Abstract: Systemic mastocytosis (SM) is characterized by abnormal proliferation and accumulation of mast cells (MC) in various tissues and organs, predominantly skin, bone marrow (BM) and visceral organs. The extent of organ infiltration and subsequent organ damage is the basis for the classification of SM into indolent SM (ISM), smoldering SM (SSM), SM with associated clonal hematologic non-MC lineage disease (SM-AHNMD), aggressive SM (ASM) and MC leukemia (MCL). Depending on the subtype of SM, cell source (BM or peripheral blood) and assay sensitivity, an acquired mutation in the receptor tyrosine kinase KIT, usually KIT D816V, is detectable in 80-90% of patients. Next-Generation Deep Amplicon Sequencing (NGS) was performed to investigate 18 candidate genes at known mutational hotspot regions as previously described. Additional mutations in genes encoding for signaling molecules (JAK2, CBL, KRAS, NRAS), transcription factors (RUNX1), epigenetic regulators (ASXL1, DNMT3A, EZH2, TET2) or splicing factors (SRSF2, SF3B1, U2AF1) are detectable in the vast majority of patients KIT D816V+ advanced SM. In order to gain more insight into clonal evolution of myeloid progenitors in indolent and advanced SM with or without AHNMD, we explored the mutation profile of single-cell-derived CFU-GM colonies - by using Sanger sequencing - in 19 KIT D816V+ SM patients (investigated colonies, n=285; median per patient, n=15; range 10-30). The study included 7 patients with ISM/SSM/ASM (0 additional mutations), 4 patients with SM-AHNMD (median 1 additional mutation, range 1-4) and 8 patients with ASM-AHNMD (median 3 additional mutations, range 1-4). KIT D816V+ CFU-GM could be identified in all 8 patients with ASM ± AHNMD but in only 20% (1/5) of patients with SM-AHNMD, while all CFU-GM colonies derived from ISM patients were completely KIT D816V negative. On the other hand, CFU-GM colonies from individual ASM ± AHNMD patients were never entirely KIT D816V+ (median 60%, range 0-95). In contrast to KIT D816V, additional mutations were identified in CFU-GM colonies from all 12 multi-mutated (A)SM-AHNMD patients and many of these mutations were present in 100% of the CFU-GM-derived colonies analyzed. In 8 patients, different subclones with variable proportions of the number of mutated genes were identified and allowed to generate putative evolutionary trees. These mutations included TET2 and SRSF2 mutations in 6/6 and 4/4 patients with (A)SM-AHNMD. In contrast, ASXL1 mutations were not identified in all TET2/SRSF2 positive CFU-GM colonies suggesting that they are likely to occur later than TET2 and SRSF2. When additional mutations and KIT D816V were detected concomitantly in individual single-cell-derived CFU-GM colonies, the overall frequency of CFU-GM colonies positive for additional mutations was always higher than those that were KIT D816V positive, indicating that the KIT D816V mutation was a secondary event. In contrast to advanced SM, all ISM patients were negative for additional mutations. CFU-GM colonies and also microdissected CD15+ cells derived from BM biopsies were entirely KIT D816V negative in these patients, highlighting that the KIT D816V mutation may be restricted to other (probably later) stages of stem cell development and possibly only to the MC lineage. In general, the relative proportion of KIT D816V+ progenitors correlated well with established parameters for quantification of disease burden, e.g. BM MC infiltration, serum tryptase levels and KIT D816V allele burden in individual patients. In conclusion, the presence of multi-mutated myeloid non-MC lineage progenitors of the CFU-GM-type suggests an initial clonal expansion at an early stage of hematopoietic development due to mutations other than KIT D816V with a subsequent phenotype modification towards SM due to a later acquisition of KIT D816V. In contrast, ISM/SSM is not affected by mutations at the CFU-GM level which may at least in part explain its excellent prognosis. Disclosures Schnittger: MLL Munich Leukemia Laboratory: Other. Horny:Novartis: Consultancy, Honoraria. Haferlach:MLL: Equity Ownership. Valent:Novartis: Consultancy. Reiter:Novartis: Consultancy, Honoraria.

