ALBERT

Description: Aim/Background CNS relapse in DLBCL is a devastating complication and the optimum strategy for prevention remains unclear. Methods We performed a multi-centre, retrospective analysis of CNS relapse rates in patients (pts) identified at risk of CNS relapse according to the type of CNS-directed prophylaxis administered. Pts receiving initial therapy for DLBCL between 1996 and 2011 (to allow ≥2 years of follow up) were included; DLBCL after histologic transformation of low-grade lymphoma and HIV-associated DLBCL were included, however pts with Burkitt or Burkitt-like lymphoma or CNS involvement at diagnosis were excluded. Selection for CNS prophylaxis strategy was by the primary managing hematologist if they fulfilled ≥2 of the following criteria: 1) multiple extranodal sites 2) raised serum LDH 3) B-symptoms, OR involvement of specific high-risk anatomical sites. We compared 3 prophylaxis strategies: prior to 2003 intrathecal (IT) methotrexate (MTX) in conjunction with CHOP chemotherapy “group 1” was the main strategy; from 2003 onwards, R-CHOP (mostly) with IT MTX was followed by two cycles of high dose intravenous (IV) MTX (1-3g/m2) “group 2”; patients

Description: Introduction. Despite considerable efforts towards novel therapeutics research and discovery, outcomes in AML remain poor. Although composite Complete Remission (cCR; CR, CRp, CRi) is routinely used as a measure of therapeutic clinical activity, cCR frequently does not translate to survival benefit. On this basis, the characterization and tracking of minimal residual disease (MRD) has emerged as a tool to help better define the depth of such cCR to offer prognostic utility on AML patients likely to experience survival benefit from a given experimental therapeutic. The MDM2 antagonist idasanutlin has shown promising clinical activity in AML. Idasanutlin enhances p53 activity through antagonism of the MDM2:p53 interaction. Disruption of this protein:protein interaction inhibits MDM2 targeting of p53 for ubiquitination and degradation, thus stabilizing p53 protein to exert tumor suppressor transcriptional regulation and induction of apoptotic pathways. Patients and Methods. Trial NP28679 (NCT01773408) is a Phase 1/1b study evaluating idasanutlin as monotherapy or in combination with cytarabine in relapsed or refractory AML patients with safety as primary and complete remission as secondary endpoints, respectively (Martinelli, EHA, 2016.) Duration of response was available as exploratory clinical data for a subset of patient. Patients' pre-treatment bone marrow aspirate specimens were evaluated by multiparametric flow cytometry using an 18 marker surface antigen-based panel. MRD assessment occurred per protocol recommendation at the time of hematological malignancy response assessment (HMRA) at Day 28 and for those patients initially experiencing cCR at each subsequent HMRA. Further flow cytometry analyses were conducted for expression changes of known p53 regulated proteins in CD45+(dim) blasts correlating with drug exposure, consistent with mechanistic engagement. Results. PFS analyses for cCR patients versus non-CR Kaplan-Meier plots indicate the median for responders is 315d (95%CI: 282, NA) versus non-responders is 43.5d (95%CI: 30, NA). When MRD1% is applied as a cut-point, analytics show a statistical association with median PFS (log-rank p-value 〈 .001) at 367d (95%CI: 219, NA) for 1% is 84d (95%CI: 30, NA.) Median MRD values for CR/CRp/CRi vs PR/HI vs PD are 0.42%, 1.79%, and 19.16%, respectively, again consistent with MRD serving as a surrogate for clinical activity (log rank test for trend p-value = .001) Additionally, multiparametric flow cytometry analysis of idasanutlin pharmacodynamic (PD) protein expression changes comparing pretreatment patient blood specimens to 24 hours following first dose administration indicates that increases in both p53 and MDM2 in CD45+(dim) blasts