62

Leukemia & Lymphoma, January 2013; 54(1): 62–68© 2013 Informa UK, Ltd.ISSN: 1042-8194 print / 1029-2403 onlineDOI: 10.3109/10428194.2012.710907

∗ Th ese authors are equal contributors. Correspondence: Michael Mian, MD, Department of Hematology, Hospital of Bolzano, Via Lorenz B ö hler 5, 39100 Bolzano, Italy. Tel: � 39-471-08807. Fax: � 39-471-08703. E-mail: mian@webmaxxi.eu

Received 27 May 2012 ; revised 26 June 2012; accepted 5 July 2012

ORIGINAL ARTICLE: CLINICAL

Role of radiotherapy in patients with early-stage diff use large B-cell lymphoma of Waldeyer ’ s ring in remission after anthracycline-containing chemotherapy

Michael Mian 1 ∗ , Andr é s J. M. Ferreri 2 ∗ , Andrea Rossi 3 , Annarita Conconi 4 , Richard Tsang 5 , Mary K. Gospodarowicz 5 , Elena Oldani 3 , Massimo Federico 6 , Stefano Luminari 6 , Enrico M. Pogliani 7 , Fausto Rossini 7 , Maria E. Cabrera 8 , Maurizio Martelli 9 , Gonzalo Gutierrez-Garcia 10 , Mario Busetto 11 , Franco Cavalli 12 , Emanuele Zucca 12 , Alessandro Rambaldi 3 & Sergio Cortelazzo 1 ; for the International Extranodal Lymphoma Study Group (IELSG)

1 Division of Hematology, Azienda Ospedaliera S. Maurizio, Bolzano/Bozen, Italy, 2 Unit of Lymphoid Malignancies, Department of Onco-Hematology, San Raff aele Scientifi c Institute, Milan, Italy, 3 Division of Hematology, Ospedali Riuniti, Bergamo, Italy, 4 Division of Hematology, Department of Clinical and Experimental Medicine and BRMA, Amedeo Avogadro University of Eastern Piedmont, Novara, Italy, 5 Department of Radiation Oncology, University of Toronto, Princess Margaret Hospital, Toronto, Canada, 6 Department of Oncology and Hematology, University of Modena and Reggio Emilia, Modena, Italy, 7 Department of Hematology, University of Milan-Bicocca, Milan, Italy, 8 Hospital del Salvador, Universidad de Chile, Santiago, Chile, 9 Department of Hematology, University of Rome, Italy, 10 Servicio de Hematologia, Hospital Clinic Universitar I, Barcelona, Spain, 11 UOC Radioterapia, UOS Radioterapia Oncoematologica, Ospedale dell ’ Angelo Mestre, Italy and 12 IOSI, Bellinzona, Switzerland

Introduction

Waldeyer ’ s ring (WR) consists of lymphoid tissue located in the pharynx to the back of the oral cavity and the base of the tongue, namely the pharyngeal and tubal tonsil as well as the palatine and lingual tonsils. Th is is the primary site of dis-ease in approximately 5 – 10% of all non-Hodgkin lymphomas (NHLs), being one of the most frequent forms of mucosa-associated lymphoma [1,2]. Th e most common subtype aff ecting the WR is diff use large B-cell lymphoma (DLBCL).

Radiotherapy (RT) is considered the mainstay of treat-ment for stage I/II NHL of the WR, leading to a 5-year overall survival (OS) of 30 – 50%. Nevertheless, � 40% of these patients experience relapse at sites not included in the primary radia-tion fi eld, leading to disseminated disease and death [3 – 5]. In the 1980s, Avil é s et al . conducted a randomized phase III trial [4] providing some evidence that combined modality treat-ment (CMT) leads to a signifi cant survival advantage when compared to chemotherapy (CHT) or RT as single strategies. However, this trial was conducted in a single center more than 20 years ago and included only patients with stage I disease; only half of the patients had DLBCL, and CMT con-sisted of primary RT followed by consolidation CHT, which is the inverse sequence with respect to the currently used chemoradiation treatment for DLBCL. Due to the rarity of

