International Scholarly Research Network ISRN Oncology Volume 2012, Article ID 652682, 7 pages doi:10.

5402/2012/652682

Research Article Prognostic Biomarkers and EBV Infection Research in Diffuse Large B-Cell Lymphoma of the Palatine Tonsils
Marinho Marques,1 Estela Luz,2 Michael Hummel,3 Maria das Gracas Vieira,2 4 Maria Cristina Oliveira,4 Eduardo Martins Netto,2 Ivana Luz,2 Regina C lia Bahia, e and Iguaracyra Araujo2
1 Servico

de Hematologia, N cleo de Oncologia da Bahia, Avenida Adhemar de Barros 123, Ondina, 40170-110 Salvador, u BA, Brazil 2 Hospital Universit rio Professor Edgard Santos, Universidade Federal da Bahia, Rua Augusto Viana s/n, 40110-060 Salvador, a BA, Brazil 3 Institute of Pathology, Charit Universit tsmedizin Berlin, Campus Benjamin Franklin, Hindenburgdamm 30, e a 12200 Berlin, Germany 4 Servico de Hematologia, Hospital Aristides Maltez, Avenida D. Jo o VI 332, Brotas, 40285-001 Salvador, BA, Brazil a Correspondence should be addressed to Marinho Marques, [email protected] Received 8 October 2011; Accepted 13 November 2011 Academic Editors: M. Emoto and N. A. Franken Copyright 2012 Marinho Marques et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Diuse large B-cell lymphoma represents approximately 30%40% of all diagnoses of non-Hodgkins Lymphoma and may represent up to 80% of all lymphomas that arise in the palatine tonsils. Several studies have attempted to correlate clinical, laboratorial, and tissue factors with the prognosis of the lymphomas, such as the International Prognostic Index, the tissue expression of some proteins, and the lymphocyte count at the time of diagnosis, as well as to correlate Epstein-Barr virus (EBV) infection with worse prognoses. Patients with palatine tonsil DLBCL, from Salvador, Bahia, Brazil, were studied in order to identify prognostic factors. Twenty-four patients with DLBCL were studied. The factors that negatively inuenced the patients survival rates were the lymphocyte count at the time of diagnosis <1.000/mm3 and the Bcl-2 protein expression. There was no CD5 expression in these lymphomas, and neither was there an association with EBV infection.

1. Introduction
The palatine tonsils, along with the nasopharyngeal lymphoid tissue, the base of the tongue, and the oropharyngeal wall make up Waldeyers ring. This ring is located at the entrance of the respiratory and digestive tract, being the second most common site of extranodal lymphomas, after the gastrointestinal tract [1, 2]. These tumors represent 15 to 20% of all lymphomas and half of the head and neck lymphomas. Approximately 50% of Waldeyers ring lymphomas arise in the palatine tonsils in presentation, and in approximately 20% of the cases they are bilateral [3]. Most lymphomas found in the palatine tonsils are the B-cell type, and, of these, diuse large B-cell lymphoma (DLBCL) represents most of the cases, reaching as much as 80% in some of the groups studied [3, 4].

Although morphologically indistinct, some molecular studies support the hypothesis that DLBCL makes up a heterogeneous group of lymphomas that has dierent prognostic implications [5]. Classically, the International Prognostic Index (IPI) has been used to predict the survival of patients with DLBCL [6]; however, it is not useful in all cases. Studies using DNA microarray analysis show that the DLBCL gene expression prole similar to B cells germinal center would have a better clinical prognosis than the prole similar to activated B cells [5, 7]. Transposing this gene prole to a protein expression, Hans et al. [8] proposed an algorithm to classify DLBCL patients, using three immunohistochemical markers (CD10, Bcl-6, and MUM1). The prole that was similar to the germinal center (GC) presented a better survival rate than the prole of the nongerminal center (non-GC). Other tissue markers, such

