ALBERT

Description: Elevated plasma levels of the nucleoside diphosphate kinase (NDPK) NM23-H1 are associated with poorer prognosis in acute myeloid leukemia (AML). We previously demonstrated that leukemic blasts release NM23-H1, which binds to more differentiated myeloid cells inducing their secretion of inflammatory cytokines, including IL-1β, that promote survival and proliferation of leukemic blasts1. Both AML and myelodysplastic syndrome (MDS) patients are prone to infections due to impaired hematopoiesis that is worsened by treatment. NDPKs are highly evolutionarily conserved raising the possibility that bacterial/fungal NDPKs could mediate the same survival effect on malignant AML/MDS blasts and exacerbate disease progression. To test this, we generated recombinant NDPKs (rNDPKs) from bacteria and fungi associated with common infections in these patients (E. coli, S. aureus, S. pneumoniae, K. pneumoniae, C. albicans). Cytokine production and survival responses of primary AMLs to these proteins were indistinguishable from their response to rNM23-H1. This activity was independent of NDPK enzyme activity since mutant rNM23-H1 and bacterial and fungal rNDPKs with impaired oligomerization, kinase or exonuclease activity elicited the same cytokine and survival response. Toll like receptors (TLRs) are the major family of human DAMP/PAMP receptors and IL-1β secretion is closely associated with TLR-4 mediated activation of the NLRP3 inflammasome in monocytes. We therefore postulated that NM23-H1 and pathogen derived NDPKs act as novel damage- and pathogen- associated molecular pattern (DAMP, PAMP) molecules. We confirmed that fluorescently labelled rNM23-H1 and S. pneumoniae rNDPK bound selectively to monocytes in peripheral blood. Using in vitro generated monocytes (vitamin D3 differentiated THP-1 cells) we demonstrated that both wild type and mutant rNM23-H1 and bacterial/fungal rNDPKs induced activation of caspase-1 and cleavage of pro-IL-1β into its active form. Secretion of IL-1β was inhibited by antagonists/inhibitors of TLR4, NLRP3 and caspase-1 indicating the involvement of the TLR4-NLRP3 inflammasome axis is mediating the NDPK response. Unlike the canonical NLRP3-inflammasome pathway that leads to monocyte cell death by pyroptosis, rNM23-H1 and rNDPKs did not lead to cell death indicating that rNDPKs are responsible for the activation of the alternative inflammasome. In our earlier studies, and those of others, we demonstrated that not all AML primary samples responsed to NM23-H1 in vitro. We have observed that non responders to NM23-H1 also do not respond to pathogen derived rNDPKs. In contrast, we have observed uniform responses in terms of cytokine release in all normal peripheral blood. We hence hypothesized that the non-rNDPK-responding AML samples may reflect the absence of monocytes in culture. To test this, we generated conditioned media using normal donor leukocytes, in presence or absence of a TLR-4 antagonist to inhibit the IL-1β production. The conditioned media was then used to culture primary AML samples, in parallel with rNDPK in unconditioned media. All the samples analyzed showed increased survival in rNDPK conditioned media even whilst some did not respond to rNDPK in unconditioned media. In summary, our data demonstrate for the first time that NM23-H1 and bacterial/fungal NDPKs are novel DAMPs/PAMPS that signal via TLR4 in monocytes. We further demonstrate that this interaction results in activation of the alternative NLRP3 inflammasome and subsequent cleavage and secretion of IL-1β without death by pyroptosis. Our data showing that bacterial/fungal NDPKs can promote survival of AML blasts indicates that rather than just being a consequence of AML associated immunosuppression, infections may drive the progression and AML. These findings have important implications in the clinical management of AML and its precursor myelodysplastic syndromes (MDS). Lilly AJ, et al.Cancer Res. 2011;71(3):1177-86.Gaidt MM, et al. Immunity. 2016;44(4):833-46 Disclosures Drayson: Abingdon Health: Consultancy, Equity Ownership.