Abstract: B cell chronic lymphocytic leukemia (B-CLL) is characterised by an accumulation of malignant B cells, and impaired humoral and cellular immune responses. Evasion strategies of leukemic cells appear to involve down-regulation of co-stimulatory molecules as well as increased resistance to apoptosis. Here we provide data supporting a novel concept to treat B-CLL with a humanized, superagonistic monoclonal antibody specific for CD28 (TGN1412). Superagonistic anti-CD28 antibodies have been shown to activate human T cells in vitro without requirement for engagement of the T cell antigen receptor (Luhder et al., J. Exp. Med. 2003. 197(8):955–66). Indicative of their activation, TGN1412-triggered T cells from healthy donors upregulate, among other activation markers, CD40L, that has been reported to promote anti-leukemic effects when ectopically expressed on B-CLL cells (Wierda et al., Blood. 2000. 96 (9): 2917–2924). In this report, the responses of PBMCs from B-CLL patients to soluble TGN1412 were examined. We show that in a dose-dependent fashion, polyclonal T cell activation was induced by TGN1412 including proliferation, cytokine production and induction of activation markers such as CD25, CD71, CD134 (Ox40), CTLA-4 (CD152) and CD154 (CD40L). Significantly, modulation of malignant B-CLL cells was also observed. MHC class II molecules (HLA-DR), CD95 and the co-stimulatory molecules CD80 and CD86, but not the proliferation marker Ki-67, were strongly up-regulated upon TGN1412 stimulation. These data suggested that improved antigen-presenting functions of B-CLL cells were induced by TGN1412. Accordingly, preliminary data indicate that B-CLL cells isolated from TGN1412 stimulated cultures induced enhanced proliferation of both allogeneic and autologous T cells, and importantly, TGN1412 activated T cells exhibited enhanced CTL-activity against B-CLL cells. In conclusion, our data suggest that TGN1412 induces polyclonal T cell expansion and activation as well as increased APC function of B-CLL cells. They imply that TGN1412 may have future therapeutic benefit for B-CLL patients.

Abstract: Signaling through the receptor tyrosine kinase kit controls proliferation and differentiation of hematopoietic precursor cells and mast cells. Somatic point mutations of the receptor that constitutively activate kit signaling are associated with mastocytosis and various hematopoietic malignancies. We generated a Cre/loxP-based bacterial artificial chromosome transgenic mouse model that allows conditional expression of a kit gene carrying the kitD814V mutation (the murine homolog of the most common mutation in human mastocytosis, kitD816V) driven by the kit promoter. Expression of the mutant kit in cells of adult mice, including hematopoietic precursors, caused severe mastocytosis with 100% penetrance at young age frequently associated with additional hematopoietic (mostly B lineage–derived) neoplasms and focal colitis. Restriction of transgene expression to mature mast cells resulted in a similar mast cell disease developing with slower kinetics. Embryonic expression led to a hyperproliferative dysregulation of the erythroid lineage with a high rate of perinatal lethality. In addition, most adult animals developed colitis associated with mucosal mast cell accumulation. Our findings demonstrate that the effects of constitutive kit signaling critically depend on the developmental stage and the state of differentiation of the cell hit by the gain-of-function mutation.

Abstract: Abstract 2859 Point mutations in the kinase domain of the receptor tyrosine kinase KIT at position 816 (D816V 〉 95%; D816H, D816Y, others 〈 5%) play a central role in the pathogenesis, diagnosis and targeted treatment of systemic mastocytosis (SM). However, accurate numbers on the overall frequency of KIT D816V in the diverse SM subtypes remain largely unknown because of i) the rarity of disease, ii) highly variable mast cell burden in bone marrow (BM) and peripheral blood (PB) and iii) inadequate sensitivity of disparate assays. We therefore developed an allele-specific quantitative RT-PCR (RQ-PCR) with an enhanced sensitivity of 0.01–0.1% which was superior to D-HPLC (mutation detection rate: 0.5%-1%) or conventional sequencing (mutation detection rate: 10–20%). Classification of the patients into disease subtypes was performed according to the WHO 2008 classification. Irrespective of subtype, KIT D816 mutations were identified in 143/144 (99.3%) of patients (D816V, n=141, 98.6%; D816H, n=1; D816Y, n=1) with SM (indolent SM [ISM], n=62, 43.1%; smoldering SM [SSM] , n=7, 4.9%; SM with associated hematological non-mast cell disease [SM-AHNMD], n=15, 10.4%, and aggressive SM/mast cell leukemia [ASM/MCL] , n=59, 41%). As control groups, seven patients with cutaneous mastocytosis, 20 healthy individuals and 80 patients with chronic myeloid leukemia in molecular remission were negative for the KIT D816V mutation in BM and PB. All patients with SSM (n=7), SM-AHNMD (n=14) and ASM/MCL (n=56) were positive in BM and also in PB. In contrast, 25/51 (49%) ISM patients were positive in BM but negative in PB. The KIT D816V RQ-PCR was subsequently used for quantitative assessment of the KIT D816V allele burden, expressed as the ratio KIT D816V/total KIT, in a total of 207 samples from PB (n=126) and BM (n=81) of 141 KIT D816V positive patients. In BM, the median allele burden was 8.9% (range 1.1–48.3%) in ISM, 29.7% (range 13.6–39.4%, p=0.007) in SSM, 23.3% (range 1.8–45.8%, p=0.09) in SM-AHNMD and 30.7% (range 1–99%, p 〈 0.0001) in ASM/MCL. In PB, the allele burden clearly allowed differentiation of ISM (median 