Advances in the management and monitoring of extranodal NK/T-cell lymphoma, nasal type


  • Raymond Liang

    1. S.H. Ho Professor of Haematology and Oncology, Department of Medicine and Centre of Cancer Research, Li Ka Shing Faculty of Medicine, University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong, China
    Search for more papers by this author

Raymond Liang, MD, K417 Department of Medicine, University of Hong Kong, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong, China. E-mail:


Extranodal natural killer (NK)/T-cell lymphoma, nasal type, has a unique geographic distribution. Its pathology is characterized by marked angio-invasion and tissue necrosis. A typical NK-cell phenotype is usually present: CD2+, CD3 epsilon+, CD56+, cytotoxic molecules+ and Epstein–Barr virus (EBV)+. Magnetic Resonance Imaging helps to clearly define the local involvement. Positron Emission Tomography helps to demonstrate system spread. Various prognostic variables (International Prognostic Index or the Korean Prognostic Index) should be documented. This may include quantification of plasma EBV DNA. For localized nasal disease, radiotherapy is important, although chemotherapy is often added. Sustainable remission is observed in over half of these patients. For extra-nasal or disseminated disease, systemic chemotherapy becomes the mainstay and the prognosis is usually poor. Doxorubicin-containing regimens are not entirely satisfactory and l-asparaginase containing regimens are being investigated. Patients with poor prognostic features may be considered for an early autologous haematopoietic stem cell transplant. Allogenetic transplantation is efficacious but is associated with high transplant-related mortality.

Natural killer-cell (NK-cell) neoplasms account for only a very small fraction of all non-Hodgkin lymphomas (Kwong et al, 1997a,b; Liang & Graham, 2008). Among the NK-cell disorders described in the World Health Organization (WHO) Classification of Tumours of Haematopoietic and Lymphoid Tissues, extranodal NK/T-cell lymphoma, nasal type, though still rare is the most common subtype (Kwong, 2005; Kwong & Liang, 2009). The tumour is characterized by extensive vascular destruction and prominent tissue necrosis. Most cases are genuine NK-cell neoplasms but a very small proportion of them show a cytotoxic T-cell phenotype (Kwong, 2005). Other even more uncommon NK-cell entities in the WHO Classification include the rapidly fatal aggressive NK-cell leukaemia and the clinically indolent chronic NK-cell lymphoproliferative disorders (Kwong & Liang, 2009). Patients with aggressive NK-cell leukaemia often run a rapidly downhill course and have an extremely poor clinical outcome. On the other hand, chronic lymphoproliferative disorders of NK-cells have a more chronic clinical course (Oshimi, 2007).

Extranodal NK/T-cell lymphoma, nasal type usually presents in the upper aero-digestive tract, especially the nasal cavity. Historically, different names were used to describe this extraordinary tumour (Liang et al, 1987, 1990, 1995;Liang, 2000, 2006). It has been called lethal midline granuloma, midline malignant reticulosis and polymorphic reticulosis (Ho et al, 1990a; Chim et al, 1999a). The lesion was not specified in the previous lymphoma classification systems: Rappaport, Kiel or the Working Formulation. In the Revised European and American Lymphoma (REAL) Classification, the term ‘angiocentric lymphoma’ has been used (Harris et al, 1994). However, it is well recognized that angiocentricity is not pathognomonic of NK-cell neoplasms.


Extranodal NK/T-cell lymphoma, nasal type has a unique geographic distribution (Chan et al, 1997b). It affects mainly populations in Asia (e.g. China, Japan and Korea) and Central/South America (e.g. Mexico and Peru). This lymphoma accounts for 5–10% of all the non-Hodgkin lymphomas seen in these populations (Au et al, 2005a). It is rarely diagnosed in other Western populations. The tumour is also uncommon in South Asia, Middle East and Africa and its aetiology remains uncertain. Genetic factors may be important (Au et al, 2005b). As the tumour has a very strong association with Epstein–Barr virus (EBV) infection, irrespective of the patients’ ethnic origin, it is likely that the virus also plays an important pathogenetic role (Ho et al, 1990b; Tao et al, 1994, 1995; Au et al, 2008).