Description: Elevated circulating levels of NM23-H1 have been associated with poor prognosis in haematological malignancies including AML and non-Hodgkin's lymphomas. In diffuse large B-cell lymphoma (DLBCL), several studies have demonstrated that both elevated circulating plasma levels and intra-tumoural levels of NM23-H1 correlate with poorer prognosis. These observations bring into question whether the mechanistic link of NM23-H1 expression with DLBCL prognosis depends an intrinsic intracellular mechanisms or is mediated by release of NM23-H1. We have shown that DLBCL lymphoma cell lines (Farage, HT, OCI-LY1, OCI-LY3, OCI-LY7, SUDHL4, SUDHL5 SUDHL6, U2932) express both NM23-H1 protein and the highly related protein NM23-H2. We also demonstrated that DLBCL cell lines release significant levels of NM23-H1 into their extracellular environment but not NM23-H2. Release of NM23-H1 was highly variable between cell lines with some releasing very high levels and some releasing much lower levels. Thus, DLBCL cell lines represent appropriate models to investigate the role of high and low levels of NM23-H1 on prognosis. The DLBCL cell lines HT and OCI-LY1 where selected for further study as representative of high and low level releasers of NM23-H1, respectively. CRISPR-Cas9 knockout of NM23-H1 (KO-NM23-H1) in HT and OCI-LY1 DLBCL cells had no impact on cell growth, viability or immuno-phenotype either in normoxic or hypoxic cultures. Thus NM23-H1 appears to not be required intrinsically by DLBCL cell lines. We therefore considered that the link between higher expression and release of NM23-H1 with prognosis is mediated via tumour-host environment interactions. We used an NSG mouse model transplanted subcutaneously with 1x106 CRISPR control (CTRL)-HT, CTRL-OCIL-Y1, KO-NM23-H1-HT or KO-NM23-H1-OCI-LY1 cell lines. We did not observe significant differences in tumour growth between CTRL and KO-NM23-H1 cells in the low NM23-H1 expressing OCI-LY1 DLBCL cell line. However, the high-expressing KO-NM23-H1-HT cells had significantly slower tumour progression and increased host survival when compared to CTRL-HT cells. We interpret to indicate that NM23-H1 release above a certain threshold provides a tumour growth advantage. Reduced lymphocyte:monocyte ratios (LMR) are also associated with poor prognosis in DLBCL. To investigate a potential functional link between NM23-H1 release and monocyte behaviour we first exposed peripheral blood leucocytes to Alexa 647 labelled fluorescent recombinant-NM23-H1 and found that monocytes but not neutrophils, T-cells or B-cells bound NM23-H1. We also observed that monocyte viability and survival were elevated in serum free cultures when supplemented with rNM23-H1, an observation that might indicate that elevated circulating NM23-H1 and reduced LMR in poor prognosis DLBCL may be mechanistically linked. We next co- cultured CTRL-HT or koNM23-H1-HT in a 1:1 ratios with purified monocytes from healthy human donors. Principle component analyses of 27 cytokines simultaneously measured by luminex assays identified that IL-1β, IL-6, IL-8, MIP-1α, MIP-1β and TNF-α were elevated when monocytes were co-cultured with CTRL-HT cells compared to co-culture with koNm23-H1-HT cells. In contrast, IP-10 was found elevated in the co-culture with koNm23-H1-HT cells. These observations indicate potentially important cross talk between the malignant DLBCL cells and innate immune cells of the host. Consistent with this, IL-6 and IL-8 serum levels have been shown to be elevated in pretreatment DLBCL patients compared to control subjects and MIP-1α, IL-6, and IL-8 serum levels have a greater association with DLBCL than follicular lymphoma. In conclusion ours is the first study to investigate potential mechanistic links between the associations of elevated NM23-H1 and poor prognosis in DLBCL and indicate a role for cross-talk between tumour cells and innate immune cells. Our data also indicate that NM23-H1 release from DLBCL cells may be a driving component in driving reduced LMRs that are also associated with poor prognosis. Whether reduced LMR and NM23-H1 release are casually related or not, the possibility that patients with both elevated monocytes and elevated circulating NM23-H1 levels are likely to represent a group with particularly poor prognosis requires urgent investigation. Disclosures Drayson: Abingdon Health: Consultancy, Equity Ownership. Rushworth:Abbvie: Research Funding; Janssen: Research Funding.