is associated with steady state drug exposure levels ([Spearman's Correlation =0.27; p=.013] and [Spearman's Correlation=0.24; p=.026,] respectively.) Interestingly, these changes in protein expression for p53 and MDM2 are associated with orthologous PD changes in serum protein macrophage inhibitory cytokine 1 (MIC-1) ([Spearman's Correlation=0.23; p=.035] and [Spearman's Correlation=0.22; p=.042,] respectively.) These PD changes are mechanistically consistent with enhanced p53 resulting from diminished MDM2 ubiquitination and degradation of p53. Conclusions. In summary, the results presented here are consistent with multiparametric flow cytometry MRD assessment as an early indication aligning with cCR in relapsed/refractory AML patients treated with idasanutlin. Further, assessment for association with progression-free survival indicates that the lowest quartile patients by MRD flow cytometry measurement (1%). As such, the data presented here support inclusion of MRD assessment by flow cytometry as a prognostic indicator to provide guiding information for assignment of depth of AML patient response. We will continue to monitor this biomarker for diagnostic potential as a prognostic indicator of survival-based outcomes in future randomized clinical studies of idasanutlin. Disclosures Lanza: Roche-Genentech: Employment. Martinelli:Pfizer: Consultancy, Speakers Bureau; MSD: Consultancy; Novartis: Speakers Bureau; Amgen: Consultancy, Speakers Bureau; Genentech: Consultancy; Celgene: Consultancy, Speakers Bureau; Ariad: Consultancy, Speakers Bureau; Roche: Consultancy, Speakers Bureau; BMS: Speakers Bureau. Yee:Novartis Canada: Membership on an entity's Board of Directors or advisory committees, Research Funding. Jukofsky:Roche Pharma: Employment. Reis:Roche Pharma: Employment. Blotner:Roche Pharma: Employment. Drummond:Pfizer: Honoraria, Speakers Bureau; celgene: Honoraria, Research Funding, Speakers Bureau; Novartis: Honoraria, Research Funding, Speakers Bureau; BMS: Honoraria, Speakers Bureau; Roche: Honoraria, Speakers Bureau. Vey:Sunesis: Honoraria. Dickinson:GlaxoSmithKline: Consultancy, Research Funding. Kelly:Novartis: Consultancy, Speakers Bureau; Amgen: Consultancy, Membership on an entity's Board of Directors or advisory committees; Pharmacyclics: Consultancy, Speakers Bureau. Theron:Roche Pharma: Employment. Venstrom:Roche-Genentech: Employment. Middleton:Roche Pharma: Employment. Chen:Roche Pharma: Employment. Kinnersley:Roche-Genentech: Employment, Equity Ownership. Nichols:Roche Pharma: Employment. Pierceall:Roche Pharma: Employment.

Description: Abstract 2652 Introduction Despite improvements in cure rates for patients with diffuse large B-cell lymphoma (DLBCL), up to 40% relapse after achieving initial remission, mostly within 18 months from treatment. There is no consensus as to role, or most appropriate form of post-remission surveillance. Our aim was to explore the role of Positron Emission Tomography combined with computer tomography (PET-CT) scanning in the follow up of patients with diffuse large B-cell lymphoma (DLBCL) achieving complete metabolic response (CMR) after primary therapy, identify patterns of relapse and define a risk-adapted strategy. Results We included 116 patients with de novo DLBCL treated at our centre between 2002 and 2009 with a negative post-treatment PET-CT, and at least one surveillance PET-CT scan. International Prognostic Index (IPI) was