0, range 0–15.2%) vs. SSM (median 8.8%, range 0.3–42.5%, p=0.001), ISM vs. SM-AHNMD (median 11.5%, range 0.2–64.9%, p 〈 0.001) and ISM vs. ASM/MCL (median 33%, range 0.53–74.0%; p 〈 0.001). The vast majority of ISM patients (47/51; 92%) had an allele burden in PB 〈 10%. ISM patients with an allele burden ≥10% collectively showed overlapping symptoms of diarrhea with (n=2) or without (n=1) histologically confirmed bowel infiltration, mild cytopenia (n=1), splenomegaly (n=1) and/or abdominal lymphadenopathy (n=1) potentially indicating more advanced disease. The wide heterogeneity of the allele burden in SSM and SM-AHNMD patients reflects the clinical heterogeneity in these subtypes. Of interest, the correlation of clinical findings and allele burden also revealed that patients with ASM-CMML had a significantly higher allele burden than ASM patients without monocytosis (45.5% vs. 30.9%, p=0.0029). In order to evaluate the use of RQ-PCR for monitoring of residual disease, serial measurements of the allele burden in PB were performed in patients after chemotherapy with cladribine (n=1) or allogeneic stem cell transplantation (SCT, n=2). The patient on cladribine revealed a decrease from 42% to 5% within 9 months which was associated with a decrease of serum tryptase levels from 610 μg/l to 129 μg/l and a decrease of mast cells in BM from 70–80% to 10%. When the patient relapsed 18 months after completion of chemotherapy with progressive organomegaly and thrombocytopenia, the allele burden had also risen to 35%. In the two allografted patients, allele burden was 35% and 45%, respectively, prior to SCT. Both patients are currently in complete remission without detectable KIT D816V mutation 17 and 12 months after allogeneic SCT. We therefore conclude that the measurement of the KIT D816V allele burden in SM patients from cDNA is a useful complementary tool for diagnosis and subclassification of SM. In the new treatment era of successful use of tyrosine kinase inhibitors and allogeneic SCT, it may also become very useful for monitoring of response to treatment in addition to established parameters like C-findings, serum tryptase and BM histology. Disclosures: No relevant conflicts of interest to declare.

Abstract: Abstract 4100 In some patients with suspected hypereosinophilic syndrome (HES), clonality of eosinophils may be proven by identification of an acquired chromosome or molecular abnormality leading to the diagnosis of chronic eosinophilic leukemia (CEL). The most common molecular aberrations are fusion genes with involvement of PDGFRA, e.g. FIP1L1-PDGFRA (FP), or PDGFRB, e.g. ETV6-PDGFRB. Molecular testing for FP by RT-PCR or FISH is nowadays performed early in the diagnostic work-up of suspected non-reactive eosinophilia. However, eosinophilia is also present at variable frequency in patients with systemic mastocytosis (SM) and other subtypes of myeloproliferative neoplasms (MPN). Recurrent molecular markers for those entities are KIT D816V (80-90% positivity in SM) and JAK2 V617F (60-70% positivity in MPN). We therefore sought to evaluate the relative frequency of FP (by RT-PCR), KIT D816V (by D-HPLC plus direct sequencing) and JAK2 V617F (by ARMS-PCR) in 300 samples from patients with suspected HES/CEL and to correlate molecular findings with clinical features. Molecular abnormalities were identified in 42 (14%) cases; 22 (7.3%) were positive for FP, 14 (4.6%) for KIT D816V and 6 (2.0%) for JAK2 V617F, respectively. Most baseline clinical characteristics, e.g. leukocytes, absolute and relative number of eosinophils, hemoglobin, platelets or splenomegaly, were not different between the three entities. Significant differences were found regarding age, gender, serum tryptase levels and course of disease. FP positive patients were significantly younger (p 〈 0.001) and exclusively male while a female prepoderance was observed for KIT and JAK2 mutated patients. Significantly elevated serum tryptase levels (normal value 〈 11.4μ g/l) were found in all cases of FP and KIT D816V positive patients. However, serum tryptase levels 〉 50μ g/l were almost exclusively seen in KIT D816V positive SM patients. Aggressive SM (ASM) was diagnosed in 6 of 14 (43%) KIT D816V positive patients due to characteristic bone marrow morphology and the presence of diverse C-findings (e.g. anemia 〈 10g/dl, n=2, and/or thrombocytopenia 〈 100×109/μ l, n=7). Frequent additional clinical features included lymphadenopathy (n=9) and urticaria pigmentosa (n=6). Two ASM patients died within first year of diagnosis while 21 (95%) FP positive CEL patients are in complete molecular remission on imatinib after a median treatment time of 28 months (range 8–149). We conclude that the serum tryptase level is an important diagnostic and prognostic marker in eosinophilia. We suggest that FP negative HES patients should be screened for KIT D816V and JAK2 V617F point mutations, both of which are potentially targetable by small molecule inhibitors. FIP1L1-PDGFRA KIT D816V JAK2 V617F Number of patients 22 (7.3%) 14 (4.6%) 6 (2.0%) Age (median, years) 4418–73 6342–81 7269–87 Gender (m/f) 22/0 6/8 4/6 Leukocytes (median, range, ×109/l) 13.57.2–85.6 10.06.8–124.0 26.812.7–61.1 Eosinophil (median, range, ×109/l) 6.51.4–34.8 2.21.2–100.0 10.03.2–16.5 Eosinophils (%) 469–74 246–81 378–61 Hemoglobin (median, range, g/dl) 12.67.1–15.8 11.77.9–15.7 13.511.4–16.2 Platelets (median, range, ×109/l) 16034–375 11215–945 17434–364 Splenomegaly 14/14 (100%) 13/13 (100%) 4/4 (100%) Serum tryptase 〉 50μ g/l 1/13 (8%) 8/9 (88%) not done 〉 100μ g/l 0/13 (-) 7/7 (100%) not done Disclosures: Erben: Novartis: Honoraria, Research Funding.