Description: The Medical Research Council Myeloma IX Trial (ISRCTNG8454111) examined traditional and thalidomide-based induction and maintenance regimens and IV zoledronic acid (ZOL) and oral clodronate (CLO) in 1960 patients with newly diagnosed multiple myeloma. Overall survival (OS) and skeletal-related event (SRE) data have been reported for the overall trial population. The present analysis investigated optimal therapy regimens for different patient populations in Myeloma IX. Patients were assigned to intensive or nonintensive treatment pathways and randomized to induction cyclophosphamide, vincristine, doxorubicin, and dexamethasone (CVAD) versus cyclophosphamide, thalidomide, and dexamethasone (CTD; intensive) or melphalan and prednisolone versus attenuated oral CTD (CTDa; nonintensive). Patients were also randomized to ZOL or CLO. In the nonintensive pathway, CTDa produced better responses and lower SRE rates than melphalan and prednisolone. ZOL improved OS compared with CLO independently of sex, stage, or myeloma subtype, most profoundly in patients with baseline bone disease or other SREs. In patients treated for ≥ 2 years, ZOL improved OS compared with CLO from randomization (median not reached for either; P = .02) and also from first on-study disease progression (median, 34 months for ZOL vs 27 months for CLO; P = .03). Thalidomide-containing regimens had better efficacy than traditional regimens, and ZOL demonstrated greater benefits than CLO.

Description: Key Points The use of ZOL is better than CLO in the improvement of SREs and survival in symptomatic myeloma patients at diagnosis. Response category posttransplant may influence the impact of bisphosphonate therapy.

Description: Key Points Free light chain ratio, M-protein concentration, and immunosuppression predict progression of MGUS to lymphoid malignancies.

Description: Background Renal impairment is a life threatening complication of myeloma with up to 20-25% of patients presenting with renal dysfunction. Outcome is poor as a result of a high early mortality. Around 28% of newly diagnosed myeloma patients with renal failure do not survive beyond 100 days compared with 10% overall. Studies have shown that within weeks of diagnosing myeloma with renal failure, treatment with dexamethasone alone or combined with bortezomib lowers serum free light chain (sFLC) levels by more than 50% in half of patients; achieving lower sFLC levels in this early period is associated with a greater chance of being alive and dialysis free at 100 days. Methods OPTIMAL is a randomised, multi-centre phase II trial of newly diagnosed myeloma patients with renal impairment. Renal impairment defined as 18 years old, chronic kidney disease (CKD) stage 4 or 5, not pregnant or risk of pregnancy for child bearing women, or partner of male participants, free of malignancies for 〉2 years, able to comply with all trial requirements and give fully informed consent. Patients were randomised to receive 4 cycles of either Bortezomib, Bendamustine and Dexamethasone (BBD) or Thalidomide, Bendamustine and Dexamethasone (BTD); all participants received bendamustine and dexamethasone in three week cycles. Treatment period for participants receiving 4 cycles of therapy was 12 weeks. Participants not considered suitable for autologous stem cell transplant (ASCT) could be given a further two cycles of treatment (up to 6 cycles in total) in their respective arms. The trial was powered to detect 23% differences in the percentage of patients achieving 〉50% reduction in sFLC between treatment arms, e.g. from 60% to 83%, with 80% power and a 5% 2-sided significance level, recruiting 60 patients in each arm. At the pre-planned interim analysis, the data and safety monitoring committee endorsed the closure of the trial as a 60% difference in sFLC was detected and there was no obvious benefit for the BTD arm. This was also endorsed by the trial steering committee and trial management group. Co-primary endpoints were sFLC response from baseline to week 6 (after receiving two cycles of trial treatment) and renal response according to the modified International Myeloma Working Group (IMWG) criteria after receiving four cycles of trial treatment. Secondary endpoints included haematological responses, toxicity and overall survival. Results OPTIMAL recruited 31 patients between March 2015 and March 2019 from seven centres within the UK; 16 on BBD and 15 on BTD. Fifty two per cent of patients were ≤70 years old, 55% male, 35% were CKD stage 4 and 65% were CKD stage 5, 48% had planned autologous-stem cell transplantation, 75% had ECOG performance status 0 or 1, 29% were on dialysis and 90% were ISS stage III. Serum free light chain response was assessed in 29 patients where samples were available at screening and at the end of two cycles of treatment. 