2 as Bcl-2, CD5, and p53 protein have been referred to as prognostic predictors in DLBCL [911]. A lymphocyte count <1.000 cells/mm3 at the time of diagnosis has recently been described as associated with a worse prognosis in DLBCL [12, 13]. Another controversial prognostic tissue marker has been the detection of Epstein-Barr virus (EBV) infection in neoplastic cells. While in pediatric Hodgkins Lymphoma this infection is associated with a better prognosis [14], in adults the infection in DLBCL seems to be associated with a worse prognosis [15]. EBV belongs to the herpes virus family, is transmitted through contact with saliva, and infects mostly B lymphocytes and occasionally oropharyngeal epithelial cells. Serum-epidemiological studies show that more than 90% of the patients with DLBCL, worldwide, are infected with EBV [16]. This infection has been detected in approximately 9% of patients with DLBCL [15]. However, in head and neck lymphomas, this index seems to be higher, reaching 90% in Egypt [17]. In Brazil, this infection has been associated with pediatric lymphomas originating in germinal center cells [18], although the index of this infection in adults with DLBCL has not yet been studied. Considering that the use of IPI has been shown to be insucient as the only prognostic marker in DLBCL, we evaluated other laboratory and tissue markers in patients with DLBCL of palatine tonsils, who came from a reference institution for the diagnosis and treatment of patients with cancer in Northeastern Brazil.

ISRN Oncology
Table 1: Antibodies used for immunohistochemical study. Antibody CD3 CD5 CD10 CD20 Bcl-2 Bcl-6 MUM1 p53 Ki67 Clone F7.2.38 4C7 56C6 L26 124 P1F6 MUM1p PAb1801 MIB-1 Dilution 1 : 100 1 : 50 1 : 50 1 : 100 1 : 50 1 : 20 1 : 50 1 : 100 1 : 50 Source DakoCytomation Novocastra Novocastra DakoCytomation DakoCytomation Novocastra DakoCytomation Novocastra DakoCytomation

2. Patients and Methods

2.1. Patient Selection. The patients were selected at the Pathological Anatomy Service of the Aristides Maltez Hospital in Salvador, Bahia, Brazil. All patients with a diagnosis of DLBCL of the palatine tonsils were included. The diagnoses were carried out between January of 1999 and December of 2006, on patients with illness primary site in the tonsils, or on those who presented the main tumor mass in the same area. The diagnoses were reviewed by a hematopathologist (I.A.) according to the criteria established by the lymphoid neoplasia classication of the World Health Organization 2008 [19]. The clinical data were gathered from medical chart records, and patients with incomplete medical charts were excluded. The staging was obtained using the Ann Arbor criteria [20], and the IPI was obtained according to previously established parameters [6]. This current work was approved by the Ethics Committee for Human Subjects Research and complies with the principles of the Helsinki Declaration. 2.2. Immunohistochemical (IHC) Study. The tissue used in the study had been set in formalin. Serial 4 m sections of tissue blocks were mounted on silanized slides, deparaphined with xylene, and cleansed in alcohol. The streptavidinbiotin-peroxidase technique was used, with previous, humid heat antigen recovery. A commercially available panel of antibodies was used (Table 1).

For the antibodies CD3, CD5, CD10, CD20, Bcl-2, Bcl-6, and MUM1, the cases with more than 10% of the tumor cells marked were considered positive. For the Ki-67 analysis, the cases were semiquantied on a positivity scale of 0 to 100%, according to the quantity of marked tumor cells per eld, in an increase of 400x. The p53 protein analysis was carried out using the Sannino and Shousha score [21]. The IHC classication according to the Hans et al. algorithm [8] in GC and non-GC prole was carried out using the C10, Bcl-6, and MUM1 markers. Patients who were positive for CD10 alone, or positive for CD10 and Bcl-6, were considered as GC proles. Patients who were negative for CD10 and Bcl-6 were considered as non-GC proles. Patients who were negative for CD10 and positive for Bcl-6 were classied after the MUM1 analysis. The cases that were negative for MUM1 were classied as GC proles, and the positive cases were classied as non-GC proles. 2.3. In Situ Hybridization. In situ hybridization for the detection of EBER 1 and 2 transcription of EBV was carried out in RNA-free conditions, using specic probes marked with digoxigenin, as previously described [18]. For positive controls, tissue from patients with Burkitt lymphoma and infectious mononucleosis was used, previously identied as positive for EBV. The sign considered positive was located in the nucleus. 2.4. Statistical Analyses. To compare the dierence between the two proportions, the chi-square test and Fisher exact test were used. The dierences between two means were analyzed by the Mann-Whitney test. The Kaplan-Meier survival analysis and the log-rank test were used to study the prognostic signicance of the utilized biomarkers. The overall survival (OS) was calculated from the date of diagnosis to the last evaluation, or date of death. Event-free survival (EFS) was calculated from the date of diagnosis until date of death, disease progression, or end of clinical followup. The P value was considered signicant when <0.05. The SPSS software version 16.0 was used to carry out all the calculations.