Description: Introduction. Despite years of translational research and discovery, outcomes in Acute Myeloid Leukemia remain poor. The MDM2 antagonist idasanutlin has shown promising clinical activity in solid tumors and acute leukemias. Idasanutlin enhances the activity of the tumor suppressor, p53, through antagonism of MDM2:p53 interaction. Such protein:protein disruption abolishes MDM2 targeting of p53 for ubiquitination and degradation. Subsequent stabilization of the p53 protein allows it to exert tumor suppressor transcriptional regulation and induction of apoptotic pathways. Identification of patients in whom functional p53 activation drives efficacy may translate into improved outcomes for patients treated with idasanutlin. Patients and Methods. Trial NP28679 (NCT01773408) is a Phase 1/1b study evaluating idasanutlin (as monotherapy or in combination with cytarabine) in relapsed or refractory AML patients, with clinical response as the primary endpoint. To identify biomarkers of response, patients' pre-treatment peripheral blood specimens were evaluated for MDM2 protein expression levels in leukemic blasts and leukemic stem cells. Association of MDM2 percent cell positivity with clinical outcomes was evaluated using intracellular flow cytometry gated on CD45dim leukemic blasts and CD45dim/CD117+/CD34+ leukemic stem cells. Results. We observed in a proof-of-principle training data set that MDM2 expression in leukemic blasts was associated with patients more likely to exhibit Complete Remission (CR, CRp, CRi) versus Progressive Disease (PD, HI, PR) [n=61; Wilcoxon p = .0041 (AUC = .74; 95%CI[.57, .91])] TP53 mutational status alone was not tightly associated with patient response (Fisher's Exact Test p = .19 [AUC = .60; 95%CI [.52,.67.])] When MDM2 protein expression and TP53 mutational status were analyzed as co-variates, an enhanced association with patient CR was observed (AUC = .76; 95% CI [.59, .93.]) A separate group of patients treated with an optimized idasanutlin formulation was utilized as a validation set. Association of CR with MDM2 expression in blasts was also shown in this separate patient population [n=24; Wilcoxon p = .0052 (AUC = .82; 95% CI [.64, 1.00.]) Comparable results were observed for proof-of-principle and validation set analyses of response association with MDM2 protein expression in AML patients CD45dim/CD117+/CD34+ stem cell subpopulations. Conclusions. In summary, training and validation sets reveal that MDM2 protein expression in leukemic blasts and stem cells are associated with idasanutlin-induced CR in patients with AML. We will continue to monitor this potentially predictive biomarker in future randomized clinical studies of idasanutlin. On a broader level, the data presented here support the concept that leukemic blasts may be "oncogene-addicted" to MDM2 in AML, i.e. that AML tumor cells may rely on a dominant oncogene for growth and survival, such that inhibition of the function of this specific oncogene is sufficient to halt the neoplastic phenotype. Thus, the concept of oncogene addiction may apply not only to mutated kinases but also to ubiquitin ligases such as MDM2. Disclosures Reis: Roche Pharma: Employment, Equity Ownership. Jukofsky:Roche Pharma: Employment, Equity Ownership. Chen:Roche Pharma: Employment, Equity Ownership. Martinelli:BMS: Speakers Bureau; MSD: Consultancy; Roche: Consultancy; ARIAD: Consultancy; Novartis: Speakers Bureau; Pfizer: Consultancy. Zhong:Roche Pharma: Employment, Equity Ownership. So:Roche Pharma: Employment, Equity Ownership. Drummond:Baxalta: Membership on an entity's Board of Directors or advisory committees; Novartis: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Celgene: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Gilead: Membership on an entity's Board of Directors or advisory committees. Assouline:Pfizer: Consultancy; BMS: Consultancy; Novartis: Consultancy. Hashemyan:Roche Pharma: Employment, Equity Ownership. Theron:Roche Pharma: Employment, Equity Ownership. Blotner:Roche Pharma: Employment, Equity Ownership. Rueger:Roche Pharma: Employment, Equity Ownership. Middleton:Roche Pharma: Employment, Equity Ownership. Vey:Celgene: Honoraria; Roche: Honoraria; Janssen: Honoraria. Nichols:Roche Pharma: Employment, Equity Ownership. Chen:Roche Pharma: Employment, Equity Ownership. Pierceall:Roche Pharma: Employment, Equity Ownership.