81% of patients on BBD achieved vGPR compared to 23% on BTD, Fisher's p=0.006, table 1. Nine patients were on dialysis at the time of screening (6 on BBD and 3 on BTD). Complete or partial renal response was achieved by 2 (50%) of patients on BBD compared to 1 (11%) on BTD, Fisher's p=0.05, table 2. Two patients on BBD arm reported reversibility of dialysis dependency after four cycles of treatment. Two patients not previously on dialysis required dialysis after 4 cycles of BTD. Seven deaths were reported from the total 31 patients (5 (31%) on BBD arm and 2 (13%) on BTD arm). There were 33 reported serious adverse events (SAEs) 14 on BBD and 19 on BTD. Conclusion OPTIMAL demonstrated a significant increase in the number of sFLC responders after the first 2 cycles for those patients allocated BBD compared to BTD; this trend continued when assessing renal response after 4 cycles with more patients being dialysis independent after receiving BBD. Funding: Project funded by NAPP Pharmaceuticals, JANSSEN-Cilag Ltd and Bloodwise (formerly named Leukaemia and Lymphoma Research). Disclosures Ramasamy: Oncopeptides and Sanofi: Honoraria, Membership on an entity's Board of Directors or advisory committees; Abbvie: Honoraria, Membership on an entity's Board of Directors or advisory committees; Takeda: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Amgen: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Janssen: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; NAPP Pharmaceuticals Ltd.: Research Funding; Janssen-Cilag Ltd.: Research Funding. Drayson:Abingdon Health: Consultancy, Equity Ownership.

Description: Introduction: Multiple myeloma (MM) is more common in men than women but the mechanism(s) driving this are not understood. In our previous study (Myeloma IX) we found sex disparities in the cytogenetic lesions present in myeloma cells at the time of diagnosis and that female sex was associated with reduced overall survival in the context of treatment with traditional chemotherapy (CVAMP/MP) and thalidomide combinations. Here, we evaluate sex differences in almost 4000 patients recruited to the UK NCRI Myeloma XI trial, in which treatment exposure to lenalidomide predominated. Methods: Myeloma XI recruited newly diagnosed patients of all ages, with pathways for transplant eligible (TE) and ineligible (TNE) patients. An induction randomisation compared the triplet combination of cyclophosphamide, lenalidomide and dexamethasone to a similar combination with thalidomide (CRD vs CTD). Eligible patients underwent autologous stem cell transplant (ASCT) and in both pathways a maintenance randomisation compared lenalidomide (+/-vorinostat) until disease progression versus observation. We compared baseline characteristics of males and females using Fisher's Exact test for categorical characteristics and the Wilcoxon-Mann-Whitney test for continuous characteristics with p1 lesion present. Results: Of 3894 patients enrolled in the trial 2268 (58%) were male and 1626 (42%) were female, in keeping with the known sex disparity in MM. There was no difference in the median age, WHO performance status, ethnicity and most laboratory values of the two groups. Females were more likely to have the molecular risk lesions t(14;16) and del(17p) and had proportionately more HiR and UHiR disease, Table 1. Despite these differences, PFS and OS from induction randomisation did not significantly differ (PFS: Males 25 months, [95% CI 24, 26], Females 24 months, [95% CI 22, 25] and OS: Males 67 months, [95% CI 62, 70], Females 70 months, [95% CI 64, 74]). Molecular lesions that have been associated with outcome