The most typical clinical presentation of ENKL is ulceration and destruction because of the characteristic angioinvasion and necrosis. The median onset age is the mid-40s to 50 years. This disease is quite uncommon in children and adolescents. There is a male predominance, especially nasal cases (Au et al. 2009; Lee et al. 2005a, b, 2006; Suzuki et al. 2010). Because around 80 % of cases occur in the upper airway tract, common symptoms are nasal obstruction, ulceration in the nasal cavity and palate, hemorrhage, and discharge. Severe destruction of the hard palate can cause midline perforation, which is why it was previously called “lethal midline granuloma” (Fig. 7.2a, b). Extranasal disease occurs in the skin, gastrointestinal tract, testis, and other organs (Au et al. 2009; Lee et al. 2005a, b; Suzuki et al. 2010). In cutaneous ENKL, the most typical lesion is nonhealing ulceration (Fig. 7.2c). Gut perforation, bleeding, and acute abdomen are common symptoms in cases of gastrointestinal tract involvement. The lesions are characterized endoscopically by erosion, ulceration, or ulceroinfiltrative lesions without mass formation (Fig. 7.2d) (Kim et al. 2007). Interestingly, there are clinical differences between nasal and extranasal ENKL even though they share the same histological features. Patients with extranasal ENKL have more adverse clinical features including advanced stage, elevated lactate dehydrogenase (LDH) level, presence of B symptoms, and poor performance status (Au et al. 2009; Lee et al. 2005a; Suzuki et al. 2010).

Hepatic dysfunction caused by cytokines and/or hepatosplenic involvement can occur even in clinically stage I or II disease. Hemophagocytosis presenting as fulminant multiorgan failure with elevated LDH level, fever, and pancytopenia can occur; this is sometimes misdiagnosed as sepsis due to secondary infection from the ulcerative lesions.

No specific staging system for ENKL has been proposed. Ann Arbor staging is usually applied even though it is not satisfactory. For initial assessment, a complete history and physical examination are necessary (Kwong et al. 2009). Recent studies show the prognostic importance of local tumor invasiveness and regional lymph node involvement (Lee et al. 2006; Suzuki et al. 2010; Kim et al. 2005). The use of appropriate computed tomography or magnetic resonance imaging may provide good information about the extent of the disease. The role of positron emission tomography (PET) remains controversial. ENKL lesions are generally fluorodeoxyglucose avid. Therefore, it seems to be helpful to detect other systemic lesions before treatment. The role of posttreatment PET requires further studies (Kako et al. 2007; Fujiwara et al. 2011; Khong et al. 2008). To detect bone marrow involvement accurately, EBER in situ hybridization should be performed routinely in all patients with ENKL (Kwong et al. 2009; Lee et al. 2007). CNS involvement has been a concern for a long time because this lymphoma usually involves the nasal and paranasal sinuses. A recent report showed CNS involvement in <6 % of cases. Among patients at low risk (groups I and II by the KPI), CNS involvement was noted in <2 % of cases. Therefore, routine CNS evaluation is not recommended, although the role of CNS evaluation has not been determined for patients in KPI groups III and IV (Kim et al. 2010).

EBV DNA can be released from the apoptotic tumor cells. Serial monitoring of circulating EBV DNA can be a valuable marker of tumor burden and disease control. The test is recommended for staging, if quantitative Polymerase chain reaction (PCR) is available (Au et al. 2004; Kim et al. 2009a; Suzuki et al. 2011).

The IPI has good power for predicting the prognosis. However, most patients are classified into low- and low-intermediate-risk groups by IPI even though they have a poor prognosis. Besides the IPI, the following parameters including local tumor invasiveness, extra-upper aerodigestive tract origin, high proliferation index of the tumor tissue, and quantification of circulating EBV DNA seem to be prognostically relevant for individual patients (Kim et al. 2002, 2005, 2009a; Au et al. 2004; Suzuki et al. 2011). Therefore, other prognostic models have been proposed (Lee et al. 2006; Suzuki et al. 2010). Currently, the KPI, which is based on B symptoms, stage, LDH level, and regional lymph node involvement, has the strongest predictive power (Fig. 7.3) (Au et al. 2009; Lee et al. 2006).

Anthracycline-based chemotherapies such as CHOP or CHOP-like regimens are not efficient because of high expression of P-glycoprotein. Instead, drugs that are not affected by P-glycoprotein are recommended as a part of a polychemotherapy regimen. l-asparaginase-containing chemotherapies have produced promising outcomes recently. In cases of localized disease, the addition of radiation is important.