Description: Background: Pre-existing thrombocytopenia in MDS/AML is worsened during the initial cycles of azacitidine (AZA) therapy, resulting in bleeding risk and possible platelet transfusion. Eltrombopag (EPG) is an oral TPO-receptor agonist. In vitro, it has anti-proliferative effects on AML blasts. Aim: 
To assess the safety of escalated doses of EPG in patients undergoing AZA for MDS/AML Method
An investigator-initiated phase-II, single arm, study of EPG with AZA. Inclusion: relapsed or de-novo MDS/CMML/AML (blasts 5-30%); or symptomatic cytopenia; or blasts 31-50% if 〉/=65 years or previously-treated disease; and platelets Result
Of 25 patients, 10 had prior therapy, which was chemo in 7; of these 6 had blasts 〉/=10%, 2 with blasts 20% or above. Of the 15 de-novo patients, 7 had blasts 〉/=20% (AML) and a further 2 had blasts between 10 and 19%. The median platelet count was 38x10^9/L (range 8-127). A median 11(2-24) cycles AZA and 6 cycles of EPG were delivered. One patient developed GrII EPG-related LFT abnormalities (resolved). Grade 3 fatigue was attributed to the combination in one patient. Thrombocytosis (〉450 x109/L) resulting in EPG cessation occurred in 6 (at 50, 50, 150 and three at 200mg), without complications. 10 patients experienced reversible skin yellowing. Response/improvement was seen in 18 (72%): 7CR, 3CRm, 5HI-P,1 HI-N,2 with 〉50% blast reduction from 〉20% (8%). 5 patients had progression at first response. There were 19 events in total (17 progressions, 2 with worsening haematology). Median PFS was 12.0 months (95% CI 5.6 to 24.3 months). Median OS was 15.3 months (95% CI 11.9 to 31.7 months). Platelet improvement was seen in 54% (13/24) of patients with baseline platelets

Description: Background Patients (pts) with TrIL have inferior outcomes compared with de novo diffuse large B-cell lymphoma (DLBCL), and their optimum follow up is not well defined. We sought to determine the utility of surveillance PET-CT in pts with TrIL achieving CMR after primary therapy and identify patterns of relapse. Methods We performed a retrospective analysis of pts with TrIL treated at Peter MacCallum Cancer Centre between 2002 and 2012 who achieved CMR after primary therapy who had ³1 subsequent surveillance PET-CT. In the period analysed, departmental protocol recommended 6-monthly scans for pts in CMR for the first 2 years, then annually until 5 years after completion of therapy, if there was intention to intervene on detection of subclinical relapse. A positive scan suggested relapsed lymphoma, with true positive results requiring either biopsy confirmation or unequivocal scan progression. A false positive scan was refuted by biopsy and/or follow up showing resolution of areas of increased FDG uptake. A negative scan was interpreted as negative for relapsed lymphoma: true negatives had no clinical relapse and false negatives manifest relapse within three months from the date of the scan. Indeterminate scans were recorded if determination could not be made. Results The cohort included 55 pts with TrIL: 38 underwent stem cell transplant (autologous, n= 37; allogeneic, n=1) as consolidation; 17 did not. (Table 1). After a median follow-up of 34 (range 3 – 101) months, the actuarial 3-year progression free (PFS) and overall survival (OS) were 77% (95% CI 62 – 86%) and 88% (75 – 94%) respectively. Multiple potential prognostic factors including IPI, stage, serum LDH, timing of transformation (simultaneous diagnosis of transformation versus delayed) and type of indolent histology were not predictive of PFS. Of 180 surveillance PET-CT scans, there were 153 true negatives, 4 false positives, 1 false negative, 7 indeterminate and 15 true positives (Table 2). Considering indeterminate scans as false positives, the specificity of PET-CT for detecting relapse was 93%, sensitivity 93%, positive predictive value 54% and negative predictive value 99%. Of the 15 pts who experienced disease relapse, 7 (47%) were subclinical (i.e. detected by surveillance PET-CT scans) whilst 8 (53%) were suspected on the basis of clinical symptoms. Although 5% of scans in the first 2 years detected a subclinical relapse, all of these were either biopsy or clinically shown to be low-grade lymphoma. All 8 symptomatic relapses (at 2 – 42 months), in contrast were DLBCL. Conclusion In pts with TrIL achieving CMR, PET-CT detects subclinical relapses of low-grade histology with high sensitivity but with a moderate false-positive rate. This is of limited clinical benefit as the initiation of further therapy in these circumstances is rarely based on imaging findings alone. In contrast, all DLBCL relapses in our cohort were accompanied by clinical symptoms. Thus, surveillance imaging of pts with TrIL achieving CMR is not indicated. PET-CT should be reserved for evaluation of suspected relapse only. Disclosures: No relevant conflicts of interest to declare.