The pathological features of this lymphoma, regardless of their primary sites, are remarkably similar (Chan et al, 2008). It is characterized by significant vascular damage and destruction. Histologically, there is a diffuse pleomorphic cellular infiltrate. The cell size tends to be quite variable. With very few exceptions, angiocentricity is usually a prominent feature. There is easily recognisable vascular invasion and occlusion by tumour cells, resulting in marked tissue ischaemia and necrosis (Fig 1).

Figure 1.

 Nasal NK/T-cell lymphoma with typical angiocentricity (Haematoxylin & Eosin, 400×).

The tumour usually has a characteristic NK-cell phenotype and is typically CD2 positive. Surface CD3 is negative but cytoplasmic CD3 epsilon is positive (Fig 2) and CD56 is usually positive (Fig 3). Cytotoxic molecules, including perforin, granzyme B and TIA-1, are also positive (Fig 4). CD56 is a useful NK-cell marker but is not specific, as it may also be expressed in some peripheral T-cell lymphomas (Chan et al, 1996, 2008). The availability of fresh or frozen tissue is helpful in differentiating an NK-cell from the less common cytotoxic T-cell phenotype. Negative surface CD3 and positive cytoplasmic CD3 epsilon are typical of true NK-cell neoplasm. Fresh frozen tissue specimens have to be available to make the distinction. The NK-cell phenotype is also supported by the finding of germ-line T-cell receptor gene rearrangement. On the other hand, the presence of a clonal T-cell receptor gene rearrangement suggests a cytotoxic T lymphocyte phenotype instead (Chan et al, 2008).

Figure 2.

 Nasal NK/T-cell lymphoma: CD3 epsilon positive.

Figure 3.

 Nasal NK/T-cell lymphoma: CD56 positive.

Figure 4.

 Nasal NK/T-cell lymphoma: cytotoxic molecule TIA1 positive.

Epstein–Barr virus infection is almost always demonstrable in the tumour cells by in-situ hybridization (ISH) for EBV-encoded early small RNA (EBER) (Fig 5) (Chan et al, 2008). The virus is found in clonal episomal form, supporting its important pathogenetic role. The virus also provides a good marker for diagnosis of this lymphoma (Ho et al, 1990b; Loke et al, 1992).

Figure 5.

 Nasal NK/T-cell lymphoma: EBER positive by in-situ hybridization.

Many different clonal cytogenetic and genetic abnormalities have been described in this lymphoma but they are usually non-specific (Siu et al, 1999, 2000, 2002a,b; Shen et al, 2002, 2003; Iqbal et al, 2009). There are suggestions that some tumour suppressor genes are involved and are possibly important in the pathogenesis. These genes may be deactivated by gene methylation (Siu et al, 2003; Iqbal et al, 2009).

Clinical features

This lymphoma is usually diagnosed in adults. Their median age is in the fifth decade, with a male to female ratio of three to one (Kwong & Liang, 2009). The tumour affects primarily the nasal cavity, nasopharynx, paranasal sinuses, tonsils, hypopharynx and larynx. It has to be distinguished from other lymphomas of the Waldeyer ring and paranasal sinsuses, which are often a diffuse large B-cell lymphoma. The lymphoma may extend locally to surrounding tissues and organs, such as the orbit and hard palate. Systemic dissemination is often late. Common metastatic sites include skin, gastrointestinal tract and genital organs (Au et al, 2008). In most case, the central nervous system and the bone marrow are not initially affected (Wong et al, 2001). Occult marrow involvement is occasionally detectable by ISH for EBV DNA (EBER staining) (Chim et al, 2005). Pancytopenia may sometimes be observed as the result of haemophagocytosis in patients without recognisable marrow involvement by the lymphoma (Hang et al, 2007). Common clinical presentations include a nasal mass, nasal obstruction and nasal bleeding. Other presenting symptoms include hoarseness of voice, dysphagia, proptosis, ophthalmoplegia, halitosis, airway obstruction and dysphonia, depending on the sites and extent of involvement.

In a small proportion of patients, the tumour affects primarily a non-nasal site, such as skin, gastrointestinal tract and genital tract (Au et al, 2009; Chan et al, 1997a; Chim et al, 1999b, 2001, 2002; Tse et al, 2002). The pattern is similar to the anatomical distribution of the metastatic sites of a primary nasal NK/T-cell lymphoma. For these patients, a careful examination of the nasal cavity is mandatory and may detect occult nasal involvement.

