Clinicopathologic Analysis of Localized Nasal/Paranasal Diffuse Large B-Cell Lymphoma

Diffuse large B-cell lymphoma (DLBCL) comprises 2 molecularly distinct subgroups of non-germinal center B-cell-like (non-GCB) and germinal center B-cell-like (GCB) DLBCLs, with the former showing relatively poor prognosis. In the present study, we analyzed the clinicopathological features of 39 patients with localized nasal/paranasal DLBCL. Immunohistochemistry-based subclassification revealed that 11 patients (28%) were of the GCB-type according to Hans’ algorithm and 11 (28%) were of the GCB-type according to Choi’s algorithm. According to both Hans’ and Choi’s algorithms, the non-GCB type was predominant. Nevertheless, prognosis was good. Overall survival did not differ significantly between the GCB and non-GCB subgroups (Hans’ algorithm: p = 0.57, Choi’s algorithm: p = 0.99). Furthermore, the prognosis of localized nasal/paranasal DLBCL was better than that of other localized extranodal DLBCLs. The prognosis of extranodal DLBCL is usually considered poorer than that of nodal DLBCL. However, in our study, no difference was noted between patients with localized nasal/paranasal DLBCL and patients with localized nodal DLBCL. In conclusion, although the non-GCB subtype is thought to show poor prognosis, in our study, the prognosis for localized nasal/paranasal DLBCL patients was good irrespective of subclassification.


Introduction
Although heterogeneous in nature, diffuse large B-cell lymphoma (DLBCL) can be classified into 2 distinct subtypes on the basis of genetic profiling: the germinal center B-cell-like (GCB) phenotype and the non-germinal center B-cell-like (non-GCB) phenotype [1,2,3]. Notably, patients belonging to the former group have a better prognosis than those belonging to the latter group. Hans et al. reported that these DLBCL subtypes can be easily distinguished on the basis of immunohistological staining for CD10, BCL6, and MUM1 proteins [2]. Later, Choi et al. added 2 new antibodies, FoxP1 and GCET1 [4], and Choi's algorithm is reported to achieve better prognostic classification than Hans' algorithm [5]. Extranodal non-GCB-like DLBCL is generally characterized by poor prognosis regardless of its localized disease, but localized primary non-tonsillar oral DLBCL exhibits favorable prognosis even in cases of the non-GCB subtype [6]. Nasal/ paranasal DLBCL is uncommon, and the GCB and non-GCB subtypes of this disease have not yet been examined. In this study, we aimed to clarify the clinicopathological features of localized nasal/paranasal DLBCL.

Patients
We selected 39 Japanese patients diagnosed with localized nasal/paranasal DLBCL between 1995 and 2010 and reviewed our institution's pathology department database to obtain the medical records of these patients. We only evaluated localized lymphomas, because the primary sites of advanced lymphomas are difficult to determine. All 39 cases were diagnosed as primary extranodal DLBCLs. Patients were defined as having extranodal DLBCL when the disease was confined to one or more extranodal sites and showed no (or only minor) nodal involvement after the staging procedures [7,8]. This group of patients was then compared with 39 patients with localized nodal DLBCLs diagnosed at our institution [9]. The samples and the medical records (clinical history, treatment and survival data) used in our study was approved by the Institute Review Board (IRB) at Okayama University. Written informed consent was waived by our institutional review board, since our study was limited to the use of excess human tissue samples and medical records.

Statistical Analysis
All data were analyzed using STATA software (version 9.0; Stata Corp., College Station, TX, USA). Actuarial overall survival curves were calculated using the Kaplan-Meier method, and differences were examined using the log-rank test to determine significant prognostic factors [10]. Overall survival was defined as the time from diagnosis to death from any cause or to the last follow-up visit.

Therapeutic Response and Outcome
Follow-up clinical data were available for 28 patients. The duration of follow-up ranged from 1 to 125 months (mean, 29 months). Fifteen patients were initially treated with chemotherapy plus irradiation, 9 were treated with chemotherapy alone, and 2 were treated with irradiation alone. Twenty-two patients achieved complete remission. Although 7 patients relapsed, 3 of

Comparison of the Clinicopathological Characteristics of Localized Nasal/Paranasal DLBCL and Localized Nodal DLBCL
The clinicopathological characteristics of nasal/paranasal DLBCL and localized nodal DLBCL are summarized in Table 5. Nasal/paranasal DLBCL patients showed good prognosis. The slight difference in the overall survival between these patients and patients with localized nodal DLBCL was not significant (p = 0.30) (Fig. 4). Moreover, analysis using the x2-test revealed a significant difference between the 2 groups with regard to age distribution and immunophenotype. Localized nasal/paranasal DLBCL patients were more likely to be more than 60 years old than localized nodal DLBCL patients (p = 0.018). In addition, localized nasal/paranasal DLBCL patients showed significantly higher positivity for MUM1 than localized nodal DLBCL patients according to Choi's algorithm (p = 0.00023, x2-test).