Abstract Consolidation radiotherapy (cRT) in patients with stage I/II diff use large B-cell lymphoma of the Waldeyer ’ s ring (WR-DLBCL) in complete remission (CR) after induction chemotherapy (CHT) is often associated with relevant acute and chronic toxicity, and its impact on survival remains to be defi ned. A total of 184 patients in CR after anthracycline-based chemotherapy were retrospectively analyzed: 62 underwent CHT alone (CHT group), while 122 (66%) patients were referred to cRT (CHT � RT group). After a median follow-up of 54 months, 36 patients (20%) experienced relapse: 19% in the CHT group and 20% in the CHT � RT group. At the time of analysis 47 (76%) CHT patients and 97 (80%) CHT � RT patients were alive. Five-year overall survival (OS), disease-free survival (DFS) and lymphoma-specifi c survival (LSS) were 80%, 74% and 86%, respectively. Five-year OS was signifi cantly prolonged in the CHT � RT group, while DFS and LSS were similar between groups. This discrepancy was attributed to a high percentage of deaths due to unrelated causes in CHT patients. cRT does not prolong LSS in patients with early-stage WR-DLBCL in CR after anthracycline-containing chemotherapy. An inter-national confi rmatory trial is warranted.

Keywords: Lymphoma , Waldeyer , chemotherapy , radiotherapy , DLBCL

Diff use large B-cell lymphoma of the Waldeyer's ring 63

WR-DLBCL, these data were confi rmed only in small, under-powered retrospective studies [5,6], sometimes including other histologic subtypes and/or other primary sites of disease [7 – 9]. Post-CHT RT of head and neck tumors is usu-ally associated with relevant acute and chronic side eff ects. Among others, xerostomia, the main side eff ect, leads to a serious impairment in quality of life of aff ected patients [10]. Importantly, hypothyroidism and second cancers arising within the irradiated volume constitute major late complica-tions, often with a negative eff ect on survival. Th is is of par-ticular importance in patients with WR-DLBCL who usually exhibit good prognosis and high cure rates. Th us, studies of large series of limited-stage WR-DLBCL aimed at identifying the best candidates to be managed with CHT alone, avoiding toxicity, are needed.

Herein, we report a large retrospective study of WR-DLBCL performed under the sponsorship of the International Extra-nodal Lymphoma Study Group (IELSG) with the primary goal of identifying patients in complete remission (CR) after primary anthracycline-based CHT who need consolidation RT (cRT) to achieve long-term remission and cure.

Patients and methods

Study population From 1985 to 2005, 14 international cancer centers referring to the IELSG collected clinical and therapeutic data retro-spectively from 184 consecutive patients aff ected by stage I/II WR-DLBCL who achieved a CR after upfront anthracycline-containing chemotherapy. Primary DLBCL of the WR was defi ned as DLBCL that aff ected primarily the WR with (stage II) or without (stage I) involvement of neck lymph nodes (cervical and supraclavicular). Patients with stage II and mediastinal and/or axillary lymphadenopathies were excluded. Staging procedures included at least complete physical examination, bone marrow biopsy and enhanced computed tomography (CT). Since in the early 1980s not all centers were equipped with a CT device, ultrasonography was used in some cases. Stage was defi ned according to the Ann Arbor system [11]. Histologic diagnosis was performed according to the World Health Organization (WHO) classifi -cation of 2001 [12], and therefore the histologic specimens of cases assessed before 2001 were revised by the various centers. All patients were treated with upfront anthracycline-containing CHT, achieving a CR. Sixty-two patients did not receive any additional treatment (CHT group), while 122 patients received additional cRT (CHT � RT group).