Clinical staging

There is currently no standard staging system for extranodal NK/T-cell lymphoma, nasal type. Like the other extranodal lymphomas, the Ann Arbor staging is unsatisfactory for this lymphoma. Stage I disease may include patients with a small tumour in the nasal cavity, as well as those with a large tumour invading locally to adjacent structures, including the paranasal sinuses and the nasopharynx (Kwong & Liang, 2009). Their prognosis is quite different.

A T-staging system originally designed for sinonasal B-cell lymphoma has been used and recommended for primary nasal NK/T-cell lymphoma (Robbins et al, 1985; Ooi et al, 2000; Kwong & Liang, 2009). T1 indicates that the tumour is confined to the nasal cavity. T2 is a tumour with an extension to the maxillary antra, anterior ethmoid sinus or hard palate. T3 is the presence of extension to posterior ethmoid sinus, sphenoidal sinus, orbit, superior alveolar bone, cheeks or superior buccinator space. T4 is involvement of the inferior alveolar bone, inferior buccinator space, infra-temporal fossa, nasopharynx or cranial fossa. This staging system provides a good indication of the extent of local tumour involvement. Patients with T1/T2 disease have been shown to have a better clinical outcome (Ooi et al, 2000).

The International Prognostic Index (IPI: age, stage, lactate dehydrogenase level, extranodal sites and performance status) was initially derived for diffuse large B-cell lymphoma. It also seems to be clinically useful for extranodal NK/T-cell lymphoma, nasal type. When applied retrospectively to cohorts of previously diagnosed extranodal NK/T-cell lymphoma, nasal type, the IPI appeared to be prognostically relevant (Chim et al, 2004; Au et al, 2009).

The Korean group has also designed a prognostic model for nasal NK/T-cell lymphoma. The parameters included in the Korean Index are stage, lactate dehydrogenase, systemic symptoms and regional lymphadenopathy. This has also been shown to be useful in predicting prognosis (Lee et al, 2006; Au et al, 2009).

Diagnosis and initial assessment

The initial presentation of a nasal NK/T-cell lymphoma often mimics a nasal sinus infection. The diagnosis can be easily missed or delayed. A high index of suspicion by clinicians is essential for early diagnosis. History and physical examination may reveal relevant symptoms and signs (Au et al, 2008; Kwong & Liang, 2009). For primary nasal disease, a careful evaluation by an otorhinolaryngologist is mandatory. Flexible nasal panendoscopy should be performed and multiple biopsies obtained from the involved and suspicious areas. Because of the anatomical site and the intrinsic property of the tumour, the biopsy specimens tend to be small and necrotic, making accurate pathological interpretation very challenging. Repeated examinations may be required and a larger piece of biopsy specimen preferred. The specimens should be sent to an experienced pathologist fresh instead of fixed for optimal analysis. A correct diagnosis often follows a meticulous morphological examination together with appropriate immunophenotyping and EBV zstudies (Au et al, 2008; Chan et al, 2008; Kwong & Liang, 2009).

Radiological assessment and monitoring

Computed tomography (CT) is conventionally used to assess the extent of local invasion of this lymphoma and also to detect distant metastases. However, it appears that magnetic resonance imaging (MRI) may define local soft tissue and bony involvement better (Ooi et al, 2000; Ou et al, 2007). The usefulness of positron emission tomography (PET) has also been evaluated. PET, when performed together with a CT scan, may define very well the tumour margin and at the same time detect occult systemic involvement (Fig 6) (Kako et al, 2007; Berk et al, 2008; Karantanis et al, 2008; Khong et al, 2008). Optimal imaging is critical for good radiotherapy planning and is also essential for monitoring of response to therapy.

Figure 6.

 PET/CT imaging showing nasal NK/T-cell lymphoma affecting right nasal cavity.

Circulating plasma EBV DNA level

Epstein–Barr virus infection is consistently associated with extranodal NK/T-cell lymphoma, nasal type. Like other EBV-associated tumours, the lymphoma may release EBV DNA into the circulation. This is mediated through apoptosis of proliferating tumour cells (Au et al, 2004; Kwong & Liang, 2009). Quantification of circulating plasma EBV DNA level by quantitative polymerase chain reaction serves as a surrogate marker of tumour load. It is potentially useful for initial prognostic determination, monitoring of disease response and detection of disease progression or relapse. A high level of circulating EBV DNA is correlated with higher tumour load, more extensive disease, poorer response to therapy and inferior survival (Au et al, 2004; Hsieh et al, 2007; Ishii et al, 2007).

