Accepted Manuscript. To appear in: Clinical Lymphoma, Myeloma and Leukemia. Received Date: 7 September Accepted Date: 18 December 2017

Transcription

1 Accepted Manuscript Consolidative Radiotherapy to Residual Masses After Chemotherapy is Associated With Improved Outcome in Diffuse Large B-cell Lymphoma. a Retrospective, Population-Based Study Øystein Fluge, Bård Mannsåker, Anders Torp, Ingvil Mjaaland, Lars Helgeland, Jan Klos, Olav Mella, Sigbjørn Berentsen, Peter Meyer PII: S (17) DOI: /j.clml Reference: CLML 1042 To appear in: Clinical Lymphoma, Myeloma and Leukemia Received Date: 7 September 2017 Revised Date: 27 November 2017 Accepted Date: 18 December 2017 Please cite this article as: Fluge Ø, Mannsåker B, Torp A, Mjaaland I, Helgeland L, Klos J, Mella O, Berentsen S, Meyer P, Consolidative Radiotherapy to Residual Masses After Chemotherapy is Associated With Improved Outcome in Diffuse Large B-cell Lymphoma. a Retrospective, PopulationBased Study, Clinical Lymphoma, Myeloma and Leukemia (2018), doi: /j.clml This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

2 Consolidative radiotherapy to residual masses after chemotherapy is associated with improved outcome in Diffuse Large B-cell lymphoma. A retrospective, population-based study. Authors Øystein Fluge, 1 Bård Mannsåker, 2 Anders Torp, 3 Ingvil Mjaaland, 4 Lars Helgeland, 5 Jan Klos, 6 Olav Mella, 1 Sigbjørn Berentsen, 7 Peter Meyer 4 Addresses 1 Department of Oncology and Medical Physics, Haukeland University Hospital, Bergen, Norway; 2 Department of Oncology and Palliative Medicine, Nordland Hospital, Bodø, Norway; 3 Department of Otolaryngology - Head and neck Surgery, Sørlandet Hospital, Kristiansand, Norway; 4 Department of Hematology and Oncology, Stavanger University Hospital, Stavanger, Norway; 5 Department of Pathology, Haukeland University Hospital; 6 Department of Pathology, Stavanger University Hospital; 7 Department of Internal Medicine, Haugesund Hospital, Haugesund, Norway. Keywords Diffuse Large B-cell lymphoma, External radiation, Response status, Prognosis, Survival Correspondence Øystein Fluge, MD, PhD Department of Oncology and Medical Physics Haukeland University Hospital Jonas Lies vei 65, N-5021 Bergen, Norway oystein.fluge@helse-bergen.no; oystein.fluge@gmail.com Telephone: Fax:

3 Microabstract The role of radiotherapy (RT) in Diffuse Large B-cell lymphomas with advanced disease is unclear. In this retrospective study, two cooperating university hospitals in Western Norway had different practices in use of RT, but similar patient characteristics and chemotherapy. RT to a residual mass after chemotherapy, in low-intermediate risk groups with advanced disease, was associated with improved cancer-specific and overall survival. Abstract Background: The role of consolidative radiotherapy (RT) in advanced Diffuse Large B-cell lymphomas (DLBCL) is not established. Methods: In a population-based retrospective analysis of patients with DLBCL in Western Norway during , 170 consecutive patients admitted to Haukeland University Hospital (HUS) and 94 to Stavanger University Hospital (SUS) were included. Mean age was 64 years (range years), 56% were male, 30% had stage I-II, 48% stage III-IV, and 22% had primary extranodal disease. Results: There were no differences between hospitals in patient characteristics, use of rituximab, number of chemotherapy courses or cumulative doses, or in distribution of response categories after chemotherapy. The use of RT was significant different: 23% received RT at SUS and 65% at HUS (p<0.001). For 219 patients with IPI-score 0-3, 5-year cancer-specific survival (CSS) was 67% at SUS and 81% at HUS (p=0.012). For 73 patients with complete response after chemotherapy there were no differences in survival between patients with and without RT. For 138 patients with any residual mass after chemotherapy, there were highly significant differences in favor of receiving RT (n=81) versus no RT (n=57): 5-year CSS 89% versus 69% (p<0.001), and 5-year OS 82% versus 59% (p=0.005). The impact of RT to residuals was evident in most subgroups, mainly in low to intermediate risk, but not in high-risk (IPI 4-5) patients. Conclusion: With the limitations of a retrospective study, these data suggest that consolidative radiotherapy may improve survival in DLBCL patients with a residual mass after chemotherapy, also in advanced disease. 2

4 Introduction Diffuse large B-cell lymphoma (DLBCL) is the most common type of aggressive lymphoma. 1 The prognosis of DLBCL has improved considerably with addition of the anti-cd20 monoclonal antibody rituximab to chemotherapy. The treatment approach has been summarized in review articles. 2,3 In DLBCL stage I-II disease, the role of consolidation radiotherapy (RT) after limited chemotherapy has been investigated in several randomized trials, 4,5 and the addition of rituximab to chemotherapy followed by RT in a phase II study. 6 Several retrospective studies of stage I-II DLBCL have suggested outcome benefit for the use of RT. 7-9 Thus, RT may be used as an integral part of the treatment of localized stage I-II disease, however still with controversies. 10,11 The role of RT in advanced DLBCL (stages III-IV) has not been clearly established from randomized trials, although several retrospective or prospective studies have suggested a role for consolidation RT In recent review articles discussing the role of RT in DLBCL, consolidation RT was suggested to be beneficial at achieving local disease control and improving overall outcomes, especially in specific clinical settings such as initially bulky disease, in advanced disease with residual masses after chemotherapy, or with extranodal bone involvement. 20,21 All lymphoma treatment in Norway is organized through the governmental hospitals in the four Regional Health Care Trusts. In Western Norway, serving a population of 1.1 million, two hospitals, Stavanger University Hospital (SUS) and Haukeland University Hospital HUS), are responsible for the management, in cooperation with local hospitals in Haugesund and Førde. The patients are treated according to national lymphoma guidelines. As quality assurance of patient management, we in collaboration performed a retrospective study of all patients registered with DLBCL in the Western part of Norway during the time period As the initial analyses showed a discrepancy in outcome between patients treated at the two institutions, an additional cohort of the patients treated in was investigated, and a thorough analysis of the material was done, trying to identify any differences in practice that could explain the findings. The patient characteristics, use of chemotherapy and distribution of response categories were very similar between the two hospitals SUS and HUS, while there proved to be a highly significant difference in use of RT, especially to residual masses after chemotherapy in stage III-IV disease. This made the data suitable for investigating the possible role of radiotherapy in DLBCL treatment. 3