Abstract: Abstract 717 Purpose: Combined modality treatment consisting of 4 cycles of chemotherapy (CT) followed by involved field radiotherapy (IF-RT) is the standard treatment for early unfavourable HL. In our prior trial for this group of patients (HD8), overall survival (OS) and freedom from treatment failure (FFTF) at 5 years were 91% and 83%, respectively. The HD11 trial thus addressed two major questions: (1) improving outcome by intensifying CT (4xABVD vs. 4xBEACOPPbaseline; Bbas) and (2) defining the best radiation dose (30Gy vs. 20Gy IF-RT). Patients and methods: Between May 1998 and January 2003, 1395 eligible patients aged 16–75 years with untreated early unfavourable stage HL (CS I, IIA with at least one of the risk factors large mediastinal mass (a), extranodal disease (b), elevated ESR (c) or ≥ 3 nodal areas (d); IIB with risk factors c and/or d) were randomized into one of the following 4 treatment arms: 4xABVD + 30Gy (A), 4xABVD + 20Gy (B), 4x Bbas + 30Gy (C) or 4x Bbas + 20Gy (D). Since there are strong indications for an interaction between CT- and RT-doses, a comparison of pooled treatment arms (A+B vs. C+D for comparison of 4×ABVD vs. 4× Bbas and A+C vs. B+D for comparison of 30Gy IF-RT vs. 20Gy IF-RT) would be misleading. Therefore all treatment arms were analysed separately. Results: Patient characteristics were well balanced between the 4 arms (median age 33 years, 49% male, 6% stage I, 29% B-symptoms). CT- and RT-related acute toxicity occurred significantly more often in the arms with the more intensive therapy (CT: 74.1% vs. 51.8%; RT: 12.3% vs. 5.5%). The complete remission rate 3 months after end of therapy was 94.1% for the whole group and did not differ significantly between the 4 arms. The 5-year estimate of FFTF (primary endpoint) is 85.0% (OS 94.5%, PFS 86.0%). Bbas is more effective than ABVD if followed by 20Gy IF-RT (5y-FFTF difference 5.7%, 95%-CI [0.1%; 11.3%]). This effect does not exist in combination with 30Gy IF-RT (5y-FFTF difference 1.6% [-3.6%; 6.9%] ). Similar results are observed for the RT-question: After 4 cycles of Bbas, 20Gy is not inferior to 30Gy (5y-FFTF difference -0.1%, 95%-CI [-5.1%; 4.9%]), whereas after 4xABVD, a relevant inferiority of 20Gy cannot be excluded (-4.0% [-9.5%; 1.4%] ). Conclusion: A reduction of RT dose from 30Gy to 20Gy IF-RT seems to be justified only in combination with Bbas, but not with a less effective chemotherapy such as 4xABVD. Patients will benefit from an intensified CT such as Bbas only in combination with 20Gy IF-RT but not with 30Gy IF-RT. Disclosures: No relevant conflicts of interest to declare.