3. Results
3.1. General Characteristics of the Patients. During the time of the study 567 diagnoses of non-Hodgkins Lymphoma were

ISRN Oncology
Table 2: Clinical features of the patients. Parameters Gender Male Female Age (mean) Stage I/II III/IV B symptoms Performance status <2 2 Serum LDH Normal Elevated IPI 0 and 1 2 3 4 and 5 Frequency (%) 11 (45.8) 13 (54.2) 60 (range 1586) 13 (45.8) 6 (25) 12 (50) 23 (95.8) 1 (4.2) 16 (66.7) 2 (8.3) 15 (62.5) 2 (8.3) 1 (4.2) 0

3 Patients with GC prole presented better OS (64.2 versus 46.5 months) and EFS (63.5 months versus 46.1 months) than patients with non-GC prole, although with no statistical signicance (P = 0.36 and P = 0.33, resp.). The isolated positivity for the markers used in the Hans algorithm (CD10, Bcl-6 and MUM-1, Figures 2(a), 2(b) and 2(c), resp.) also did not have signicant inuence on patients survival rates. The Bcl-2 protein expression (Figure 2(d)) was found in 54.2% of the patients, and these were signicantly older than the patients who were negative (65 versus 44 years old, resp., P = 0.02). Patients who were positive for Bcl-2 presented signicant worse OS (38 months versus 80.8 months, resp.) and EFS (37.6 months versus 79.1 months, resp.) rates than negative patients (P = 0.03 and P = 0.04, resp. (Figure 3)). The p53 protein expression was found in 45.8% of the patients with predominance of the male gender (P = 0.03) and young patients (mean age of 49 years in positive patients versus 63 years in negative patients, P = 0.08). Patients with p53 protein expression presented greater OS and EFS rates than negative patients (68.6 months versus 43.3 months and 67.1 months versus 43.3 months, resp.), but there was no statistical signicance (P = 0.22 and P = 0.26, resp.). The mean Ki67 was 60%, varying from 30 to 100%. There was no CD5 expression in any of the patients studied. 3.4. In Situ Hybridization. All the patients presented negative in situ hybridization for EBER-1 and 2 transcription of EBV.

LDH: lactate dehydrogenase; IPI: International Prognostic Index.

made in the study institution. Of these, 253 were classied as DLBCL (44.6%). Twenty-six diagnoses of DLBCL of the palatine tonsils were made (4.6% of the total amount of DLBCL), and for this present work 24 cases of DLBCL of the palatine tonsils were studied, due to the exclusion criteria. Table 2 shows the characteristics of the study patients. Most of the patients were treated with a schema based on anthracycline (CHOP or similar ones). The average followup time of the patients was 43 months (range 1 to 104 months). The overall and event-free survival rates were on average of 43.5 months to 39.5 months, respectively. Most of the patients presented a low-risk IPI (Table 2), and, if compared to patients with none or 1 factor versus >1 factor, this index did not signicantly inuence the OS (71.6 months versus 31 months, P = 0.30, resp.) and the EFS (71.2 months versus 30 months, P = 0.26, resp.) of the patients. 3.2. Lymphocyte Count at the Time of Diagnosis. The mean lymphocyte count at the time of diagnosis was 1.980 cells/mm3 , varying between 354 and 3.922 cells/mm3 , and 18.2% of the patients presented a lymphocyte count <1.000 cells/mm3 . Patients with a lymphocyte count at the time of diagnosis 1.000 cells/mm3 presented OS (74.9 months versus 16 months) and EFS (74.7 months versus 8.2 months) signicantly greater than patients with a count lower than this value (P = 0.005 and P = 0.001, resp.) (Figure 1). 3.3. Immunohistochemistry. According to the algorithm of Hans et al. [8], 10 patients were classied with GC prole (41.6%), and 14 were classied with non-GC prole (58.4%).