5 Patients and methods This population-based retrospective study was approved by the Regional Ethics Committee in Norway ( , and 2011/855), and by the Norwegian Social Science Data Services (14866). All patients alive received written information, with passive consent. In 2009 and 2011, we retrospectively reviewed the medical records of all DLBCL patients admitted to the two university hospitals (SUS and HUS), for the time period The database was updated in April Among 264 patients included, 246 had histologically verified DLBCL, 4 T-cell rich DLBCL, 7 primary mediastinal DLBCL, and 7 follicular lymphoma grade 3B according to the WHO classification. 22 Patients with transformed low-grade lymphoma, Burkitt lymphomas, or primary CNS lymphomas were not included. The study period was before PET/CT was generally available in the Western part of Norway, and included the time period when rituximab was implemented as standard treatment in addition to chemotherapy ( ). Patient characteristics were determined from the medical records as interpreted by the treating physicians and radiologists. These characteristics included presence of bulky disease, staging based on the Ann Arbor classification, 23 including CT scans and bone marrow biopsy, International Prognostic Index (IPI)-scores, 24 and response categories. 25 We did not undertake a formal review of all CT scans prior to, during, and after end of chemotherapy. Thus, as judged retrospectively, the interpretations of either complete response (CR) or complete response unconfirmed (CRu) may to some extent vary between physicians. To investigate the role of RT to residual masses after chemotherapy, we categorized patients as either complete response (CR) versus patients with any residual mass. The latter included partial response (PR), complete response unconfirmed (CRu), but also more uncertain residual masses such as relatively small residuals unchanged in size over time (i.e. between 3 and 6 chemotherapy courses), residual masses with no signs of vital lymphoma on biopsy, or residual masses negative by PET/CT after chemotherapy. The chemotherapy regimens used were R-CHOP or CHOP-based (the latter in 2003, before rituximab with chemotherapy was implemented as standard therapy). Patients received consolidation RT at the discretion of the treating physician. When administered, consolidation RT was generally started 3-4 weeks after chemotherapy was finished, usually given to one or sometimes two target volumes, usually as conformal RT with photon beams, and with a fraction dose of 1,8 Gy to totally 36 Gy. 4

6 Statistical analyses Categorical characteristics were compared between patients with or without RT as part of the treatment, using Chi-square statistics. Continuous variables were assessed using independent t-tests. Survival was calculated using the Kaplan-Meier product limit method, with p-values from the log-rank test. Survival was defined as the time from the date of start of lymphoma treatment, to the date of event (death) or date of last follow-up (censored). For cancer-specific (i.e. lymphoma-specific) survival (CSS) events were defined as death from lymphoma, or death from any other cause with lymphoma. Overall survival was defined as death from any cause. All probability values were two-sided. Statistics were performed using the SPSS ver. 24. Results An overview of the study is shown in Figure 1. The upper part of Figure 1 summarizes patient and treatment characteristics for the patient cohort, also by hospitals SUS and HUS, with the resulting hypothesis that the observed difference in 5-year CSS between hospitals could be ascribed to a highly significant difference in use of consolidative RT. The mid part of Figure 1 summarizes patient and treatment characteristics, and survival, for patients with or without RT in primary treatment, also by response status after chemotherapy. The lower part of Figure 1 focuses on patients with stage II-III-IV with a residual mass after chemotherapy, with the aim of identifying patients with a possible benefit from consolidative RT. The patient cohort Patient and treatment characteristics A total of 264 patients were identified with DLBCL at the two university hospitals SUS (n=94) and HUS (n=170) between January 2003 and December The mean age was 64.4 years (range years), 117 (44.3%) were women and 147 (55.7%) were men. Stage I- II were present in 30.4%, stage III-IV in 47.9%, and 21.7% had primary extranodal (PE) I-II disease. Bulky tumor was present in 78 patients (29.5%), and 103 (39%) had B-symptoms. Lactate dehydrogenase (LDH) was above normal range in 152 patients (57.6%). International Prognostic Index (IPI)-score 0-1 was detected in 41.7%, IPI-score 2-3 in 41.3% and IPI-score 4-5 in 15.5%. There were no significant differences between the two hospitals SUS and HUS for any of these patient characteristics. 5

7 Among 260 patients receiving chemotherapy, the mean number of chemotherapy courses was 5.78 (range 1-10). The chemotherapy regimens were R-CHO(E)P based in 200 patients (76%) and CHO(E)P based in 55 (21%), while five patients received other chemotherapy regimens, and four patients aged above 85 years did not want chemotherapy. During the chemotherapy period, 24 patients (9.2%) died, and 25 patients (9.6%) had primary progressive disease (PD). There were no significant differences in number of chemotherapy courses, in cumulative doses (mg/m 2 ) of cyclophosphamide or doxorubicin, or in use of rituximab, between the two hospitals SUS and HUS (data not shown). Excluding four patients not given chemotherapy, response evaluation after end of chemotherapy was assessed as complete response (CR) in 73 out of 260 patients (28.1%). Any residual mass was detected in 138 patients (53.1%), including 22 patients (8.5%) with partial response (PR), and 116 patients (44.6%) with CRu. The latter CRu category also included 46 patients with uncertain residual masses, comprising 17 patients with residuals but without histologic evidence of lymphoma on biopsy, 4 patients with residuals but negative PET/CT scan, and 25 relatively small residuals with no change in size over time during the chemotherapy period (e.g. between 3 and 6 courses). There were no differences between broad response categories between the two hospitals SUS and HUS (data not shown). While there were no differences between hospitals SUS and HUS for patient characteristics, use of chemotherapy or rituximab, or response categories after chemotherapy, there were highly significant differences in use of consolidative RT. For the analyses of consolidative RT, patients with PD or patients having died during the chemotherapy period were excluded, because they were not at risk for receiving consolidative RT. Among 74 patients at SUS, 17 (23.0%) received consolidative RT, while among 141 patients at HUS, 92 (65.2%) received RT (p<0.001). By pooled data from the two hospitals, RT in the primary treatment was given to 109 out of 215 patients (50.7%), and the indications were as follows: in 50 patients (23.3%) to a residual mass/initially bulky disease, in 33 patients (15.3%) to extranodal disease, and in 23 patients (10.7%) as initially planned treatment in stage I-II after chemotherapy. In addition, 3 patients who did not want chemotherapy received palliative RT. Outcome The median follow-up (time interval between start of lymphoma treatment and time of death or last contact) was 5.5 years (range years) for all 264 patients. Among 139 patients alive at censoring, the median follow-up was 7.3 years (range years), and in 125 patients dying before April 2015, the median follow-up was 1.5 years (range years). 6