Abstract: Abstract 716 Background: There has been an ongoing debate on the best treatment for patients with early favourable Hodgkin lymphoma (HL). Open questions include the choice between combined modality treatment or chemotherapy only, the number of chemotherapy cycles needed and the optimal radiation dose. The GHSG thus conducted a randomized study for patients with early-stage favourable Hodgkin lymphoma (HD10) in which these questions were addressed. Methods: HD10 was an international prospectively randomized multicenter trial comparing 2 and 4 cycles of ABVD as well as 20Gy or 30Gy involved field radiotherapy (IFRT) in a 2 × 2 statistical design. Between 5/1998 and 1/2003, a total of 1370 patients from 329 centers were randomized into four arms: 4 × ABVD + 30Gy; 4 × ABVD + 20Gy; 2 × ABVD + 30Gy; 2 × ABVD + 20Gy. All patients had their initial histology reviewed by a lymphoma expert panel. Documentation was complete in more than 99,1% of cases for this final analysis. Results: Patients were equally balanced for age, gender, stage, histology, performance status and risk factors between arms. There were significant differences in major toxicity (WHO grade III/IV) between 4 × ABVD and 2 × ABVD in the overall number of events (52% vs 33%) including leukopenia (24% vs 15%) and hair loss (28% vs 15%). In terms of radiation dose, there also was a difference in toxicity between 30Gy and 20Gy IFRT (all events: 8.7% vs 2.9%), dysphagia (3% vs 2%), mucositis (3.4% vs 0.7%). Complete remission was achieved in 97% of patients treated with 4 × ABVD, 97% with 2 × ABVD, 99% after 30Gy and 97% after 20Gy. With a median follow-up of 79–91 months, there was no significant difference between 4 × ABVD and 2 × ABVD in terms of overall survival at 5 years (OS: 4 × ABVD 97.1%; 2 × ABVD: 96.6%), freedom from treatment failure (FFTF: 93.0% vs 91.1%) and progression free survival (PFS: 93.5% vs 91.2%). For the radiotherapy question, there were also no significant differences between patients receiving 30Gy IFRT and those with 20Gy IFRT in terms of OS (97.6% vs 97.5%), FFTF (93.4% vs 92.9%) and PFS (93.7% vs 93.2%), respectively. Importantly, there was also no significant difference in terms of OS, FFTF and PFS when all four arms were compared. Conclusion: Two cycles of ABVD followed by 20Gy IFRT is the new GHSG standard of care for Hodgkin patients in early favourable stages. Disclosures: No relevant conflicts of interest to declare.

Abstract: Introduction Salvage high dose chemotherapy (HDCT) followed by autologous stem cell transplantation (ASCT) is used in fit patients with relapsed multiple myeloma (RMM) in clinical practice. However, the role of this approach in the era of continuous novel agent based treatment has not been defined in randomized trials. The ReLApsE trial compared lenalidomide/dexamethasone (Rd) re-induction, salvage HDCT/ASCT and lenalidomide (R) maintenance with standard continuous Rd in a randomized controlled multicenter trial. Methods Between 2010 and 2016, 282 patients were randomized of whom 277 constituted the intention-to-treat (ITT) population (arm B/A n=139/138). Arm B received 3 cycles of Rd (lenalidomide 25 mg, day 1-21; dexamethasone 40 mg, day 1, 8, 15, 22; 4 week cycles) re-induction, HDCT (melphalan 200 mg/m2), ASCT and R maintenance (10 mg daily) until progression (PD). Arm A was treated with Rd until PD. In both arms stem cells were harvested after the 3rd Rd cycle if no back-up transplant was available. Key inclusion criteria were 1-3 prior therapy lines, age ≤ 75 years, time to PD ≥ 12 months in case of front-line HDCT/ASCT and WHO PS ≤ 2. The primary endpoint was progression free survival (PFS). Secondary endpoints included overall survival (OS), response rates and toxicity. ISRCTN16345835, Eudra CT-No: 2009-013856-61. Results Arm B and A were balanced regarding age (median 61.3 vs. 62.2 years), ISS (I/II/III in 62.6/24.4/13% vs. 59.7/31/9.3%) and WHO PS (0/1/2 in 69.1/30.9/0% vs. 76.1/23.2/0.7%). Almost all patients had only 1 prior therapy line (arm B: 94.2% vs. arm A: 93.5%) and had received front-line HDCT/ASCT (92.8% vs. 94.2%). More patients in arm B had high risk cytogenetic aberrations (HR-CA; 42.9% vs. 31.6%) based on a higher frequency of t(4;14) (20.2% vs. 10.1%). The overall response rate (≥ partial response; ORR) for arm B and A was 77.9% and 74.6% (p=0.57) with 49.3% and 47.1% (p=0.81) achieving ≥ very good partial response as best response. Within a median follow up of 36.3 months, 183 PFS events and 76 deaths occurred. Median PFS in the ITT population was 20.7 