Th is study was performed following the principles of the Declaration of Helsinki, and was approved by our institu-tional review board.

Statistical analysis Th e study endpoint was the evaluation of diff erences in relapse rate, OS, disease-free survival (DFS) and lymphoma-specifi c survival (LSS) according to the diff erent treatment modalities (CHT vs. CHT � RT) and stages of disease (I vs. II). Th e actuarial durations of OS, DFS and LSS were plot-ted as curves according to the Kaplan – Meier method and defi ned according to Cheson et al . [13]. A log-rank test was

used to investigate the impact on survival of categorical vari-ables. Cox ’ s proportional hazards model [14] was applied to evaluate the impact on survival of treatment and other clini-cal variables. Th e χ 2 test was performed to assess the signifi -cance of diff erences between categorical variables. Medians were compared with the Mann – Whitney U -test.

A p -value of � 0.05 was considered statistically signifi -cant. All statistical analyses were performed with Statisti-cal Package for the Social Sciences (SPSS) software v.17.0.1 (SPSS, Inc., Chicago, IL).

Results

Patient characteristics Th e main patient characteristics of both treatment groups are summarized in Table I. Median age was 62 years (range, 14 – 85) for the CHT group and 51 years (range, 17 – 85) for the CHT � RT group. Despite a higher prevalence of patients � 60 years of age among the CHT group (58% vs. 31%), fewer than 30% of the patients of both groups had an intermediate- or high-risk International Prognostic Index (IPI � 1). A modi-fi ed IPI � 1 according to Miller et al . [15] was present in 47% of patients. In both groups, B symptoms occurred in � 15% and bulky disease ( � 10 cm) in � 10% of cases. No statistical diff erence was observed in stage distribution between the two treatment groups.

Treatment All patients were treated with a median of six cycles of an anthracycline-based chemotherapy regimen (3 – 12 cycles). Fifty-six percent of patients received � 6 cycles of CHT and 44% received � 6 cycles. Among patients treated without RT, 66% received � 6 cycles and 34% received � 6 cycles. Among patients treated with CHT followed by RT, 52% received � 6 cycles and 48% received � 6 cycles ( p � 0.12). Only fi ve patients received rituximab, and therefore this information was not considered for further analysis. Intrathecal chemo-therapy as central nervous system prophylaxis was admin-istered in 2/57 (3%) of the CHT group and 3/114 (3%) of the CHT � RT group.

Sixty-two (34%) patients did not receive any additional treatment after CHT (CHT group), while 122 (66%) received

Table I. Clinical characteristics of 244 patients according to treatment group.

Parameter

CHT group ( n � 62)

CMT group ( n � 122)

n Valid % n Valid %

Median age (range), years * 62 (14 – 85) 51 (17 – 85)Age � 60 years ∗ 36 62 58 38 122 31Male 34 62 55 75 122 61Female 28 45 47 39B symptoms 9 61 15 14 114 12Stage II 49 62 79 97 122 79Performance status 4 60 7 6 119 5LDH � UNL 7 55 13 11 99 11Bulky disease ( � 10 cm) 3 58 5 10 114 9ESR � UNL 20 48 42 29 84 34 β 2 -Microglobulin � UNL 3 41 7 6 51 12

LDH, lactate dehydrogenase; UNL, upper normal limit; ESR, erythrocyte sedimentation rate; CHT, chemotherapy; CMT, combined modality treatment. ∗ p -Value � 0.001.

64 M. Mian et al.

cRT. RT consisted of 4 – 9 MV photons, with a median dose of 4000 cGy (range 3000 – 5250). Th e irradiated volume included the WR in all cases, alone in 23% of irradiated cases, and asso-ciated with laterocervical � supraclavicular nodes in the rest. Sixty-nine percent of patients with stage I disease were irra-diated exclusively to the WR, while prophylactic lymph node irradiation was indicated in the rest. Th e neck was irradiated in 88% of patients with stage II disease. All patients but two completed the planned cRT: they interrupted RT because of toxicity after 1200 cGy and 1600 cGy, respectively.

