe ES, Ralfkiaer E. Extranodal NK/T-cell nasal sinuses. Curr Diagn Pathol 1995;2:236–55. lymphoma, nasal type. In: Jaffe ES, Harris NL, Stein H, 38. Barnes L. Midfacial destructive diseases. In: Barnes L, Vardiman JW, editors. World Health Organization classifi- editor. Surgical pathology of the head and neck. 2nd cation of tumours. Pathology and genetics of tumours of ed. New York: Marcel Dekker; 2001. p. 759–86. haematopoietic and lymphoid tissues. Lyon: IARC Press; 39. Harrison DF. Non-healing granulomata of the upper respi- 2001. p. 204–7. ratory tract. Br Med J 1974;4:205–9. 19. Sheahan P, Donnelly M, O’Reilly S, Murphy M. T/NK cell non- 40. Yamanaka N, Kataura A, Sambe S, Minase T, Ishii Y. Hodgkin’s lymphoma of the sinonasal tract. J Laryngol Otol Midfacial T cell lymphoma: characterization by monoclonal 2001;115:1032–5. antibodies. Ann Otol Rhinol Laryngol 1985;94:207–11. 20. Drake-Lee AB, Milford CA. A review of the role of radiology 41. Chan JK, Ng CS, Lau WH, Lo ST. Most nasal/nasopharyngeal in non-healing granulomas of the nose and nasal sinuses. lymphomas are peripheral T-cell neoplasms. Am J Surg Rhinology 1989;27:231–6. Pathol 1987;11:418–29. 348 J.P. Rodrigo et al.

42. Gaulard P, Henni T, Marolleau JP, Haioun C, Henni Z, Voisin 43. Heffner DK. Idiopathic midline destructive disease. Ann MC, et al. Lethal midline granuloma (polymorphic reticu- Otol Rhinol Laryngol 1995;104:258., [Letter]. losis) and lymphomatoid granulomatosis. Evidence for a 44. Mills SE, Gaffey MJ, Frierson HF. Tumors of the upper monoclonal T-cell lymphoproliferative disorder. Cancer aerodigestive tract and ear. Washington, DC: Armed 1988;62:705–10. Forces Institute of Pathology; 2000. p. 219–25, 239–41.