Management of localized disease

For patients with clinically localized nasal disease, radiotherapy plays a very important role and should be given to all patients if permissible (Cheung et al, 2002, 2003; Li et al, 2006; Wang et al, 2007; Huang et al, 2008). The beneficial effect of radiotherapy is usually prompt and the overall response rate following this treatment is over 80%. Primary radio-resistance is uncommon. About 70% of all treated patients achieve complete remission. However, despite its initial efficacy, a high relapse rate of about 50% is still observed. Most of the relapses are local and usually occur early within the first year of treatment. Systemic failure occurs in a small proportion of the patients. These systemic relapses are not always associated with local recurrence.

The radiotherapy treatment demands careful planning and an optimal radiation dose (Isobe et al, 2006). A CT or, more preferably, an MRI should be used routinely to assist radiotherapy planning. A total dose ranging from 30 to 60 Gy with fractional doses of 1·5–2·5 Gy are often used. The optimal total dose is probably around 45–55 Gy. The radiation field should include at least the nasal cavity and the nasopharynx with a tumour margin of 1–2 cm. However, if the tumour is locally advanced, a wider radiation field becomes necessary. In those cases, some areas, such as the nasal sinuses and the cervical lymph nodes, may need to be included. The eyes should be spared if possible to preserve vision. This may prove to be difficult if the orbit is extensively involved. In this case, initial chemotherapy to shrink the tumour before radiotherapy is preferred. Radiotherapy is well tolerated by most patients, including the elderly. Common side effects include dermatitis, mucositis and transient loss of taste. In some patients, persistent nasal symptoms, such as discharge and bleeding, may still be observed as the result of radiotherapy, many years after successful treatment.

In view of the relatively high relapse rate when treated with radiotherapy alone, combination chemotherapy is often added to radiotherapy as part of the initial treatment (Au et al, 2008; Li et al, 2008; Kwong & Liang, 2009). CHOP (cyclophosphamide, doxorubicin, vincristine and prednisolone) chemotherapy is commonly used. Chemotherapy is often given first before radiotherapy. If this is the case, it is important that radiotherapy should not be delayed for too long. The patient must be monitored closely for disease response or progression while on treatment, with interval flexible nasal endoscopic examination and radiological imaging (CT or MRI). On reassessment, suspicious lesions should be biopsied. Random biopsies of the normal looking mucosa are recommended. All specimens should be carefully studied by an experienced pathologist. Morphological interpretation of these post-therapy specimens can be quite challenging. Immunophenotyping and EBER staining by ISH may help to detect residual disease. Incomplete response predicts an inferior clinical outcome (Au et al, 2008; Chan et al, 2008; Kwong & Liang, 2009).

In general, sustained remission is achievable in over half of the patients with localized nasal disease. Chemotherapy alone is not ideal for patients with a localized tumour and radiotherapy is an essential part of the management. For patients with a truly localized lesion and a small tumour load, the benefit of adding chemotherapy to radiotherapy may be doubtful (Kwong & Liang, 2009). Indeed, for this group of patients, giving initial radiotherapy of sufficient dose followed by ‘adjuvant’ chemotherapy is a reasonable and well accepted approach. Concurrent chemo-radiotherapy for localized nasal disease is also being studied in clinical trials. It would be interesting to see whether this will lead to superior clinical outcome with acceptable toxicities (Ribrag et al, 2001; Li et al, 2006; Suzuki et al, 2008).

For patients with primarily a localized non-nasal tumour, radiotherapy may also provide good local control of the lymphoma. However, it is recognized that non-nasal NK/T-cell lymphoma tends to disseminate early and rapidly. Therefore, initial systemic chemotherapy is preferred regardless of clinical stage. Local radiotherapy is still useful as an adjunct or for palliation. Primary non-nasal disease is generally associated with a poorer clinical outcome. Unlike patients with primarily localized nasal disease, sustained remission is achievable in only a very small fraction of them (Au et al, 2009).