8 For all 264 patients, 5-year overall survival (OS) was 62% (SE 0.03%). The 5-year cancerspecific survival (CSS) was 69% (SE 0.03%), with no significant difference between the two hospitals (64% at SUS versus 72% at HUS, p=0.14). There was a significant difference in cancer-specific survival between the two hospitals for the 219 patients with IPI-score 0-3 (5- year CSS 67% at SUS versus 81% at HUS, p=0.012). Among 41 patients with IPI-score 4-5, there were no differences in 5-year CSS (40% at SUS versus 39% at HUS, p=0.58), or in 5- year OS (30% at SUS versus 39% at HUS, p=0.33). With versus without radiotherapy Patient characteristics by consolidative RT For analyses including consolidative RT, patients with PD (n=25) or patients dying during the chemotherapy period (n=24) were excluded. In addition, four patients aged years who refused chemotherapy were excluded. Thus 211 patients were included in analyses with or without radiotherapy, comprising 138 patients with a residual mass after chemotherapy, and 73 patients with complete response (CR). For 73 patients with CR, there were no significant differences between those not receiving RT and those receiving RT, for any of the patient characteristics (Table 1), but with a trend for difference in distribution of stage (Table 1). Among the 138 patients with any residual mass after chemotherapy, the groups receiving RT (n=81), or not receiving RT (n=57), were well balanced regarding age, IPI-scores (0-3 vs 4-5), bulky disease, or elevated LDH. There was a higher proportion of women, and lower proportions of stage III-IV and of B-symptoms, in those receiving RT (Table 1). Comparing chemotherapy treatment characteristics, with or without RT, analyses were limited to stages II-III-IV to make meaningful comparisons (Table 1). Both in patients with CR, and in patients with a residual mass after chemotherapy, there were no significant differences between the No RT and With RT groups, either in mean doxorubicin or cyclophosphamide cumulative doses, or in the use of rituximab, or in the mean number of chemotherapy courses, (Table 1). Outcome by consolidative radiotherapy Among 211 patients (all stages I-IV and primary extranodal (PE), and all response categories, but excluding PD and patients dying during the chemotherapy period) 5-year CSS was 82% in those not receiving RT and 89% in those receiving consolidative RT (p=0.044) (Figure 2A). 7

9 For the 73 patients with CR after chemotherapy, we found no significant differences without RT versus with RT, in 5-year CSS (98% versus 92%), or in 5-year OS (85% versus 92%), respectively (Figure 2B, Table 1). In contrast, in the 138 patients with any residual mass after chemotherapy, there were highly significant differences in outcome between those not receiving RT and those receiving consolidative RT. Among 57 patients not receiving RT, 5-year CSS was 69%, while among 81 patients receiving RT 5-year CSS was 89% (p<0.001) (Figure 2C, Table 1). Correspondingly, 5-year OS was 59% in those without RT and 82% in those with RT (p=0.005) (Figure 2D, Table 1). Stage II-III-IV disease with a residual mass The difference in use of consolidative RT between the two hospitals SUS and HUS was most evident for patients with stages II-III-IV disease and a residual mass after chemotherapy. This group comprised 94 patients, and their characteristics, with or without consolidative RT, are shown in Supporting Table S1A. There were no significant differences for mean age, sex, stage (II versus III versus IV), IPI-score (0-2 versus 3-5), bulky disease, elevated LDH (< 300 versus > 300), number of chemotherapy courses, or doxorubicin or cyclophosphamide cumulative doses. There was a lower frequency of B-symptoms in those receiving RT (Supporting Table S1A). The Kaplan-Meier plots for these 94 patients with stages II-III-IV disease and a residual after chemotherapy are shown in Figure 2E (CSS) and Figure 2F (OS), with a 5-year CSS 88% in those with RT versus 68% in those without RT (p=0.003) (Figure 2E, Table 2). 5-year OS was 84% in those with RT, versus 57% in those without RT (p=0.014) (Figure 2F). Out of these 94 stage II-III-IV patients, 60 had a clear residual mass (PR or CRu). In these 60 patients (i.e. excluding the 34 uncertain residuals), 41 patients (68%) received consolidative RT. At time of censoring, cancer-specific death had occurred in 5 out of 41 patients (12%) given RT, and in 10 out of 19 patients (53%) not given RT (p<0.001) (Figure 2G), with the corresponding plot for overall survival in Figure 2H. In 34 out of the 94 patients, the CRu residuals were interpreted by the treating physician as uncertain, of these only 3 patients (9%) received RT. These 34 patients included 12 with biopsy negative residuals, 19 with small residuals unchanged over time during chemotherapy, and 3 patients with PET negative residuals. Relapse occurred in 12 (35%) out of these 34 patients, including 4 with biopsy-negative residuals. At censoring, cancer-specific death had 8