months in arm B and 18.8 months in arm A without a statistically significant difference (HR 0.87; 95% CI 0.65-1.16; p=0.34). Median OS was not reached (NR) in arm B vs. 62.7 months in arm A (HR 0.81; 95% CI 0.52-1.28; p=0.37). In arm B, 41 patients (29.5%) did not receive the planned HDCT/ASCT. Thus, exploratory landmark (LM) analyses from HDCT and the contemporaneous Rd cycle 5 in arm A were performed (median interval from randomization to HDCT/Rd cycle 5: 117/122 days; n=103[B]/114[A] ). They showed a trend towards superior PFS (23.3 vs. 20.1 months; HR 0.74; p=0.09) and significantly superior OS (NR vs. 57 months; HR 0.56; p=0.046) in arm B vs. A. Multivariate analyses revealed significant associations of treatment in arm B with superior LM PFS (HR 0.6; p=0.01) and LM OS (HR 0.39; p=0.006). Other factors in the LM multivariate models showing significant associations with survival were HR-CA (PFS, OS), number of prior therapy lines (PFS), and age (PFS). The ORR in arm B after HDCT/ASCT was significantly higher than in arm A after Rd cycle 5 (82.3% vs. 69.6%; p=0.04). Grade ≥3 adverse events were reported in 83% (arm B) and 74.5% (arm A; p=0.11). Grade ≥3 leukopenia/neutropenia was reported in 61.5 vs. 24.8% (p 〈 0.001), grade ≥3 thrombopenia in 45.2 vs. 11% (p 〈 0.001) and grade ≥3 mucositis in 10.4% vs. 2.1% (p=0.005). No significant difference in grade ≥3 infections/infestations (33.3 vs. 27.6%; p=0.3) was observed. Eleven patients died on protocol treatment (B: 4 vs. A: 7). No deaths occurred in the HDCT/ASCT phase. Conclusions This is the first RCT comparing salvage HDCT/ASCT with continuous novel agent based treatment. No significant PFS or OS difference was observed in the overall trial population. However, HR-CA were more frequent in the HDCT/ASCT arm and ~30% of patients did not receive the planned HDCT/ASCT. Landmark analyses from the time of HDCT indicate superior PFS and OS in patients actually undergoing salvage HDCT/ASCT. Salvage HDCT/ASCT was safe with an expected increase in hematological as wells as gastrointestinal toxicity but without treatment-related mortality in patients up to the age of 75 years in this multicenter trial. However, the number of patients not undergoing salvage HDCT/ASCT and the approval of more active Rd-based triplet regimens after the initiation of this trial prevents definite conclusions on the role of salvage HDCT/ASCT. Disclosures Goldschmidt: Amgen: Consultancy, Research Funding; Novartis: Honoraria, Research Funding; ArtTempi: Honoraria; Janssen: Consultancy, Honoraria, Research Funding; Sanofi: Consultancy, Research Funding; Mundipharma: Research Funding; Takeda: Consultancy, Research Funding; Celgene: Consultancy, Honoraria, Research Funding; Bristol Myers Squibb: Consultancy, Honoraria, Research Funding; Adaptive Biotechnology: Consultancy; Chugai: Honoraria, Research Funding. Baertsch:Takeda: Consultancy; Novartis: Consultancy, Research Funding. Raab:Celgene: Consultancy, Honoraria; Novartis: Consultancy, Honoraria, Research Funding; BMS: Consultancy, Honoraria, Research Funding; Amgen: Consultancy, Honoraria, Research Funding. Hillengass:Novartis: Honoraria, Other: Advisory Board; Sanofi: Research Funding; Takeda: Honoraria, Other: Advisory Board; Janssen: Honoraria, Other: Advisory Board; amgen: Consultancy, Honoraria, Other: Advisory Board; BMS: Honoraria, Other: Advisory Board; Celgene: Consultancy, Honoraria, Other: Advisory Board, Research Funding. Graeven:AbbVie: Honoraria; Roche: Membership on an entity's Board of Directors or advisory committees. Fenk:Takeda: Honoraria; Bristol-Meyers Squibb: Honoraria, Other: travel grant; Celgene: Honoraria, Other: Travel grant, Research Funding; Janssen: Honoraria; Amgen: Honoraria. Haenel:Novartis: Honoraria; Roche: Honoraria; Amgen: Honoraria; Takeda: Honoraria. Scheid:Novartis: Honoraria, Research Funding; Takeda: Honoraria, Research Funding; Janssen: Honoraria; Celgene: Honoraria; BMS: Honoraria; Amgen: Honoraria. Salwender:Celgene: Honoraria, Other: travel suppport, Research Funding; Janssen: Honoraria, Other: travel support, Research Funding; Novartis: Honoraria, Other: travel suppport, Research Funding; Takeda: Honoraria; Amgen: Honoraria, Other: travel suppport, Research Funding; Bristol-Myers Squibb: Honoraria, Other: travel suppport, Research Funding. Weisel:Amgen, Celgene, Janssen, and Sanofi: Research Funding; Amgen, BMS, Celgene, Janssen, and Takeda: Honoraria; Amgen, BMS, Celgene, Janssen, Juno, Sanofi, and Takeda: Consultancy, Membership on an entity's Board of Directors or advisory committees.