Management of disseminated disease

Chemotherapy is the primary treatment for patients with systemic dissemination. Following the practice of treating B-cell lymphoma, CHOP chemotherapy is commonly used also for treating extranodal NK/T-cell lymphoma, nasal type, but the result is far from satisfactory (Au et al, 2008; Suzuki et al, 2008). There is no evidence that the more complicated doxorubicin containing regimens, such as BACOP (bleomycin, adriamycin, cyclophosphamide, vincristine, prednisone, m-BACOD (methotrexate, bleomycin, adriamycin, cyclophosphamide, vincristine, dexamethasone), MACOP-B (methotrexate with leucovorin rescue, doxorubicin, cyclophosphamide, vincristine, prednisolone and bleomycin) or ProMACE-CytaBOM (bleomycin, cyclophosphamide, cytarabine, doxorubicin, etoposide, leucovorin, methotrexate, prednisone, vincristine), are better in this setting. The high incidence of primary drug resistance with this tumour may be related to the expression of the multi-drug resistance (MDR) gene (ABCB1) with an increase in the amount of P-glycoprotein in the lymphoma cells (Yamaguchi et al, 1995). It may also be due to the marked tissue ischaemia and necrosis, resulting in poor penetration of the cytotoxic agents (Kwong & Liang, 2009).

It is therefore necessary to look for a better treatment alternative. Preliminary data have suggested that l-asparaginase is a good drug for treating this type of NK/T-cell lymphoma (Nagafuji et al, 2001; Matsumoto et al, 2003; Obama et al, 2003; Jaccard et al, 2009). The drug is not affected by P-glycoprotein. l-asparaginase works by depleting asparagines in the tumour cells. The tumour cells lack l-asparagine synthetase and are susceptible to the drug. Good responses were observed when l-asparaginase was used alone or in combination with prednisolone for relapsed or refractory cases (Yong et al, 2003, 2006, 2009). The combination of l-asparaginase, vincristine and prednisolone, has yielded a response rate of about 50% for nasal NK/T-cell lymphoma. This drug appears to have good activity against NK-cell tumours.

A retrospective study reported the treatment results of an l-asparaginase containing regimen as salvage therapy for 45 Chinese patients with refractory or relapsed extranodal NK/T-cell lymphoma, nasal type (Yong et al, 2009). There was an impressive complete response rate of 56%. The overall response rate was 82%. The 5-year survival was 67%. The treatment was well tolerated. l-asparaginase related side-effects included myelosuppression, impaired liver function, hyperglycaemia and allergic reaction (Yong et al, 2009). Similar benefit was also observed in Western patients. Fifteen patients with relapsed, refractory or disseminated extranodal NK/T-cell lymphoma and leukaemia were treated. An objective response was observed in 13 of them (87%), including seven complete responses (47%) (Jaccard et al, 2009).

The Japanese group has designed an l-asparaginase containing combination chemotherapy regimen called SMILE. The SMILE protocol consists of Steroids, Methotrexate and folinic acid, Ifosphamide, l-asparaginase and Etoposide. Pre-emptive use of granulocyte colony-stimulating factor was found to be essential with this regimen because of significant myelosupporession. The initial experience with this SMILE protocol has been encouraging. When used for treating refractory and relapsed cases, the overall response rate was 67% with a 50% complete response (Yamaguchi et al, 2008). The regimen is now being investigated for first line use in patients with both localized and disseminated NK/T-cell lymphomas (Yamaguchi et al, 2008).

In view of the radio-sensitivity of the tumour, it is important to note that local radiotherapy remains a useful adjunct even for patients with disseminated disease. It may be used to treat areas of bulky disease. Also, it is effective for symptomatic relief (Au et al, 2008; Kwong & Liang, 2009).

Autologous or allogeneic haematopoietic stem cell transplantation

High dose chemotherapy and autologous haematopoietic stem cell transplantation (HSCT) is widely used for treating all types of lymphoma. However, the experience of using an autologous transplant for treating extranodal NK/T-cell lymphoma, nasal type, is still limited. Further clinical trials are necessary before firm recommendations can be made. In view of the high rate of relapse following chemotherapy and radiotherapy and the fact that the bone marrow is often not involved, autologous transplant as consolidation therapy for this tumour is an attractive option (Li et al, 2004). Autologous transplantation appears to provide a survival benefit mainly to patients who have attained a complete remission at the time of transplant. Therefore, initial experience suggested that patients with poor prognostic features should be considered early for auto-transplant, preferably at the time of first complete remission (Au et al, 2003; Kim et al, 2006).