10 occurred in 10 out of 34 patients (29%), including 4 deaths after 5 years follow-up. Of these 10 patients with lymphoma-specific death, 9 did not receive consolidative RT. 5-year CSS for these 94 patients with stage II-III-IV disease and a residual mass, with or without RT, and by subgroups, is shown in Table 2. There was significantly improved CSS by receiving consolidative RT across most subgroups. Generally, except for high-risk patients with IPI-score 4-5, 5-year CSS was approximately 20% higher in the groups receiving consolidative RT (Table 2). The benefit in 5-year CSS with RT was most evident for the lowintermediate risk groups (Table 2). In the high-risk patients with IPI-score 4-5, 5-year CSS was 86% with RT compared to 79% without RT (p=0.47). In these 94 patients, we investigated the prognostic impact of the other clinical variables for CSS (Supporting Table S1B) and OS (Supporting Table S1C). There were no significant differences in 5-year CSS for subgroups within age (< 65 vs 65 years), sex, stage (II vs III vs IV), IPI-score (0-3 vs 4-5), bulky disease, or LDH (< 300 U/L vs 300), or within cyclophosphamide (< 4500 vs > 4500 mg/m 2 ) or doxorubicin (< 300 vs > 300 mg/m 2 ) cumulative doses (Supporting Table S1B). For 5-year OS, there was as expected a significant impact for age, otherwise no significant differences for subgroups within the other clinical variables (Supporting Table S1C). Discussion In this retrospective study comprising virtually all adult patients with DLBCL in the Western part of Norway during the time interval , patient and treatment characteristics were very similar between the two university hospitals SUS and HUS, except for the use of radiotherapy during primary treatment, especially to a residual mass after chemotherapy and most evident in stage III-IV disease. When analyzing the complete patient cohort, there was no significant difference in 5-year CSS between hospitals. However, when excluding the high-risk patients (IPI 4-5) patients, a significant difference in CSS was seen in favor of HUS. There were no significant differences between hospitals for clinical prognostic factors, and further no differences in use of rituximab, in number of chemotherapy courses, or in doxorubicin and cyclophosphamide cumulative doses. Thus, we hypothesized that the observed improved cancer-specific (i.e. lymphoma-specific) survival at HUS, for low-intermediate risk DLBCL patients, could be explained by a highly significant difference in the use of consolidative RT to a residual mass after chemotherapy. 9

11 This study was population-based, with all patients admitted to the two hospitals in the 6-year period included. The hospitals followed the same national lymphoma guidelines for use of chemotherapy, with similar treatment characteristics for chemotherapy, but with a very different use of radiotherapy during primary treatment. Of note, use of radiotherapy and especially to DLBCL in stages III-IV, was not specified in detail in national lymphoma treatment programs, thus the utilization of radiotherapy in this clinical setting became highly dependent on local traditions and preferences. While the use of chemotherapy and rituximab usually is described thoroughly in national and international protocols for clinical studies of DLBCL with advanced disease, the use of RT is often only briefly mentioned. The main limitation of this study is the retrospective design. Response categories have been reported as interpreted by the treating physician according to the patient files. There may be differences in interpretations between radiologists and individual treating physicians. This could be a limitation of the study, however it reflects decision making in daily practice outside of formal clinical studies. We focused on the specific group of patients with DLBCL stages II-III-IV with any residual mass after chemotherapy, with the aim to elucidate use of consolidative RT in this group. To be able to analyze distribution of patient and treatment characteristics, possible prognostic factors, CSS and OS, and effects of consolidative RT specifically to this group, patients with no residual mass (complete response) were excluded. DLBCL stage I and primary extranodal disease were excluded because in these patients RT may often be initially planned as part of the treatment strategy, and therefore be less dependent on the response status after chemotherapy. Patients with primary progressive disease (PD) and patients dying during the chemotherapy period were excluded because these were not at risk for consolidative RT. In patients with a clear residual mass (PR or CRu), there were highly significant and surprisingly large differences of approximately 20% (varying from 15% to 40% in subgroups) in CSS in favor of consolidative RT. These differences in survival could not be explained by uneven distribution of clinical prognostic factors between the groups with and without RT, by differences in chemotherapy courses, chemotherapy dose intensity, or by the use of rituximab. Thus, the data in this study although limited by retrospective design, strongly suggest a benefit of consolidative RT to a residual mass after chemotherapy. Whether the group of patients with an uncertain residual mass could benefit from radiotherapy, remains elusive from this study. Only 3 patients (9%) out of the 34 patients with uncertain residuals received consolidating RT. Among the 31 patients with uncertain 10

12 residuals not receiving consolidative RT, 5-year CSS was 84%. Four late lymphoma-specific deaths occurred after the 5-year follow-up, so during the complete follow-up a total of 9 lymphoma-specific deaths occurred, and the 8-year CSS was 64%. This contrasts to 90% 5- and 8-year CSS among 60 patients with clear residual masses given consolidative RT, and 41% 5- and 8-year CSS for patients with clear residual masses not receiving RT. Among the patients with uncertain residuals, lymphoma relapse occurred in 4 out of 12 patients with a negative biopsy after chemotherapy, in 7 out of 19 patients with small residuals unchanged in size during chemotherapy, and in 1 out of 3 patients with negative PET imaging after chemotherapy. This could indicate that also patients with uncertain residuals may benefit from consolidative RT. Although to our knowledge no randomized study has been published investigating the role of RT in advanced DLBCL, several retrospective studies have addressed this issue. The role of consolidative RT after R-CHOP therapy was investigated in a retrospective study of 469 DLBCL patients among whom 60% had advanced disease, reporting a 5-year OS and progression-free survival (PFS) of 89% and 76% for stage III-IV patients treated with consolidative RT, versus 66% and 55% in those without RT. 18 Among 43 patients with residual masses by CT-scan but a negative PET after chemotherapy receiving RT, OS and PFS were comparable to patients who achieved a complete response. However, in contrast to our present study, the authors also reported a significant benefit of RT in patients with a complete response after chemotherapy. 18 A retrospective study performed before the rituximab-era, in DLBCL stage III-IV patients with bulky or semi-bulky lesions in complete remission after chemotherapy, found improved outcome in those receiving consolidative RT. 15 Also addressing the role of consolidation RT in DLBCL stage III-IV with negative imaging after R-CHOP chemotherapy (assessed by PET/CT, gallium with CT, or only CT scans), consolidation RT (median 25 Gy) to involved sites was associated with improved in-field control and event-free survival, but with no difference in OS compared to those without RT. 17 In patients with no residual mass and complete response to chemotherapy, we found no significant benefit from consolidative RT, for either CSS or OS. A prospective study in elderly DLBCL patients compared the best arm of the RICOVER-60 trial, with involved-field RT (36 Gy) to initial bulky or extranodal sites, to a cohort (RICOVER-noRTh) receiving the same chemotherapy without RT, and reported improved outcome with consolidative RT