Relapses After a median follow-up of 54 months (range, 3 – 216), 36 patients (20%) experienced relapse: 11 (19%) in the CHT group and 25 (20%) in the CHT � RT group. No local relapses were detected in patients with stage I disease; recurrences involved mostly secondary (distant) sites in both treatment

groups (Table II). Only one case of central nervous system (CNS) relapse was observed: this patient was managed with CHT � RT and had not received CNS prophylaxis. Overall, the 5-year DFS was 74% [Figure 1(A)], and no signifi cant dif-ferences between the two treatment groups were observed [5-year DFS: 73% in CHT group and 75% in CHT � RT group; p � 0.61; Figure 1(B)]. Relapse patterns and rates did not diff er signifi cantly between subgroups with diff er-ent stages of disease (Table II), with a 5-year DFS of 84% for patients with stage I disease and 71% for patients with stage II disease ( p � 0.12).

Th e impact on disease control of cRT was also analyzed separately in subgroups of patients divided according to stage of disease. In patients with stage I disease, the addition of cRT was not associated with signifi cant improvement in DFS, with a 5-year DFS of 73% and 89% ( p � 0.46), respec-tively, for CHT and CHT � RT subgroups [Figure 1(C)]. In

Figure 1. Disease-free survival curves for the whole series (A), for patients grouped according to treatment modality (B), for patients with stage I disease (C) and for patients with stage II disease (D) divided according to treatment modality.

Diff use large B-cell lymphoma of the Waldeyer's ring 65

common in patients with stage II lymphoma treated with CHT with respect to patients with stage II lymphoma treated with CHT � RT (16% vs. 4% of patients; p � 0.01).

Multivariate analysis In order to verify whether the observed survival diff erences in univariate analysis were independent from other clini-cal parameters, multivariate analyses for OS, DFS and LSS including age ( � 60 vs. � 60 years), performance status (Eastern Cooperative Oncology Group [ECOG] score � 1 vs. � 1), lactate dehydrogenase (LDH) serum level (elevated vs. normal), stage (I vs. II), bulky disease � 10 cm (yes vs. no) and use of cRT (yes vs. no) were performed. In all three mod-els, cRT was not independently associated with outcome. Elevated LDH was the only independent prognostic factor for DFS ( p � 0.02), OS ( p � 0.003) and LSS ( p � 0.02).

Th e analysis of the impact of cRT in subgroups of patients divided according to LDH serum level confi rmed that com-bined treatment did not provide a survival advantage in both low- and high-risk patients (data not shown).

Discussion

Th e current standard of care for localized WR-DLBCL con-sists of an anthracycline-containing chemotherapy regimen sustained by cRT on the initial site of disease. However, cRT often leads to relevant acute and long-term side eff ects [4,10]. Th is IELSG study shows that cRT could be omitted in patients with stage I and stage II WR-DLBCL who achieve a CR after anthracycline-based chemotherapy. Moreover, this large series shows that patients with elevated LDH need an intensifi ed treatment.

Th e strengths of this study were the large number of cases, the long follow-up time and the homogeneity of selected patients according to stage and histotype. Th is series is representative of the whole population aff ected by this lymphoma, with patient characteristics comparable to those of previously reported series [6,7], even if the distribu-tion of some variables may be slightly diff erent with respect to earlier reports, because most published studies included a small number of patients and various histologic subtypes [6,7,16]. Th e main limitations of this study are the same as those of every retrospective study: a mostly long accrual period, incomplete information on some variables consid-ered for analysis, and management heterogeneity. However, retrospective analysis of a large international series seems to be a valid strategy to improve the management of a

patients with stage II disease, the addition of cRT was not associated with signifi cant DFS advantage, with a 5-year DFS of 74% and 71% ( p � 0.67), respectively [Figure 1(D)].