remained prognostic in both sexes, with a stepwise reduction in PFS and OS with cumulative risk lesions. PFS: Males SR 29 months, HiR 23 months, UHiR 16 months (p 〈 0.0001), Females SR 27 months, HiR 18 months, UHiR 17 months (p = 0.0007); OS: Males SR 77 months, HiR 59 months, UHiR 34 months (p 〈 0.0001), Females SR 82 months, HiR 54 months, UHiR 41 months (p 〈 0.0001). There was, however, no significant difference in PFS or OS when we compared males and females with a given cytogenetic lesion or cytogenetic risk. There was a significant difference in the proportion of patients of either sex who continued through the trial and underwent ASCT in the TE pathway (Males 72%, Females 67%; p = 0.031), but no significant difference in those that entered the maintenance randomisation (TE: Males 56%, Females 50%, p = 0.107; TNE Males 45%, Females 42%, p = 0.249). There was no significant PFS or OS difference by sex when analysed within each treatment pathway (TE, TNE), induction regimen (CTD, CRD) and maintenance approach (lenalidomide maintenance, observation). Conclusions: Females had a higher proportion of the adverse molecular risk lesions del(17p) and t(14;16) and were more likely to have HiR and UHiR disease. In the context of Myeloma XI trial treatment this did not correspond to a difference in PFS or OS, either overall or within the induction or maintenance randomisation treatment options (even though males were more likely to undergo ASCT). This suggests that in women the treatment delivered may have been able to overcome some of the adverse effect of the risk lesions present or that other factors affecting outcome were more important. on behalf of the NCRI Haematological Oncology Clinical Studies Group. Disclosures Cairns: Celgene, Amgen, Merck, Takeda: Other: Research Funding to Institution. Davies:Janssen, Celgene: Other: Research Grant, Research Funding; Amgen, Celgene, Janssen, Oncopeptides, Roche, Takeda: Membership on an entity's Board of Directors or advisory committees, Other: Consultant/Advisor. Boyd:Novartis: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Celgene: Consultancy, Honoraria; Takeda: Consultancy, Honoraria. Cook:Celgene, Janssen-Cilag, Takeda: Honoraria, Research Funding; Janssen, Takeda, Sanofi, Karyopharm, Celgene: Honoraria, Speakers Bureau. Drayson:Abingdon Health: Consultancy, Equity Ownership. Gregory:Abbvie, Janssen: Honoraria; Amgen, Merck: Research Funding; Celgene: Consultancy, Research Funding. Jenner:Abbvie, Amgen, Celgene, Novartis, Janssen, Sanofi Genzyme, Takeda: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau. Jones:Celgene: Honoraria, Research Funding. Kaiser:Takeda, Janssen, Celgene, Amgen: Honoraria, Other: Travel Expenses; Celgene, Janssen: Research Funding; Abbvie, Celgene, Takeda, Janssen, Amgen, Abbvie, Karyopharm: Consultancy. Owen:Celgene, Janssen: Honoraria; Celgene, Janssen: Consultancy; Celgene: Research Funding; Janssen: Other: Travel expenses. Russell:DSI: Consultancy, Honoraria, Speakers Bureau; Jazz: Consultancy, Honoraria, Speakers Bureau; Pfizer Inc: Consultancy, Honoraria, Speakers Bureau; Astellas: Consultancy, Honoraria, Speakers Bureau. Morgan:Bristol-Myers Squibb, Celgene Corporation, Takeda: Consultancy, Honoraria; Amgen, Janssen, Takeda, Celgene Corporation: Other: Travel expenses; Celgene Corporation, Janssen: Research Funding. Jackson:Celgene, Amgen, Roche, Janssen, Sanofi: Honoraria. Pawlyn:Amgen, Celgene, Takeda: Consultancy; Amgen, Janssen, Celgene, Takeda: Other: Travel expenses; Amgen, Celgene, Janssen, Oncopeptides: Honoraria. OffLabel Disclosure: CTD/CRD induction therapy for myeloma, Lenalidomide maintenance 10mg 21/28 days