An international retrospective analysis has been performed to determine survival benefit of autologous transplant on patients with extranodal NK/T-cell lymphoma, nasal type (Lee et al, 2008a). A total of 47 transplanted patients were included in the study. The patients were followed up for a median period of 116·5 months. It was concluded that autologous transplantation conferred a survival advantage to patients transplanted in complete remission. It was recommended that those with poor risk factors at diagnosis should be considered for autologous transplantation after achieving a complete remission (Lee et al, 2008a).

Allogeneic transplantation using either a human leucocyte antigen-matched sibling/family donor or a matched unrelated donor is a tempting alternative to autologous transplantation. There is a theoretical advantage of graft-versus-lymphoma effect. The tumour almost invariably expresses EBV antigens, providing an extra allo-reactive target. The experience of allogeneic transplantation for this lymphoma is also limited. Some small series have demonstrated a disease-free survival of 30–50% but with a high transplant-related mortality of about 25% (Murashige et al, 2005; Suzuki et al, 2006; Suzuki et al, 2008; Kwong & Liang, 2009; Sato et al, 2008). A national survey on allogeneic transplantation for EBV-associated NK/T-cell lymphoproliferative diseases in Japan demonstrated that an allogeneic transplant could be curative with either a conventional or reduced intensity transplant (Sato et al, 2008).

Alternative stem cell sources have been used for transplanting these patients with some success. They included cord blood and mismatched donors (Suzuki et al, 2008). There was a recent report of successful unrelated cord blood transplantation for treating a relapsed case of advanced extranodal NK/T-cell lymphoma, nasal type (Yokoyama et al, 2007). The transplant was performed after a second remission was achieved by re-induction with an l-asparginase containing regimen. At the time of reporting, the patient had survived in remission for 33 months after the transplant (Yokoyama et al, 2007). Further clinical studies are essential to define the precise role of allogeneic transplantation for this lymphoma (Murashige et al, 2005; Suzuki et al, 2006; Sato et al, 2008).

Other issues

Nasal NK/T-cell lymphoma is often associated with marked tissue ischaemia and necrosis. The tumour may invade surrounding tissues, resulting in severe bone and soft tissue destruction. Even after successful therapy, damage to the neighbouring bony structures results in troublesome consequences. The collapsed orbital floor can be quite disfiguring and may affect vision. A perforated hard palate may disturb swallowing. Foods and drinks may regurgitate into the nasal cavity on oral feeding. Reconstructive surgery should be considered to correct the defects (Kwong & Liang, 2009).

In some places, such as Hong Kong, where extranodal NK/T-cell lymphoma, nasal type, is relatively more prevalent, up to 10% of the population are chronic hepatitis B virus (HBV) carriers. These patients are at high risk of HBV reactivation following chemotherapy. They may go into liver failure and the mortality is high. This is a serious but preventable complication. Patients at risk should be identified by checking their HBV status before they receive chemotherapy. Pre-emptive anti-viral therapy, such as lamivudine, is very effective in preventing HBV reactivation (Liang, 2009).

Concluding remarks

Extranodal NK/T-cell lymphoma, nasal type, has a peculiar geographic distribution. There are typical clinical, morphological and immunophenotypic features for its diagnosis. The strong association with EBV provides a good diagnostic marker. Local radiotherapy is an important treatment for patients with localized nasal tumour. For patients with non-nasal or disseminated disease, systemic chemotherapy is the mainstay. The efficacy of doxorubicin-containing chemotherapy regimens is not entirely satisfactory. Experimental protocols, which include l-asparaginase, appear to be promising and are now being tested in clinical trials. Other new approaches should also be explored (Shen et al, 2007; Ishitsuka et al, 2008; Lee et al, 2008b). The use of autologous or allogeneic HSC transplantation may be beneficial in selected patients but its precise role needs to be further defined by clinical trials.


This work is supported by the Carol Chau Ting Tong Research Fund of the University of Hong Kong.