Seventeen (47%) relapsed patients received an adequate salvage treatment, and eight (22%) of them achieved a sec-ond lymphoma remission; the remaining patients did not respond and died of lymphoma progression. Th e propor-tion of patients who received (45% and 48%) and responded (27% and 20%) to salvage treatment was similar between CHT and CHT � RT groups.

Survival At the time of the last database update, 47 patients in the CHT group (76%) were alive (45 in CR); seven (11%) died of lymphoma and eight (13%) died of unrelated causes while disease-free (Table III). In the CHT � RT group, 97 were alive (80%; all in CR); 17 (14%) died of lymphoma and seven (6%) died of other causes while disease-free. Th e 5-year OS and LSS of the whole series were 80% and 86%, respectively [Figure 2(A)]. According to treatment subgroup, the 5-year OS was 67% in the CHT group and 87% in the CHT � RT group ( p � 0.06), while the 5-year LSS was 81% and 88% ( p � 0.55), respectively [Figure 2(B)].

Th e impact of cRT was analyzed separately in subgroups of patients divided according to stage of disease. In patients with stage I lymphoma, the addition of cRT did not modify OS (5-year: 86% vs. 100%; p � 0.80) or LSS (5-year: 86% vs. 100%; p � 0.57) [Figure 2(C)]. Conversely, the addition of cRT was associated with signifi cantly improved OS in patients with stage II disease, with a 5-year OS of 61% and 83% ( p � 0.01), respectively, for CHT and CHT � RT groups, whereas no dif-ference was observed in LSS (5-year: 76% vs. 85%; p � 0.52) [Figure 2(D)]. Th ere were no diff erences in the distribution of causes of death between patients divided according to stage of disease. However, deaths for unrelated causes were more

Table II. Relapse rate according to diff erent treatment groups, site of relapse and stage of disease ∗ .

Th erapy group

Site of relapse

TotalSite of primary

diseaseOther

site Unknown

CHT Stage I ( n � 13) 0 2 (15%) 2 (15%)II ( n � 49) 2 (4%) 2 (4%) 4 (8%)

Total ( n � 62) 2 (3%) 4 (6%) 5 (8%) 6 (10%)CHT � RT Stage I (n � 25) 0 3 (12%) 3 (12%)

II (n � 97) 5 (5%) 11 (11%) 16 (16%)Total ( n � 122) 5 (4%) 14 (11%) 6 (5%) 19 (16%)

CHT, chemotherapy; RT, radiotherapy. * No statistically signifi cant diff erences were observed.

Table III. Cause of death according to treatment and stage of disease ∗ .

Th erapy group

Cause of death

TotalDisease Other

CHT Stage I ( n � 13) 1 (8%) 0 1 (8%)II ( n � 49) 6 (12%) 8 (16%) 14 (28%)

Total ( n � 62) 7 (11%) 8 (13%) 15 (24%)CHT � RT Stage I ( n � 25) 3 (12%) 3 (12%) 6 (24%)

II ( n � 97) 14 (14%) 4 (4%) 18 (18%)Total ( n � 122) 17 (14%) 7 (6%) 24 (20%)

CHT, chemotherapy; RT, radiotherapy. ∗ Causes of non-lymphoma-related deaths were two brain ischemias, one heart ischemia and one accident in the CHT group while in the other group there were recorded two deaths related to carcinomas and one to cardiac failure.