13 Use of RT to DLBCL in the rituximab era was investigated retrospectively in a large American study. 19 It included 841 patients, 35% received RT after R-CHOP based chemotherapy, with improved 5-year OS and failure-free survival (FFS) with RT compared to without RT. In the MInT study, young good-prognosis DLBCL patients with stages II-IV or bulky stage I were randomized to 6 CHOP with or without rituximab, and patients with bulky or extranodal disease (40%) received consolidative RT. 26 Indirect comparison of data from the MInT study with a similar patient population from the GELA LNH03-2B trial receiving either R-CHOP-21 or a more intensified chemotherapeutic regimen but without RT 27, suggested better outcome and a possible beneficial effect of RT, underscoring the need for a randomized study. 26 The possible role of consolidation RT in DLBCL, also in advanced stage III-IV disease, has recently been discussed. 28 However, the role of RT in this clinical setting is still not established. Our study was performed before PET/CT was generally available in Western Norway. Therefore, during the study time interval only a few patients had PET imaging at the end of chemotherapy, a procedure now routinely performed in most DLBCL patients. The current status of PET imaging in DLBCL has been reviewed. 29 The ability for PET/CT to predict outcome at end of chemotherapy is supported by several retrospective studies, 30 and new criteria for response assessment using PET/CT has eliminated the CRu category. 31,32 A retrospective study of 300 DLBCL patients, however, evaluated PET versus CT for response characterization after chemotherapy. Seventy-three per cent of the patients had received R- CHOP and none had RT. Five-year OS, disease-specific survival (DSS) and progression-free survival (PFS) were higher in patients with CR by both PET and CT (85%, 87%, 84%, respectively) than in patients with CR only by PET (66%, 71%, 64%, respectively), and compared to those with residual disease by both PET and CT (35%, 40%, 42%, respectively). 33 Thus, the presence of a residual mass on CT scan, despite negative findings on PET, was associated with inferior outcome, and the authors concluded that any residual mass after chemotherapy more than 2 cm in diameter should be regarded with suspicion, and biopsy or consolidative RT should be considered. 33 Although in our study PET/CT was only performed in a minority of patients, our data support the conclusions of that study showing an impact of a residual mass and advocating consideration of consolidative RT. A recent review exploring the use of RT in DLBCL in advanced stages, suggests that withholding consolidative RT in PET-based complete responses is premature. 34 The late toxicity from RT has been recognized, especially for patients treated with large mediastinal fields. For younger patients with primary mediastinal DLBCL, efforts should 12

14 focus on eliminating or optimizing radiation therapy due to possible late toxicity. In a phase-ii study, using the dose-adjusted EPOCH-R regimen for primary mediastinal DLBCL, 5-year event-free survival (EFS) was 93% and OS 97%, thus obviating the need for consolidative RT. 35 However, in most patients with DLBCL, with a mean age approximately 65 years at diagnosis, and for example one residual mass after chemotherapy e.g. in the retroperitoneal area, the clinical impact of possible late toxicity is very different from the young patient with a bulky mediastinal tumor. In the present study, we found a surprisingly large difference in survival in favor of RT in patients with advanced disease and a residual mass after chemotherapy. The absolute differences in overall and cancer-specific survival were approximately 20%, and in univariate analyses evident across most subgroups except the high-risk (IPI 4-5) patient group. The traditional rationale for consolidative RT to residual masses after chemotherapy involves enhanced local lymphoma control by eradication of remaining clonogenic tumor cells through damage to DNA and permanent proliferative arrest. The relatively large difference in survival in favor of consolidative RT observed in this study, given to one or sometimes two target volumes, leads to the hypothesis that systemic anti-tumor effects could be activated from sequential multimodal immunochemotherapy followed by radiotherapy. Recently, the knowledge of biological mechanisms for inhibition and activation of immune responses to tumor cells, with the immune checkpoint inhibitor systems and the important role of tumor stroma, has increased rapidly. Increasing focus has also been attributed to immunogenic cell death (ICD) and the interaction between RT and tumor-specific adaptive immune responses operating also outside the radiation target volume. ICD involves a variant of apoptosis including release of damage-associated molecular patterns (DAMP) stimulating an adaptive tumor-specific immune response. 36 Doxorubicin, cyclophosphamide and RT have been associated with induction of ICD and adaptive immune responses against dying cells. 37,38 A possible synergy of RT and immune checkpoint inhibitors has been suggested. 39 Thus, the surprisingly large survival benefit associated with consolidating RT in our study might also be influenced by an RT-associated ICD and enhanced adaptive anti-lymphoma immune response to primarily surviving lymphoma cells outside of the radiated volumes. In conclusion, with the limitations of a retrospective study, the data in this study supports accumulating data indicating that consolidative radiotherapy may improve survival in patients with DLBCL of advanced stage with a residual mass after chemotherapy. 13

15 Clinical Practice Points -In patients with Diffuse Large B-cell lymphomas (DLBCL), the role of radiotherapy (RT) is considered unclear, especially in advanced stages. -In this retrospective, population-based study from two university hospitals in Western Norway, with similar patient and chemotherapy treatment characteristics, but very different use of RT, a difference in cancer-specific survival was observed. -In patients with complete response (no residual mass) after chemotherapy, use of RT was not associated with improved outcome. -In patients of stage II-III-IV with any residual mass after chemotherapy, the use of RT was associated with improved cancer-specific and overall survival, most evident in low and intermediate risk groups, but not in high-risk (IPI 4-5) patients. -In patients with DLBCL and residual masses after chemotherapy, consolidative radiotherapy should be considered. Acknowledgements The study was supported by The Norwegian Cancer Society (grant to Øystein Fluge no ), and from the Folke Hermansen Foundation. The funders had no role in study design, data collection, data analysis, data interpretation, or the decision to submit for publication. Disclosures The authors have stated no conflicts of interest. 14