Description: Abstract 335 Background Induction treatment for myeloma has focused on improving response rates and many centres have evaluated and used triplets of therapy based on either IMiD or proteosome inhibitor drugs. Other strategies rely on combining the two classes of novel agents and excellent response rates have been obtained. An alternative approach to this, which moves some way towards personalised therapy, is to use sequential combinations of the two classes of novel agents dependent on the response achieved. The aim of this approach is to maximise responses and, by so doing, improve survival. Methods In order to test this concept, we have carried out a trial, Myeloma XI, which compared responses to cyclophosphamide, thalidomide and dexamethasone (CTD) with cyclophosphamide, lenalidomide and dexamethasone (CRD). Following the randomised use of either of these triplets, given to maximum response, patients achieving no response (NC or PD) received further therapy with a triplet composed of cyclophosphamide, bortezomib and dexamethasone (CVD). In order to test if a bortezomib combination could improve the response of maximally treated patients, those with a suboptimal response (MR or PR) were randomised to either no further therapy or to CVD. For patients achieving ≥ VGPR no further induction was given. Younger fitter patients went on to receive HDM plus ASCT whereas older less fit patients did not. All patients were eligible for a maintenance randomisation to no maintenance, lenalidomide or lenalidomide/vorinostat maintenance. For patients receiving CVD, treatment was continued to a maximum of 8 cycles, with response being assessed using IMWG criteria after each cycle and therapy continued to maximum response or intolerance. Here we present the response rates for patients who have received CVD. Results At the time of this initial analysis 1424 patients have been randomised overall; 790 in the intensive pathway and 634 in the non-intensive pathway. Across both pathways, of the refractory patients with no response to induction treatment, an upgrade in response from NC or PD to ≥MR was seen after CVD treatment in 58% of patients, with 31% going on to achieve a VGPR or CR. In the other group of patients, those with a suboptimal response to induction treatment and randomised to receive CVD, 45% went on to achieve VGPR or CR. These patients reduced their paraprotein value by a mean of 74% from the start of CVD suggesting that the improvement in response is significant and not the result of patients marginally crossing the boundaries of response criteria. It seems that even in a group of maximally treated patients that response rates can be increased further by the use of a proteosome inhibitor drug. In the MRC Myeloma IX study the response rates at a similar time point for those treated with similar thalidomide based induction were PD/NC 7.1%, MR/PR 41.8%, VGPR/CR 37.5%. Assuming at least the same response rates in the current study then we calculate that it would be possible to get approximately 45% extra patients from MR/PR and 31% from PD/NC to ≥VGPR, equating to an approximate 21% increase in the ≥VGPR rate to 59% of the total patients. Conclusions We have demonstrated an improvement in response with the addition of a bortezomib based regimen in the group of patients who had a suboptimal response following induction chemotherapy with an IMiD based regimen. Particularly encouraging is the excellent rate of achievement of CR/VGPR in patients with no initial response to treatment. We anticipate that together this will significantly improve the percentage of patients achieving ≥VGPR, approximating 60% of all patients following induction, a response rate that should improve further following ASCT. Further follow up is required to know whether or not this will translate into an improvement in OS or PFS, however from previous studies we would expect this to be the case. Disclosures: Pawlyn: Celgene: Unrestricted educational grant Other. Off Label Use: Lenalidomide and vorinostat as maintenence therapy. Davies:Celgene: Honoraria, Speakers Bureau, Unrestricted educational grant, Unrestricted educational grant Other; Ortho Biotech: Honoraria, Speakers Bureau. Gregory:Celgene: Unrestricted educational grant Other. Szubert:Celgene: Unrestricted educational grant Other. Bell:Celgene: Unrestricted educational grant Other. Ouzman:Celgene: Unrestricted educational grant Other. Drayson:Celgene: Unrestricted educational grant Other. Owen:Celgene: Unrestricted educational grant Other. Jackson:Celgene: Honoraria, Unrestricted educational grant Other. Russell:Celgene: Unrestricted educational grant Other. Morgan:Ortho Biotech: Honoraria, Speakers Bureau; Celgene: Honoraria, Speakers Bureau, Unrestricted educational grant, Unrestricted educational grant Other.