66 M. Mian et al.

Figure 2. Overall survival (left column) and lymphoma-specifi c survival (right column) curves for the whole series (A), for patients grouped according to treatment modality (B) and for patients with stage I disease (C) and patients with stage II disease (D) divided according to treatment modality.

very uncommon malignancy such as WR-DLBCL. Another important limitation regards the fact that treatments used in the period 1985 – 2005 were diff erent compared with strategies used currently in patients with DLBCL. In the last decade, there has been important technological progress in RT and changes in fi eld and doses. Currently a more confor-mal dose delivery is used, often through a positron emission tomograpy (PET)-guided procedure, with consequent bet-ter tolerability. Eff ectively, incidence, extension and sever-ity of xerostomia, tooth damage and stomatitis have been remarkably improved with the use of modern techniques. However, it is more diffi cult for such technological improve-ments to result in better systemic disease control, which is the main failure pattern seen in the studied patients. Moreover, conversely to the current standard approach to patients with DLBCL, most analyzed patients were treated with CHOP (cyclophosphamide, doxorubicin, vin-cristine and prednisone) without the monoclonal anti-CD20

antibody rituximab. However, this seems not to be an impor-tant drawback in assessing the role of RT in these patients, considering that there is some retrospective evidence sug-gesting that the addition of rituximab does not provide a clear survival advantage in patients aff ected by localized extra-nodal NHL (WR-DLBCL was classifi ed as extranodal NHL in that study) [17]. Nevertheless, the addition of rituximab could contribute to improved disease control in WR-DLBCL as was the case for localized DLBCL in the Southwestern Oncology Group (SWOG) trial [18], where the addition of rituximab to a standard CHOP regimen was associated with improved outcome (4-year OS of 88% and 92%, respectively, for patients treated with CHOP and R-CHOP). However, we can hypothesize that the addition of rituximab could fur-ther reduce the clinical benefi t of consolidation RT in the management of WR-DLBCL. As an additional limitation, a central pathology review was not performed. However, all participating centers demonstrate a lengthy experience in

Diff use large B-cell lymphoma of the Waldeyer's ring 67

Figure 2. (Continued).

lymphoma diagnosis and management, also regarding the active involvement of expert hemopathologists.

Th e most important fi nding of the present study is evi-dence that the addition of cRT in patients with limited-stage WR-DLBCL in CR after anthracycline-containing chemo-therapy is not associated with improved local control, DFS and LSS. In other words, according to the present study, these patients could be managed with chemotherapy alone, avoiding the risk of severe toxicity.

Th is is supported by the observation that the majority of the patients relapsed at sites diff erent from the primary dis-ease location. Since to date no analysis with a corresponding patient cohort has been published, it is diffi cult to compare our results with historic data. However, the most similar patient setting was published in 1995 by Avil é s et al . [4], who analyzed the outcome of 316 patients with stage I aggressive NHL of the WR who underwent RT alone, anthracycline-containing chemotherapy, or both. In contrast to our expe-rience, these authors reported a relapse rate of 46% in the

CHT group compared to nearly 21% in the CHT � RT group, but that study included variable lymphoma categories, diagnosed almost 30 years previously, and used an inverted sequence of combined treatment. A more recently reported series of 121 patients with DLBCL of the tonsil suggests that the addition of cRT after chemotherapy is associated with better disease control and survival [6]. However, in this Indian study, patients were not selected on the base of response after chemotherapy, resulting in the inclusion of 20% of patients without CR at the end of treatment, and 40% of patients treated with CHT alone did not achieve a CR, ren-dering any comparison with the present study impossible. Besides, when the authors analyzed the role of cRT in the 83 (68%) patients who attained a CR to the initial CHT, they observed a signifi cant increase in DFS for patients treated with cRT, whereas OS remained unchanged (5-year OS: 98% vs. 95%).

Before the introduction of rituximab, the standard of care for localized nodal aggressive DLBCL consisted of a short

68 M. Mian et al.

Potential confl ict of interest: Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.

Th is work was supported by the Autonomous Province of Bolzano, the local health authority and Alto Adige Bolzano-AIL Onlus.