Discussion
DLBCL is the most frequent and aggressive lymphoma, representing a heterogeneous group that includes de novo large Bcell lymphomas, as well as transformed lymphomas from follicular or mucosa-associated lymphoid tissue lymphomas [11]. Recent studies have demonstrated that DLBCL can be further subclassified into 2 major prognostic categories according to Hans et al.: the GCB-type and the non-GCB-type [1,2].
However, Hans' algorithm has been superseded by a new algorithm devised by Choi et al., and results obtained using Choi's algorithm closely correlate with those of gene expression profiling for predicting prognosis [5]. In general, the non-GCBtype of DLBCL is associated with a significantly poorer prognosis than the GCB-type [1]; however, it has recently been established that this may not be true for extranodal DLBCL. Patients with localized primary non-tonsillar oral DLBCL presented with a  Table 3. Clinical and phenotypic characteristics of patients with GCB-type and non-GCB-type DLBCL.

Total (n = 39)
Hans' algorithm P Choi' algorithm p GCB (n = 11) non-GCB (n = 28) GCB (n = 11) non-GCB (n = 28) favorable clinical course despite having the non-GCB-type [6]. Similarly, in our study, the non-GCB-type of localized nasal/ paranasal DLBCL was the dominant type following subclassification according to both algorithms, but the prognosis of these patients was good. Moreover, the prognosis of localized nasal/ paranasal DLBCL was as good as that of primary cutaneous DLBCL [12] (p = 0.10) (Fig. 5) and was statistically better than that of other localized extranodal DLBCLs (CNS [13], testis [14], and adrenal gland [15]) (p = 0.0002, p = 0.0012, and p = 0.0044, respectively) (Fig. 6). Generally, extranodal non-GCB-like DLBCLs are characterized by poor prognosis, and the incidence of non-GCB-type DLBCLs among extranodal DLBCLs is 83-100%, although this value differs according to the organ of manifestation [16,17,18,19]. According to previous reports, DLBCLs of the central nervous system [16], breast [17], stomach [20], leg type [21], testis [18], and intravascular type [19] are predominantly of the non-GCB-type, an observation consistent with the finding in our study of localized nasal/ paranasal DLBCL cases. However, patients with CNS, breast, and testicular DLBCL exhibit poor prognosis, regardless of the localized disease [16,17,18,22]. As shown for primary cutaneous B-cell lymphoma, findings of genes expression analysis suggest that primary non-leg-type cutaneous DLBCL and primary cutaneous DLBCL, leg type have expression profiles similar to those of GCB-type and non-GCB-type DLBCLs, respectively [23]. Therefore, primary non-leg-type cutaneous DLBCL is predominantly associated with an excellent prognosis [12,24]. According to the recent World Health Organization (WHO) classification, subsets of DLBCLs arising in peculiar extranodal sites have been categorized as distinct disease subgroups (primary DLBCLs of the CNS, primary cutaneous DLBCLs, leg-type) or as distinct disease entities (primary mediastinal large B-cell lymphoma), on the basis of specific clinical and/or pathologic features [25,26]. When the cases in our study are included, extranodal disease is common among DLBCL patients [27]. It is thought that there are important clinical differences between nodal and extranodal DLBCL and that the most reliable distinction can be made in patients with stage I disease. For these patients, extranodal DLBCL is independently associated with poor survival [27]. Therefore, we also compared the clinicopathological profiles of localized nasal/paranasal DLBCLs with localized nodal DLBCLs. This analysis showed that localized nasal/paranasal DLBCL was associated with good prognosis and no difference was noted in the prognosis compared with localized nodal DLBCL. In recent years, the use of rituximab has improved the prognosis of DLBCL patients, and CHOP (cyclophosphamide, adriamycin, vincristine, and prednisolone) therapy combined with rituximab (R-CHOP) is currently a standard chemotherapy for DLBCL [28]. In our study, no significant difference was noted in the number of patients treated with rituximab between localized nasal/paranasal DLBCL and localized nodal DLBCL patients (p = 0.24). Therefore, the prognosis of localized nasal/paranasal DLBCLs was favorable regardless of treatment with rituximab. In conclusion, the prognosis of localized nasal/paranasal DLBCL patients was good irrespective of the disease subclassification, although the non-GCB-type of DLBCLs are usually thought to be associated with a poor prognosis.