16 Figure Legends Figure 1 Outline of retrospective study of Diffuse Large B-cell lymphoma (DLBCL) in Western Norway Abbreviations: SUS: Stavanger University Hospital; HUS: Haukeland University Hospital; DLBCL: Diffuse Large B-cell lymphoma; CSS: Cancerspecific survival; OS: Overall survival; RT: Radiotherapy; PD: Primary progressive disease; PE: Primary extranodal; CR: Complete response; PR: Partial response; CRu: Complete response unconfirmed; LDH: Lactate dehydrogenase; IPI: International Prognostic Index. Figure 2 Survival of Diffuse Large B-cell lymphoma (DLBCL) patients, with or without radiotherapy (RT). The number of patients included (stages and response categories) are indicated on each panel. Patients with primary progressive disease (PD) and those dying during the chemotherapy period were excluded. (A) Cancer-specific survival (CSS) in 211 patients; (B) Cancer-specific survival in 73 patients with complete response after chemotherapy; (C) Cancer-specific survival in 138 patients with any residual mass after chemotherapy; (D) Overall survival (OS) in 138 patients with any residual mass after chemotherapy; (E) Cancer-specific survival in 94 stage II-III-IV patients with any residual mass after chemotherapy; (F) Overall survival in 94 stage II-III-IV patients with any residual mass after chemotherapy; (G) Cancer-specific survival in 60 patients with a clear residual mass (PR or CRu) after chemotherapy; (H) Overall survival in 60 patients with a clear residual mass (PR or CRu) after chemotherapy. Tables Table 1 Patient and treatment characteristics by response status after chemotherapy, with or without radiotherapy. Table 2 5-year cancer-specific survival for patients with DLBCL stage II-III-IV disease and a residual mass, with or without radiotherapy. 15

17 References 1. Morton LM, Wang SS, Devesa SS, Hartge P, Weisenburger DD, Linet MS. Lymphoma incidence patterns by WHO subtype in the United States, Blood. 2006;107: Kubuschok B, Held G, Pfreundschuh M. Management of diffuse large B-cell lymphoma (DLBCL). Cancer Treat Res. 2015;165: Bachy E, Salles G. Treatment Approach to Newly Diagnosed Diffuse Large B-Cell Lymphoma. Semin Hematol. 2015;52: Miller TP, Dahlberg S, Cassady JR, et al. Chemotherapy alone compared with chemotherapy plus radiotherapy for localized intermediate- and high-grade non- Hodgkin's lymphoma. N Engl J Med. 1998;339: Horning SJ, Weller E, Kim K, et al. Chemotherapy with or without radiotherapy in limited-stage diffuse aggressive non-hodgkin's lymphoma: Eastern Cooperative Oncology Group study J Clin Oncol. 2004;22: Persky DO, Unger JM, Spier CM, et al. Phase II study of rituximab plus three cycles of CHOP and involved-field radiotherapy for patients with limited-stage aggressive B- cell lymphoma: Southwest Oncology Group study J Clin Oncol. 2008;26: Ballonoff A, Rusthoven KE, Schwer A, et al. Outcomes and effect of radiotherapy in patients with stage I or II diffuse large B-cell lymphoma: a surveillance, epidemiology, and end results analysis. Int J Radiat Oncol Biol Phys. 2008;72: Kwon J, Kim IH, Kim BH, Kim TM, Heo DS. Additional Survival Benefit of Involved-Lesion Radiation Therapy After R-CHOP Chemotherapy in Limited Stage Diffuse Large B-Cell Lymphoma. Int J Radiat Oncol Biol Phys Vargo JA, Gill BS, Balasubramani GK, Beriwal S. Treatment Selection and Survival Outcomes in Early-Stage Diffuse Large B-Cell Lymphoma: Do We Still Need Consolidative Radiotherapy? J Clin Oncol Campbell BA. The role of radiation therapy in the treatment of stage I-II diffuse large B-cell lymphoma. Curr Hematol Malig Rep. 2013;8: Bindra RS, Yahalom J. The important role of radiation therapy in early-stage diffuse large B-cell lymphoma: time to review the evidence once again. Expert Rev Anticancer Ther. 2011;11: Held G, Zeynalova S, Murawski N, et al. Impact of rituximab and radiotherapy on outcome of patients with aggressive B-cell lymphoma and skeletal involvement. J Clin Oncol. 2013;31: Held G, Murawski N, Ziepert M, et al. Role of radiotherapy to bulky disease in elderly patients with aggressive B-cell lymphoma. J Clin Oncol. 2014;32: Schlembach PJ, Wilder RB, Tucker SL, et al. Impact of involved field radiotherapy after CHOP-based chemotherapy on stage III-IV, intermediate grade and large-cell immunoblastic lymphomas. Int J Radiat Oncol Biol Phys. 2000;48: Ferreri AJ, Dell'Oro S, Reni M, et al. Consolidation radiotherapy to bulky or semibulky lesions in the management of stage III-IV diffuse large B cell lymphomas. Oncology. 2000;58: Wilder RB, Rodriguez MA, Tucker SL, et al. Radiation therapy after a partial response to CHOP chemotherapy for aggressive lymphomas. Int J Radiat Oncol Biol Phys. 2001;50:

18 17. Dorth JA, Prosnitz LR, Broadwater G, et al. Impact of consolidation radiation therapy in stage III-IV diffuse large B-cell lymphoma with negative post-chemotherapy radiologic imaging. Int J Radiat Oncol Biol Phys. 2012;84: Phan J, Mazloom A, Medeiros LJ, et al. Benefit of consolidative radiation therapy in patients with diffuse large B-cell lymphoma treated with R-CHOP chemotherapy. J Clin Oncol. 2010;28: Dabaja BS, Vanderplas AM, Crosby-Thompson AL, et al. Radiation for diffuse large B-cell lymphoma in the rituximab era: Analysis of the National Comprehensive Cancer Network lymphoma outcomes project. Cancer. 2015;121: Boyle J, Beaven AW, Diehl LF, Prosnitz LR, Kelsey CR. Improving outcomes in advanced DLBCL: systemic approaches and radiotherapy. Oncology (Williston Park). 2014;28: , Shi Z, Das S, Okwan-Duodu D, et al. Patterns of failure in advanced stage diffuse large B-cell lymphoma patients after complete response to R-CHOP immunochemotherapy and the emerging role of consolidative radiation therapy. Int J Radiat Oncol Biol Phys. 2013;86: Swerdlow SH, Campo E, Harris NL, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC Press Carbone PP, Kaplan HS, Musshoff K, Smithers DW, Tubiana M. Report of the Committee on Hodgkin's Disease Staging Classification. Cancer Res. 1971;31: A predictive model for aggressive non-hodgkin's lymphoma. The International Non- Hodgkin's Lymphoma Prognostic Factors Project. N Engl J Med. 1993;329: Cheson BD, Horning SJ, Coiffier B, et al. Report of an international workshop to standardize response criteria for non-hodgkin's lymphomas. NCI Sponsored International Working Group. J Clin Oncol. 1999;17: Pfreundschuh M, Kuhnt E, Trumper L, et al. CHOP-like chemotherapy with or without rituximab in young patients with good-prognosis diffuse large-b-cell lymphoma: 6-year results of an open-label randomised study of the MabThera International Trial (MInT) Group. Lancet Oncol. 2011;12: Recher C, Coiffier B, Haioun C, et al. Intensified chemotherapy with ACVBP plus rituximab versus standard CHOP plus rituximab for the treatment of diffuse large B- cell lymphoma (LNH03-2B): an open-label randomised phase 3 trial. Lancet. 2011;378: Dabaja BS, Advani R, Hodgson DC, et al. ACR Appropriateness Criteria(R) Diffuse Large B-Cell Lymphoma. American journal of clinical oncology. 2015;38: Moskowitz CH, Schoder H. Current Status of the Role of PET Imaging in Diffuse Large B-Cell Lymphoma. Semin Hematol. 2015;52: Coughlan M, Elstrom R. The use of FDG-PET in diffuse large B cell lymphoma (DLBCL): predicting outcome following first line therapy. Cancer Imaging. 2014;14: Juweid ME, Stroobants S, Hoekstra OS, et al. Use of positron emission tomography for response assessment of lymphoma: consensus of the Imaging Subcommittee of International Harmonization Project in Lymphoma. J Clin Oncol. 2007;25: Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014;32: Dabaja BS, Phan J, Mawlawi O, et al. Clinical implications of positron emission tomography-negative residual computed tomography masses after chemotherapy for diffuse large B-cell lymphoma. Leuk Lymphoma. 2013;54:

19 34. Shi Z, Esiashvili N, Flowers C, Das S, Khan MK. Renewed interest in the role of consolidative radiotherapy in advanced stage diffuse large B-cell lymphoma. Leuk Lymphoma. 2013;54: Dunleavy K, Pittaluga S, Maeda LS, et al. Dose-adjusted EPOCH-rituximab therapy in primary mediastinal B-cell lymphoma. N Engl J Med. 2013;368: Bloy N, Pol J, Manic G, et al. Trial Watch: Radioimmunotherapy for oncological indications. Oncoimmunology. 2014;3:e Golden EB, Apetoh L. Radiotherapy and immunogenic cell death. Seminars in radiation oncology. 2015;25: Bezu L, Gomes-de-Silva LC, Dewitte H, et al. Combinatorial strategies for the induction of immunogenic cell death. Frontiers in immunology. 2015;6: Pilones KA, Vanpouille-Box C, Demaria S. Combination of radiotherapy and immune checkpoint inhibitors. Seminars in radiation oncology. 2015;25: Supporting information Supporting Table S1 Table S1A Patient characteristics in 94 patients with stage II-III-IV disease and a residual mass, with or without consolidative radiotherapy Table S1B 5-year cancer-specific survival in 94 patients with stage II-III-IV disease and a residual mass, according to clinical variables Table S1C 5-year overall survival in 94 patients with stage II-III-IV disease and a residual mass, according to clinical variables 18