Abstract: Introduction The ReLApsE trial compared lenalidomide (LEN)/dexamethasone (DEX; Rd) re-induction, salvage high dose chemotherapy (HDCT), autologous stem cell transplantation (ASCT) and LEN maintenance with continuous Rd in relapsed multiple myeloma. Landmark (LM) analyses from salvage HDCT were performed due to the fact that ~30% of patients in the HDCT arm did not receive salvage HDCT/ASCT. These analyses showed a survival benefit in patients actually undergoing salvage HDCT/ASCT. Median PFS and OS from LM were 23.3 vs. 20.1 months (HR 0.74; p=0.09) and not reached vs. 57 months (HR 0.56; p=0.046) favoring the salvage HDCT/ASCT arm. Multivariate LM analyses showed significant associations of the salvage HDCT/ASCT arm with superior PFS (HR 0.6; p=0.01) and OS (HR 0.39; p=0.006). The present analysis aims to dissect treatment efficacy in relevant subgroups and provide clues for treatment stratification. Methods The ReLApsE trial (ISRCTN16345835) compared 3 Rd (LEN 25 mg, d1-21; DEX 40 mg, d1,8,15,22; 4 week cycles) re-induction cycles, HDCT (melphalan 200 mg/m2), ASCT and LEN maintenance (10 mg/d) until PD (arm B, n=139) with Rd until PD (arm A, n=138). Key inclusion criteria were 1-3 prior therapy lines, age ≤ 75, time to PD in case of front-line HDCT/ASCT (TTP1) ≥ 12 months and WHO PS ≤ 2. Exploratory subgroup analyses were performed in the ITT population for PFS/OS using an LM at HDCT (B; n=103) and the contemporaneous Rd cycle 5 (A; n=114). The median interval from randomization to LM was 117/122 days in arm B/A. Heterogeneity of treatment effect was assessed by cox regression with interaction term between treatment and subgroup factor. Results No significant differences in the PFS/OS benefit between arms were observed in subgroups according to baseline ISS (I/II/III; interaction p[i-p]=0.5/0.66), age ( 〈 /≥65 yrs; i-p=0.13/0.89), renal function (MDRD 〈 /≥ 60 ml/min; i-p=0.68/0.34), response to re-induction ( 〈 /≥ PR; i-p=0.92/0.48), prior therapy lines (1/ 〉 1; i-p=0.37/0.22), single vs. tandem front-line HDCT/ASCT (i-p=0.34/0.56), and TTP1 (12-24 vs. 24-48 vs. 〉 48 months; i-p=0.91/0.21). The subgroups according to front-line HDCT/ASCT (yes/no) differed significantly with regard to PFS/OS benefit in arm B (i-p=0.006/0.001). A significant benefit was observed in patients with front-line HDCT/ASCT treated in arm B regarding PFS (HR 0.68, p=0.03; n=107[A]/98[B] ) and OS (HR 0.43, p=0.009). Patients without front-line HDCT/ASCT constituted a very small subgroup with imbalances in baseline parameters adversely affecting arm B and had expectably inferior survival in arm B (PFS: HR 4.35, p=0.08; OS: HR 19.83, p=0.0078; n=7[A]/5[B] ). The subgroup with baseline LDH 〈 / 〉 upper limit of normal (ULN) differed significantly in PFS benefit in arm B (i-p=0.03) but not in OS benefit (i-p=0.46). Patients with LDH 〈 ULN had significantly better PFS (HR 0.61, p=0.01; n=98[A]/85[B] ) in arm B whereas no significant difference between trial arms was observed in patients with LDH 〉 ULN (PFS: HR 1.54, p=0.31; n=16[A]/18[B] ). The subgroups according to baseline cytogenetic risk and R-ISS showed a trend towards a differential benefit in arm B regarding OS (i-p=0.05 and 0.09) but not PFS (i-p=0.5 and 0.88). Patients with low risk cytogenetics (i.e. absence of t(4;14), del17p, +1q 〉 3 copies and t(14;16)) had significantly superior OS in arm B (HR 0.21; p=0.01; n=57[A]/35[B] ), whereas patients with high risk cytogenetics had no significant difference in OS according to trial arm (HR 0.82, p=0.67; n=25[A]/28[B] ). Patients with R-ISS I had significantly superior OS in arm B (HR 0.08; p=0.02; n=33[A]/25[B] ), whereas no significant difference in OS according to trial arm was seen in patients with R-ISS II (HR 0.72, p=0.42; n=52[A]/43[B] ) and R-ISS III (HR 0.65, p=0.6; n=3[A]/5[B] ). Conclusions The ReLApsE trial is the first RCT of salvage HDCT/ASCT vs. continuous novel agent treatment. In the absence of a significant survival benefit for the primary endpoint, LM analyses indicated a significant PFS/OS benefit in patients actually undergoing HDCT/ASCT. No heterogeneity of treatment effect was observed according to ISS, age, renal function, response to re-induction, prior therapy lines, single vs. tandem front-line HDCT/ASCT, and TTP1. Subgroup effects regarding PFS and/or OS benefit from HDCT/ASCT were seen favoring patients with front-line HDCT/ASCT and patients with low risk according to LDH, cytogenetics and R-ISS. Disclosures Baertsch: Takeda: Consultancy; Novartis: Consultancy, Research Funding. Raab:Celgene: Consultancy, Honoraria; BMS: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding; Amgen: Consultancy, Honoraria, Research Funding. Hillengass:Celgene: Consultancy, Honoraria, Other: Advisory Board, Research Funding; amgen: Consultancy, Honoraria, Other: Advisory Board; Novartis: Honoraria, Other: Advisory Board; Janssen: Honoraria, Other: Advisory Board; Takeda: Honoraria, Other: Advisory Board; BMS: Honoraria, Other: Advisory Board; Sanofi: Research Funding. Graeven:Roche: Membership on an entity's Board of Directors or advisory committees; AbbVie: Honoraria. Fenk:Bristol-Meyers Squibb: Honoraria, Other: travel grant; Takeda: Honoraria; Amgen: Honoraria; Janssen: Honoraria; Celgene: Honoraria, Other: Travel grant, Research Funding. Haenel:Takeda: Honoraria; Novartis: Honoraria; Amgen: Honoraria; Roche: Honoraria. Scheid:Amgen: Honoraria; BMS: Honoraria; Celgene: Honoraria; Janssen: Honoraria; Novartis: Honoraria, Research Funding; Takeda: Honoraria, Research Funding. Salwender:Janssen: Honoraria, Other: travel support, Research Funding; Celgene: Honoraria, Other: travel suppport, Research Funding; Novartis: Honoraria, Other: travel suppport, Research Funding; Bristol-Myers Squibb: Honoraria, Other: travel suppport, Research Funding; Amgen: Honoraria, Other: travel suppport, Research Funding; Takeda: Honoraria. Weisel:Amgen, Celgene, Janssen, and Sanofi: Research Funding; Amgen, BMS, Celgene, Janssen, Juno, Sanofi, and Takeda: Consultancy, Membership on an entity's Board of Directors or advisory committees; Amgen, BMS, Celgene, Janssen, and Takeda: Honoraria. Goldschmidt:Celgene: Consultancy, Honoraria, Research Funding; Takeda: Consultancy, Research Funding; Novartis: Honoraria, Research Funding; Mundipharma: Research Funding; Janssen: Consultancy, Honoraria, Research Funding; Adaptive Biotechnology: Consultancy; ArtTempi: Honoraria; Bristol Myers Squibb: Consultancy, Honoraria, Research Funding; Chugai: Honoraria, Research Funding; Sanofi: Consultancy, Research Funding; Amgen: Consultancy, Research Funding.

Abstract: Purpose: The GHSG HD9 trial had established BEACOPP escalated (BE) as new standard of care for advanced-stage HL patients by showing significant superiority in terms of failure-free survival (FFTF) and overall survival (OS) over COPP/ABVD and BEACOPP baseline (BB) (each 8 cycles). The successor study, HD12, evaluated a possible reduction in toxicity by comparing 8 cycles of BE with 4 cycles BE followed by 4 cycles BB. The second question in this trial related to the need of additional radiotherapy (RT) to initial bulk and residual disease. Patients and methods: HL patients in stage IIB with large mediastinal mass and/or E-lesions or stage III/IV were randomised according to a 2×2-factorial design between: 8BE + RT, 8BE no RT, 4BE+4BB + RT, 4BE+4BB no RT. Reviewing CT-images before and after chemotherapy treatment, fields for RT were centrally planned by a multidisciplinary diagnostic panel blinded for the randomisation arm. Primary endpoint of the trial was FFTF. Between 9/1999 and 1/2003, a total of 1,670 patients aged 16–65 were randomized. For this final analysis at a median follow up of 78 months, 99 patients were excluded (42 HL not confirmed, 20 revision of stage, 20 no study treatment or documentation, 17 others) resulting in 1,571 eligible patients. Results: Patient characteristics in the 4 groups were comparable with 49% of patients in stage III, 35% in stage IV, 68% reporting B-symptoms and 28% having a large mediastinal tumor. An IPS of 3 or greater was reported for 38% of patients, predominant histology was nodular sclerosis with 57% of cases. Treatment-related toxicity of WHO grade III/IV was observed in 97% of patients. Most prominent differences between pooled chemotherapy arms were anemia (65% 8BE vs 51% 4BE+4BB) and thrombopenia (65% vs 51%). Treatment outcome: complete remission 92.4%; early progression 2.2%; progression/relapse 7.8% (6.6% and 8.5%). A total of 156 (9.9%) deaths (72 vs 84) have been observed (22 vs 32 acute or salvage treatment toxicity; 15 vs 24 HL; 22 vs 13 secondary neoplasia). Most treatment related deaths occurred in the & gt;60 years age group, the first 4 cycles and the IPS & gt; 3 RF groups. Secondary neoplasias were observed in 77 patients (4.9%): AML/MDS 1.5% vs 1.4%, NHL 1.4% vs 0.6% and solid tumors/others 2.5% vs 2.3%. At 5 years, OS was 91%, FFTF 85.5% and progression free survival (PFS) 86.2% (Kaplan- Meier estimates). Estimates for the difference at 5 years are 1.8% for OS, 2.3% for FFTF and 2.7% for PFS favoring BE. However, there was no statistical difference between 8x BE and 4BE+4BB in all outcome parameters (p & gt;0.19, log rank test). There is also no significant difference between the RT or no-RT arms in this study with the caveat that a number of high-risk patients receiving RT based on the blinded panel decision. Conclusion: The adoption of 4BE+4BB as a new standard in the future GHSG studies will depend on a refined analysis of the total data set and will be presented.