4. Discussion
Among the non-Hodgkins lymphomas, we observed a DLBCL frequency similar to other studied series; however, the palatine tonsils were mostly attacked at a slightly greater frequency than that referred to in the literature [1, 2]. Similar to other studies, we observed a predominance of B lymphomas (mainly DLBCL) in this site, as well as more advanced age at the time of diagnosis, around 60 years [22]. We observed a predominance of tonsil lymphoma patients and low-risk IPI. When observed alone, the IPI did not show any signicance to predict survival rates, and this index might not be the most adequate prognostic factor for patients with extranodal lymphoma. However, the reduced number of patients in this study may have had an inuence on this analysis. Recent studies have shown the lymphocyte count at the time of diagnosis with a prognostic factor in DLBCL [12, 13], as already shown in other hematological neoplasias such as Hodgkins lymphoma [23], follicular lymphoma [24], and acute myeloid leukemia [25]. In the current study, this correlation was also observed, for patients with a lymphocyte count 1.000 cells/mm3 at the time of diagnosis presented signicantly greater OS and EFS rates (P = 0.005 and P = 0.001, resp.). DLBCL has several forms of presentation and is a heterogeneous entity seen as certain IHC marker expressions, which could give it a better or worse prognosis. Hans et al. [8] observed that patients classied as GC prole

ISRN Oncology

Overall survival 1 1

Event-free survival

0.8

Lymphocyte count 1000/mm3 Probability

0.8

Lymphocyte count 1000/mm3

Probability

0.6 P = 0.005

0.6 P = 0.001

0.4

0.4

0.2

Lymphocyte count <1000/mm3

0.2

Lymphocyte count <1000/mm3

0 0 20 40 60 80 Times (months)
(a)

0 100 120 0 20 40 60 80 Times (months)

(b)

100

120

Figure 1: Overall survival and event-free survival considering lymphocyte count at the time of diagnosis.

(a)

(b)

(c)

(d)

Figure 2: Immunohistochemical aspects of the palatine tonsils DLBCL: positivity for CD10 (a), Bcl-6 (b), MUM-1 (c), and Bcl-2 (d) antibodies.

ISRN Oncology
Overall survival 1 1 Event-free survival

0.8 Bcl-2 () Probability Probability 0.6

0.8 Bcl-2 () 0.6

0.4

Bcl-2 (+)

0.4

Bcl-2 (+)

0.2 P = 0.03 0 0 20 40 60 80 Times (months)

(a)

0.2 P = 0.04 0 100 120 0 20 40 60 80 Times (months)

(b)

100

120

Figure 3: Overall survival and event-free survival considering Bcl-2 protein status at the time of diagnosis.

(using an algorithm that regards the expression of C10, Bcl6, and MUM1 for classication) presented better survival rates than patients with non-GC prole. Patients, in this study, who had tonsil lymphoma and GC prole, were observed as having better survival rates; however, a statistical signicance was not reached, possibly due to the small number of patients. The isolated expression of algorithmic immunohistochemical markers did not show a correlation with survival rates. In a more consistent manner, the Bcl-2 expression in the pre-Rituximab era has proven to be an unfavorable prognostic factor [9]. In the present study this nding was conrmed, for this protein expression was linked to worse GS and EFS rates. In these patients, the Bcl-2 expression was present in a signicant way in older patients (P = 0.02), which might have inuenced OS and EFS rates negatively due to the combination of both factors. Still quite a controversial issue is the prognostic value of the p53 protein expression in patients with DLBCL. In two other Brazilian studies that researched this correlation [26, 27], only one was able to show a signicant dierence in OS [26]. In this current study, positivity was associated with better OS (with no statistical signicance), and the patients who were positive for p53 tended to be younger, which might have inuenced this observation. In the present study all the patients were negative for the CD5 marker. The expression of this marker has been reported as between 5%10% of the patients who have de novo DLBCL, granting a worse prognosis [10]. Due to the negativity we were unable to evaluate this data. However, in another study involving patients with DLBCL of the respiratory-digestive tract, the 17 patients with palatine