References Hoppe RT , Burke JS , Glatstein E , et al . Non-Hodgkin ’ s lymphoma: [1]

involvement of Waldeyer ’ s ring . Cancer 1978 ; 42 : 1096 – 1104 . Yuen A , Jacobs C . Lymphomas of the head and neck . Semin Oncol [2]

1999 ; 26 : 338 – 345 . Banfi A , Bonadonna G , Ricci SB , et al . Malignant lymphomas [3]

of Waldeyer ’ s ring: natural history and survival after radiotherapy . Br Med J 1972 ; 3 : 140 – 143 .

Aviles A , Delgado S , Ruiz H , et al . Treatment of non-Hodgkin ’ s [4] lymphoma of Waldeyer ’ s ring: radiotherapy versus chemotherapy versus combined therapy . Eur J Cancer B Oral Oncol 1996 ; 32B : 19 – 23 .

Laskar S , Mohindra P , Gupta S , et al . Non-Hodgkin lymphoma of [5] the Waldeyer ’ s ring: clinicopathologic and therapeutic issues . Leuk Lymphoma 2008 ; 49 : 2263 – 2271 .

Laskar S , Bahl G , Muckaden MA , et al . Primary diff use large B-cell [6] lymphoma of the tonsil: is a higher radiotherapy dose required? Cancer 2007 ; 110 : 816 – 823 .

Chang DT , Mendenhall NP , Lynch JW , et al . Long-term outcomes [7] for stage I-II aggressive non-Hodgkin lymphoma of Waldeyer ’ s ring . Am J Clin Oncol 2009 ; 32 : 233 – 237 .

Ezzat AA , Ibrahim EM , El Weshi AN , et al . Localized non-Hodgkin ’ s [8] lymphoma of Waldeyer ’ s ring: clinical features, mana gement, and prognosis of 130 adult patients . Head Neck 2001 ; 23 : 547 – 558 .

Saul SH , Kapadia SB . Primary lymphoma of Waldeyer ’ s ring. [9] Clinicopathologic study of 68 cases . Cancer 1985 ; 56 :157 – 166.

Chang DT , Amdur RJ , Pacholke H , et al . Xerostomia in long-[10] term survivors of aggressive non-Hodgkin ’ s lymphoma of Waldeyer ’ s ring: a potential role for parotid-sparing techniques? Am J Clin Oncol 2009 ; 32 : 145 – 149 .

Carbone PP , Kaplan HS , Musshoff K , et al . Report of the Committee [11] on Hodgkin ’ s Disease Staging Classifi cation . Cancer Res 1971 ; 31 : 1860 – 1861 .

Jaff e E , Harris N , Stein H , et al . Pathology and genetics of [12] tumors of haematopoietic and lymphoid tissues . Lyon: World Health Organization ; 2001 .

Cheson BD , Pfi stner B , Juweid ME , et al . Revised response criteria [13] for malignant lymphoma . J Clin Oncol 2007 ; 25 : 579 – 586 .

Cox DR . Regression models and life-tables . J R Stat Soc B 1972 ;[14] 34 : 187 – 220 .

Miller TP , Dahlberg S , Cassady JR , et al . Chemotherapy alone [15] compared with chemotherapy plus radiotherapy for localized intermediate- and high-grade non-Hodgkin ’ s lymphoma . N Engl J Med 1998 ; 339 : 21 – 26 .

Lopez-Guillermo A , Colomo L , Jimenez M , et al . Diff use large [16] B-cell lymphoma: clinical and biological characterization and outcome according to the nodal or extranodal primary origin . J Clin Oncol 2005 ; 23 : 2797 – 2804 .

Gutierrez-Garcia G , Colomo L , Villamor N , et al . Clinico-biological [17] characterization and outcome of primary nodal and extranodal diff use large B-cell lymphoma in the rituximab era . Leuk Lymphoma 2010 ; 51 : 1225 – 1232 .

Persky DO , Unger JM , Spier CM , et al . Phase II study of rituximab [18] plus three cycles of CHOP and involved-field radio therapy for patients with limited-stage aggressive B-cell lymphoma: Southwest Oncology Group study 0014 . J Clin Oncol 2008 ; 26 : 2258 – 2263 .