Description: Background: Bromodomains (BRDs) are domains found in a variety of proteins that recognize and bind to acetylated lysine residues in histone and other target proteins. The BRD and extra-terminal (BET) family of BRD-containing proteins bind to acetylated histone tails, alters chromatin structure and facilitates transcriptional complex localization to specific genes, thereby regulating gene transcription. The investigational agent GSK525762 is a potent small molecule inhibitor of the BET family of proteins that prevents assembly of macromolecular complexes and transcriptional response. GSK525762 inhibits growth in a broad spectrum of human hematological cancer cell lines, including cell lines derived from human patients with acute myeloid leukemia (AML), non-Hodgkin's lymphoma (NHL), and multiple myeloma (MM). GSK525762 is orally bio-available and can cause tumor reduction and improved animal survival in in vivo xenograft models of hematologic malignancies. Methods: This is a Phase I/II, open-label, 2-part study. Part 1 is a dose-escalation phase to determine the safety, tolerability, and recommended Phase 2 dose (RP2D) of GSK525762 in adult subjects with relapsed/refractory AML, NHL, and MM. Dose escalation is performed independently for each of these three cohorts. An accelerated dose titration was employed with one subject per dose level until the occurrence of a ≥Grade 2 drug-related toxicity or dose-limiting toxicity; thereafter, subjects have been enrolled in a standard 3+3 design. A Neuenschwander continual reassessment method (N-CRM) model is used at each dose escalation decision to provide guidance for the next dose escalation level. Starting dose is 5 mg GSK525762 orally once daily and dose escalation continues until the MTD is identified. All data, including safety, tolerability, pharmacokinetics (PK), and efficacy, are used to identify the RP2D. In Part 1, approximately 60-70 subjects will be enrolled (approximately 20 in each of three disease-specific cohorts); no hypothesis is being tested, and all analysis will be descriptive and exploratory. In Part 2, the clinical activity of GSK525762 (overall response rate) will be evaluated in expansion cohorts of subjects with AML, NHL, and MM. Up to 32 subjects may be enrolled into the AML and NHL cohorts, and up to 37 subjects may be enrolled in the MM cohort. Cohorts may be closed early if they do not exceed futility assessment. In addition, an exploratory cohort of subjects with double-hit lymphoma (DHL) and triple-hit lymphoma (THL) will be enrolled to evaluate clinical activity in this patient population. Additional study objectives include analysis of PK after single and repeat dosing, evaluation of pharmacodynamics (PD) and the relationship between GSK525762 exposure and safety/efficacy/PD parameters. Recruitment is ongoing across five centers (USA, UK, and Australia). Currently, 40 subjects have been enrolled (29 AML, 8 NHL, and 3 with MM). Study funded by GSK. Disclosures Stein: Seattle Genetics: Research Funding; Agios Pharmaceuticals: Other: Advisory Board, Research Funding; Celgene: Other: Advisory Board, Research Funding; Novartis: Consultancy. Huntly:Novartis: Speakers Bureau; BMS: Speakers Bureau; Ariad: Speakers Bureau; Pfizer: Speakers Bureau. Dickinson:GlaxoSmithKline: Consultancy, Research Funding. Horner:GlaxoSmithKline: Employment. Brennan:GlaxoSmithKline: Employment. Baron:GlaxoSmithKline: Employment. Kremer:GlaxoSmithKline: Employment, Equity Ownership. Dhar:GlaxoSmithKline: Employment.