tonsil lymphoma were also negative for CD5 [28], similar to the ndings of this study. Several studies in the literature have demonstrated the association between EBV infection and some subtypes of non-Hodgkins lymphoma [15, 17]. In the patients with diuse large B-cell lymphoma, this association is undergoing conrmation, with the appearance of many studies seeking to reinforce the association with EBV. There is even a correlation with the prognosis of these lymphomas, which is an unfavorable prognostic factor in EBV-positive patients, presenting worse treatment response, as well as worse overall survival, and event-free survival rates [15]. No association was found between EBV infection and DLBCL in the patients of this study. This data is dierent from what was found by Park et al., who studied 380 patients with DLBCL, where 9% presented EBV infection [15]. Also, Bahnassy et al. showed EBV infection in 90% of the fty cases of head and neck NHL studied in Egypt [17]. Although it was shown previously that the tonsils of Brazilian children had signicantly more EBV-infected cells than the tonsils of German children [18], the negativity for EBV infection was found by Wong et al. when they studied 17 patients with palatine tonsil DLBCL in Malaysia [28]. We conclude that, in our midst, the palatine tonsil diuse large B-cell lymphomas predominantly presented a non-GC prole and were not associated with EBV infection. The factors that negatively inuenced the OS and EFS rates of these patients were the lymphocyte count at the time of diagnosis <1.000/mm3 and the Bcl-2 protein expression. New studies are necessary to attempt to explain the etiology of these lymphomas and to identify other prognostic factors.

ISRN Oncology
[13] L. F. Porrata, K. Ristow, T. M. Habermann, T. E. Witzig, D. J. Inwards, and S. N. Markovic, Absolute lymphocyte count at the time of rst relapse predicts survival in patients with diuse large B-cell lymphoma, American Journal of Hematology, vol. 84, no. 2, pp. 9397, 2009. [14] M. Engel, M. F. Essop, P. Close, P. Hartley, G. Pallesen, and C. Sinclair-Smith, Improved prognosis of Epstein-Barr virus associated childhood Hodgkins lymphoma: study of 47 South African cases, Journal of Clinical Pathology, vol. 53, no. 3, pp. 182186, 2000. [15] S. Park, J. Lee, Y. Ko et al., The impact of Epstein-Barr virus status on clinical outcome in diuse large B-cell lymphoma, Blood, vol. 110, no. 3, pp. 972978, 2007. [16] J. M. Middeldorp, A. Brink, A. van den Brule, and C. Meijer, Pathogenic roles for Epstein-Barr virus (EBV) gene products in EBV-associated proliferative disorders, Critical Reviews in Oncology/Hematology, vol. 45, no. 1, pp. 136, 2003. [17] A. A. Bahnassy, A.-R. N. Zekri, N. Asaad et al., EpsteinBarr viral infection in extranodal lymphoma of the head and neckcorrelation with prognosis and response to treatment, Histopathology, vol. 48, no. 5, pp. 516528, 2006. [18] I. Araujo, H. D. Foss, M. Hummel et al., Frequent expansion of Epstein-Barr virus (EBV) infected cells in germinal centres of tonsils from an area with a high incidence of EBVassociated lymphoma, Journal of Pathology, vol. 187, no. 3, pp. 326330, 1999. [19] S. H. Swerdlow, E. Campo, N. L. Harris et al., Eds., WHO Classication of Tumours of Haematopoietic and Lymphoid Tissues, IARC, Lyon, France, 2008. [20] S. A. Rosenberg, Validity of the Ann Arbor staging classication for the non-Hodgkins lymphomas, Cancer Treatment Reports, vol. 61, no. 6, pp. 10231048, 1977. [21] P. Sannino and S. Shousha, Demonstration of oestrogen receptors in paran wax sections of breast carcinoma using the monoclonal antibody 1D5 and microwave oven processing, Journal of Clinical Pathology, vol. 47, no. 1, pp. 9092, 1994. [22] A. Krol, S. Le Cessie, S. Snijder, J. C. Kluin-Nelemans, P. M. Kluin, and E. M. Noordijk, Waldeyers ring lymphomas: a clinical study from the comprehensive cancer center west population based NHL registry, Leukemia and Lymphoma, vol. 42, no. 5, pp. 10051013, 2001. [23] D. Hasenclever and V. Diehl, A prognostic score for advanced Hodgkins disease, New England Journal of Medicine, vol. 339, no. 21, pp. 15061514, 1998. [24] M. Siddiqui, K. Ristow, S. N. Markovic et al., Absolute lymphocyte count predicts overall survival in follicular lymphomas, British Journal of Haematology, vol. 134, no. 6, pp. 596601, 2006. [25] D. Behl, L. F. Porrata, S. N. Markovic et al., Absolute lymphocyte count recovery after induction chemotherapy predicts superior survival in acute myelogenous leukemia, Leukemia, vol. 20, no. 1, pp. 2934, 2006. [26] K. B. B. Pagnano, J. Vassallo, I. Lorand-Metze, F. F. Costa, and S. T. O. Saad, p53, Mdm2, and c-Myc overexpression is associated with a poor prognosis in aggressive non-Hodgkins lymphomas, American Journal of Hematology, vol. 67, no. 4, pp. 8492, 2001. [27] F. R. Kerbauy, G. Colleoni, S. Saad et al., Detection and possible prognostic relevance of p53 gene mutations in diuse large B-cell lymphoma. An analysis of 51 cases and review of

Conict of Interests
There is no conict of interests for any of the authors involved in this study.