Bonnet C , Fillet G , Mounier N , et al . CHOP alone compared with [19] CHOP plus radiotherapy for localized aggressive lymphoma in elderly patients: a study by the Groupe d ’ Etude des Lymphomes de l ’ Adulte . J Clin Oncol 2007 ; 25 : 787 – 792 .

Reyes F , Lepage E , Ganem G , et al . ACVBP versus CHOP plus [20] radiotherapy for localized aggressive lymphoma . N Engl J Med 2005 ; 352 : 1197 – 1205 .

Paes FM , Kalkanis DG , Sideras PA , et al . FDG PET/CT of [21] extranodal involvement in non-Hodgkin lymphoma and Hodgkin disease . Radiographics 2010 ; 30 : 269 – 291 .

Qi SN , Li YX , Wang H , et al . Diff use large B-cell lymphoma: [22] clinical characterization and prognosis of Waldeyer ring versus lymph node presentation . Cancer 2009 ; 115 : 4980 – 4989 .

course of CHOP chemotherapy followed by RT, since it has been proved to be superior to further cycles of CHT without RT [15]. However, these original observations have not been confi rmed by a more recent randomized trial [19], and, as mentioned above, the addition of rituximab to CHOP chemo-therapy could further improve OS and progression-free sur-vival (PFS), with the consequent defi nition of R-CHOP � RT as the new standard of care in localized nodal DLBCL [18]. Th is standard of care could also be applied to patients with WR-DLBCL, since specifi c trials focusing on this rare condition do not exist. Our data seem to suggest that cRT can be omitted in patients with WR-DLBCL in CR after primary chemotherapy, independent of the number of delivered chemotherapy cycles. In contrast to those reported in most trials focusing on local-ized DLBCL [15,18 – 20], the present study shows an evident plateau in the survival curves, suggesting that WR-DLBCL was already curable in the pre-rituximab era.

Finally, it has been shown that fl uorodeoxyglucose (FDG) PET-CT is eff ective in detecting DLBCL of the WR [21]. Th ere-fore, the now widespread use of this imaging modality could improve the identifi cation of a CR after chemotherapy and reduce even further the role of RT in these patients.

In the present study, DFS and LSS curves for both treat-ment groups were overlapping, while OS curves were diver-gent for both treatment groups in the whole cohort and in the subgroup of patients with stage II disease, with a signifi cantly better OS for CHT � RT patients. Actually, this discrepancy should be attributed to the diff erent distribution of causes of death among subgroups. In fact, more than half of the deaths in the CHT group were due to causes diff erent from DLBCL, while in the CHT � RT group lymphoma progression was the main cause of death. A possible explanation for the relatively high rate of non-lymphoma-related causes in the CHT group could be the fact that in this group median age was higher and, therefore, these patients had a higher risk for cardiovascular events and neoplasias. Moreover, the higher percentage of patients aged � 60 years in the CHT group can probably be ascribed to the intent to avoid RT toxicity in elderly patients. We can bona fi de exclude that these fi g-ures could have introduced an interpretation bias, since the rate of relapse and the percentage of disease-related deaths were similar in both treatment groups, whereas the rate of non-lymphoma-related deaths was more than twice in the CHT group.

As previously reported [4,22], an elevated LDH serum level was the only prognostic factor for DFS, OS and LSS, suggest-ing that these patients have more aggressive disease and that additional RT does not provide suffi cient disease control for these patients. Probably, these patients will require a more intensifi ed treatment to improve outcome.

In conclusion, cRT can be omitted in patients with stage I/II WR-DLBCL in CR after anthracycline-containing che-motherapy. A confi rmatory, international prospective trial is warranted.

Acknowledgements

Th e authors thank Armando L ó pez-Guillermo and Monica Bellei for their contributions.