Description: Key Points The type of antibody secreted at relapse can serve as a marker of clonal heterogeneity. It is important to monitor for serum FLC in the suspicion of clinical relapse to ensure that FLC relapse is not missed.

Description: The role of maintenance therapy for the long term control of the plasma cell clone in patients induced into response with either intensive or conventional treatment is an important outstanding question. We addressed this in the MRC Myeloma IX study which incorporates intensive and non-intensive pathways selected according to PS and age. In the intensive pathway patients were randomised to either CTD or CVAD induction, followed by High Dose Melphalan (HDM) before being randomised to either thalidomide or no maintenance. In the non-intensive pathway patients were randomised to either MP or attenuated CTD prior to the maintenance randomisation. For patients randomised to thalidomide it was initiated at d100 following HDM or at the end of induction in the non-intensive arm with the aim of delivering 100mg daily until relapse. A dose reduction algorithm for side effects was used. Between the years of 2003–8, 820 patients were entered into the maintenance randomisation, median age 64 (intensive 59, non-intensive 73), median follow-up 32 months. Prognostic features were evenly distributed between the arms. FISH and cytogenetics were done using standard methods. Response was assessed by IWG criteria. For overall survival (OS) there was a non-significant trend in favour of the no maintenance arm, which enables us, by calculating confidence limits on the hazard ratio, to make the assertion that no maintenance could be up to 7% worse than thalidomide at 5 years (p=.005). Further analysis showed that there was no significant difference in OS in either the intensive or the non-intensive arm. The duration of time on thalidomide maintenance appeared to make no difference to OS. There was a non-significant improvement in progression free survival (PFS) across the maintenance randomisation as a whole and in the intensive pathway a significant benefit of maintenance was seen in the patients achieving less than a VGPR post initial induction therapy prior to HDM, (hazard ratio 1.9, p=.007). This PFS difference did not translate into a survival benefit because the survival after progression in the PR patients receiving maintenance thalidomide was poor (p=.002). In addition we looked at the time spent off thalidomide, the recovery time, (the time between stopping thalidomide and progression) as a possible predictor of survival after progression. Treated as a continuous variable in the Cox model this showed a trend for longer survival after relapse in those with longer recovery time (p=.056). In the non-intensive pathway a similar but less pronounced effect of thalidomide maintenance on PFS was seen. These results are consistent with a consolidation rather than a maintenance effect for thalidomide in this setting. The impact of maintenance in different cytogenetic subgroups was also determined [17p-, 13q-, 14q abnormalities including t(4;14), t(14;16), t(6;14), t(14;20) and t(11;14)]. For the 17p- group, the difference in OS between no thalidomide and thalidomide is large (HR = 4.55, p=.02) with the thalidomide patients faring worse, although this is based on only 30 patients. For the non 17p- group there is no difference in PFS (HR = 1.24, p=.37), in the 17p- group, however, the PFS is worse. In addition, of the 22, 17p- patients receiving CTD or CTDa as initial therapy, the 10 who received no thalidomide maintenance are all still alive, whereas 9/12 of those who went on to receive thalidomide maintenance have died. It seems that thalidomide given at induction and again in maintenance, may be particularly detrimental in 17p- patients. Although thalidomide maintenance may improve PFS, there is no demonstrable benefit on OS. It is important to identify 17p- in order to exclude these patients from receiving thalidomide maintenance.