Acknowledgments
The authors wish to thank all the employees of the Pathological Anatomy Service of the Aristides Maltez Hospital and Dr. William Harrington, Jr. (in memoriam) for their help in obtaining the immunohistochemical markers.

References
[1] A. Yuen and C. Jacobs, Lymphomas of the head and neck, Seminars in Oncology, vol. 26, no. 3, pp. 338345, 1999. [2] R. M. Nathu, N. P. Mendenhall, N. M. Almasri, and J. W. Lynch, Non-Hodgkins lymphoma of the head and neck: a 30year experience at the University of Florida, Head and Neck, vol. 21, no. 3, pp. 247254, 1999. [3] E. Zucca, E. Roggero, F. Bertoni, A. Conconi, and F. Cavalli, Primary extranodal non-Hodgkins lymphomas. Part 2: head and neck, central nervous system and other less common sites, Annals of Oncology, vol. 10, no. 9, pp. 10231033, 1999. [4] M. M. S. Neto, E. M. Jalil, and I. Araujo, Linfomas no-Hodgkin extranodais em Salvador-Bahia: aspectos a clnicos e classicacao histopatologica segundo a OMS 2001, Revista Brasileira de Hematologia e Hemoterapia, vol. 30, pp. 3640, 2008, http://www.scielo.br/scielo .php?script=sci arttext&pid=S1516-84842008000100010. [5] A. A. Alizadeh, M. B. Eisen, R. E. Davis et al., Distinct types of diuse large B-cell lymphoma identied by gene expression proling, Nature, vol. 403, no. 6769, pp. 503511, 2000. [6] M. Shipp, D. Harrington, and J. Anderson, The international non-Hodgkins lymphoma prognostic factors project. A predictive model for aggressive non-Hodgkins lymphoma, The New England Journal of Medicine, vol. 329, pp. 987994, 1993. [7] A. Rosenwald, G. Wright, W. C. Chan et al., The use of molecular proling to predict survival after chemotherapy for diuse large-B-cell lymphoma, New England Journal of Medicine, vol. 346, no. 25, pp. 19371947, 2002. [8] C. P. Hans, D. D. Weisenburger, T. C. Greiner et al., Conrmation of the molecular classication of diuse large B-cell lymphoma by immunohistochemistry using a tissue microarray, Blood, vol. 103, no. 1, pp. 275282, 2004. [9] J. Muris, C. Meijer, W. Vos et al., Immunohistochemical proling based on Bcl-2, CD10 and MUMI expression improves risk stratication in patients with primary nodal diuse large B cell lymphoma, Journal of Pathology, vol. 208, no. 5, pp. 714723, 2006. [10] M. Yamaguchi, M. Seto, M. Okamoto et al., De novo CD5+ diuse large B-cell lymphoma: a clinicopathologic study of 109 patients, Blood, vol. 99, no. 3, pp. 815821, 2002. [11] M. A. Piris, F. Pezella, J. C. Martinez-Montero et al., p53 and bcl-2 expression in high-grade B-cell lymphomas: correlation with survival time, British Journal of Cancer, vol. 69, no. 2, pp. 337341, 1994. [12] D. H. Kim, J. H. Baek, Y. S. Chae et al., Absolute lymphocyte counts predicts response to chemotherapy and survival in diuse large B-cell lymphoma, Leukemia, vol. 21, no. 10, pp. 22272230, 2007.

ISRN Oncology
the literature, Leukemia and Lymphoma, vol. 45, no. 10, pp. 20712078, 2004. [28] K. K. Wong, N. Prepageran, and S. C. Peh, Prognostic subgroup distribution in diuse large B-cell lymphoma of the upper aerodigestive tract, Pathology, vol. 41, no. 2, pp. 133 139, 2009.