20 Table 1 Patient and treatment characteristics by response status after chemotherapy, with or without radiotherapy Complete response (CR), n=73 Any residual mass after chemotherapy a, n=138 Patient characteristics No RT n=48 With RT n=25 P-val. No RT n=57 With RT n=81 P-val. Age, mean (range) b ( ) ( ) ( ) ( ) 0.31 b Age 0.93 c 0.70 c < 70 years, n (%) 36 (75.0) 19 (76.0) 37 (64.9) 50 (61.7) 70 years, n (%) 12 (25.0) 6 (24.0) 20 (35.1) 31 (38.3) Sex 0.16 c c Female, n (%) 15 (31.3) 12 (48.0) 22 (38.6) 46 (56.8) Male, n (%) 33 (68.7) 13 (52.0) 35 (61.4) 35 (43.2) Stage (I-II vs III-IV) 0.06 d d Stage I-II, n (%) 18 (37.5) 12 (48.0) 13 (22.8) 31 (38.3) Stage III-IV, n (%) 17 (35.4) 3 (12.0) 39 (68.4) 32 (39.5) Stage PE I-II, n (%) 13 (27.1) 10 (40.0) 5 (8.8) 18 (22.2) IPI-score (0-3 vs 4-5) 0.30 c 0.13 c IPI-score 0-3, n (%) 46 (95.8) 25 (100.0) 47 (82.5) 72 (91.1) IPI-score 4-5, n (%) 2 (4.2) 0 (0.0) 10 (17.5) 7 (8.9) Bulky 0.56 c 0.82 c No bulk, n (%) 42 (87.5) 23 (92.0) 37 (64.9) 51 (63.0) With bulk, n (%) 6 (12.5) 2 (8.0) 20 (35.1) 30 (37.0) B-symptoms 0.44 c c No B-symptoms, n (%) 35 (76.1) 21 (84.0) 27 (47.4) 59 (73.7) With B-symptoms, n (%) 11 (23.9) 4 (16.0) 30 (52.6) 21 (26.3) LDH 0.51 c 0.72 c LDH normal, n (%) 35 (72.9) 20 (80.0) 15 (26.3) 23 (29.1) LDH > normal, n (%) 13 (27.1) 5 (20.0)) 42 (73.7) 56 (70.9) Survival No RT n=48 With RT n=25 P-val. e No RT n=57 With RT n=81 P-val. e 5-year CSS, % (SE) 98% (0.02) 92% (0.05) 69% (0.06) 89% (0.04) year CSS, % (SE) 98% (0.02) 92% (0.05) 62% (0.07) 89% (0.04) < year OS, % (SE) 85% (0.05) 92% (0.05) 59% (0.07) 82% (0.04) year OS, % (SE) 80% (0.06) 75% (0.13) 47% (0.07) 64% (0.07) Chemotherapy (stage II-III-IV) f Chemotherapy with rituximab, n (%) Chemotherapy courses (number), mean (range) No RT n=33 With RT n=6 P-val. No RT n= With RT n=44 P-val. 28 (84.8) 5 (83.3) 0.93 c 44 (88.0) 33 (75.0) 0.10 c 6.67 (5-8) 6.50 (3-8) 0.76 b 7.12 (1-10) 7.23 (4-10) 0.73 b Chemotherapy doses (stage II-III-IV) f Cyclophosphamide cumulative dose g, mean (SD) Doxorubicin cumulative dose g, mean (SD) No RT n= (938) With RT n=5 P-val (1350) No RT n= b 4741 (1399) With RT n=41 P-val (1070) 0.93 b 315 (64) 291 (88) 0.47 b 307 (91) 319 (68) 0.48 b

21 Abbreviations: RT = radiotherapy; LDH: Lactate Dehydrogenase; CSS = cancer-specific survival; OS = overall survival, SE = standard error; SD = standard deviation. a Out of 264 patients included in the study, 211 were included in analyses with consolidative RT versus without RT in primary treatment. 25 patients had primary progressive disease (PD), and 24 died during the chemotherapy period, these 49 patients were excluded (not at risk for consolidative RT). In addition, 4 patients of age > 85 years who did not want chemotherapy were excluded. The 138 patients with residual masses after chemotherapy included 22 patients with partial response (PR) and 116 patients with complete response unconfirmed (CRu), the latter also comprising 46 patients with uncertain residuals either biopsy negative (n=17), small residuals unchanged over time during chemotherapy (n=25), or PET-negative after chemotherapy (n=4). b P-value from independent t-tests. c P-value from Chi-Square statistics. d P-values from Chi-Square statistics of stage I-II vs III-IV. e P-value from Kaplan-Meier analysis with log-rank test. f Treatment characteristics for 133 patients with stage II-III-IV disease, including 94 with a residual mass after chemotherapy, and 39 with complete response after chemotherapy. g Data on chemotherapy doses were available for 82 out of 94 patients with stage II-III-IV DLBCL and a residual mass after chemotherapy. Out of 39 patients with stage II-III-IV and complete response after chemotherapy, chemotherapy doses were available for 33 patients. Cumulative doses for cyclophosphamide and for doxorubicin (mg/m 2 ) are shown. Missing data: for B-symptoms 3 patients, and for LDH and IPI-score 2 patients had missing data. For chemotherapy doses (cyclophosphamide and doxorubicin) 18 out of 133 stage II-III-IV patients (15%) had missing data.

22 Table 2 5-year cancer-specific survival for patients with DLBCL stage II-III-IV disease and a residual mass, with or without radiotherapy No radiotherapy N=50 With radiotherapy N=44 N Events 5y CSS % (SE) N Events 5y CSS % (SE) P-value b Residual mass after chemo (stages II-III-IV) a, n= % (0.07) % (0.05) Cohort , n= % (0.10) % (0.07) Cohort , n= % (0.10) % (0.07) Stage Stage II, n= % (0.15) % (0.08) 0.13 Stage III-IV, n= % (0.08) % (0.06) B-symptoms No B-symptoms, n= % (0.11) % (0.05) With B-symptoms, n= % (0.09) % (0.08) 0.13 Sex Female, n= % (0.12) % (0.07) Male, n= % (0.08) % (0.07) Age Age < 60 years, n= % (0.10) % (0.05) Age 60 years, n= % (0.09) % (0.08) Bulky No bulk, n= % (0.08) % (0.05) With bulk, n= % (0.12) % (0.08) IPI-score IPI-score 0-1, n= % (0.16) % IPI-score 2-3, n= % (0.09) % (0.07) IPI-score 4-5, n= % (0.13) % (0.13) 0.47 LDH c LDH < 300, n= % (0.11) % (0.06) LDH 300, n= % (0.09) % (0.07) Abbreviations: CSS = cancer-specific (i.e. lymphoma-specific) survival; SE = standard error; IPI = International Prognostic Index; LDH = Lactate Dehydrogenase. a The 94 stage II-III-IV patients with a residual mass after chemotherapy comprised 10 patients with partial response (PR), and 84 patients with CRu including 34 patients with uncertain residuals either biopsy negative, unchanged in size over time during chemotherapy, or PET-negative after chemotherapy. Patients with complete response after chemotherapy, primary extranodal (PE) disease, stage I disease, primary progressive disease (PD), or death during chemotherapy, were excluded. Out of 44 stage II-III-IV patients receiving consolidative RT, 40 (91%) had RT to a residual mass after chemotherapy/initial bulk (including 28 patients in stage III-IV, and 12 patients in stage II), and 4 patients (9%) had RT to extranodal sites. Eleven out of 12 patients in stage II receiving RT had six or more chemotherapy courses (i.e. RT was not given as part of initially planned treatment after limited 3-4 courses chemotherapy). b P-value from Kaplan-Meier analyses with log-rank test. c Lactate dehydrogenase (LDH) normal range U/L. One patient had missing data for B-symptoms, one for LDH, and one for IPI-score.