Description: Introduction: Hypomethylating agents (HMA), including AZA, used to treat cytopenias in MDS patients (pts), can exacerbate thrombocytopenia. Such pts are usually treated with repeated platelet transfusions and AZA dose adjustments. Relieving thrombocytopenia may reduce platelet transfusion requirements and allow optimal AZA dosing. Eltrombopag is an oral TPO receptor agonist approved for the treatment of pts with chronic ITP, hepatitis C virus-related thrombocytopenia, and recurrent severe aplastic anemia. SUPPORT was a randomized, double-blind, placebo-controlled trial investigating the platelet supportive care effects of eltrombopag versus placebo in pts with intermediate (int)-1, int-2 or high-risk MDS receiving AZA. Methods :Adult pts with no previous exposure to HMA, baseline (BL) platelets

Description: Abstract 4180 High risk acute myeloid leukaemia (AML) and myelodysplastic syndrome (MDS) may be amenable to immunotherapy with adoptive transfer of chimeric antigen receptor (CAR) T cells. However, persistence and generation of central memory (CM) are essential for ongoing clinical responses (Morgan et al Science 2006). Our prior work showed that the carbohydrate antigen LewisY (LeY) is expressed in 46% of AML patients (n=33) to varying degrees, making it a suitable and novel target for CAR therapy. In our phase I study, we have assessed the trafficking, persistence and functional capacity of CAR-T cells transduced with an anti-LeY chimeric receptor gene in high-risk patients with LeY positive AML or MDS. METHOD: The CAR comprised extracellular humanized scFv recognizing the LeY Ag, linked to an extracellular CD8 hinge region, a transmembrane and cytoplasmic CD28 and CD3 zeta signalling domain. The humanized anti- LeY scFv-CD28-ζ receptor vector was produced as described previously (Westwood J et al PNAS 2005). T cells, harvested from the patient, were transduced with LeY scFv-CD28-ζ pSAMEN retroviral vector. Bulk transduced T cells (LeY-T) were re-infused into the patient after lymphodepleting fludarabine chemotherapy. AML immunophenotyping and cytogenetics were used to monitor minimal residual disease (MRD). PB and BM samples were collected prior to and post CAR-T adoptive transfer. An optimized PCR assay (sensitivity 1:1e5) for the presence of the LeY transgene (TG) was performed on genomic DNA extracted from each sample. To measure both CAR-T cell functional polarization and persistence of CAR expression, PB or BM cells were co-cultured with a cell line expressing LeY (OVCAR-3), in the presence or absence of MHC class I or II blocking antibodies. Culture supernatant was collected at 24 hours and assessed by Luminex assay for Th1 (IFN-g, IL-2), Th2 (IL-4, IL-10), pro-inflammatory (IL-6, IL-17 and TNF) cytokines and TGF-b. RESULTS: Five AML patients have been enrolled to date. Four patients have had sufficient CAR-T cells generated to provide doses of 0.5 × 109 −1.3 × 109T cells (transduced T cells: 8 – 30% of total T-cells]) with a viability of 96.1– 98.5%. 2 patients (01 and 04) had progressive disease 1 and 2 months after infusion. 2 patients (02 and 05) remain in morphologic remission with stable cytogenetic MRD, 3 and 14 months post infusion. Patients 01, 02 and 04 all had LeY TG in PB and BM samples at all time points measured, up to day 49 in patient 01 (Figure 1) and up to 10 months in patient 02. In patient 04 LeY TG was also detected in a skin biopsy (day +6) of leukemic skin infiltrate. Co-culture supernatant cytokine analysis showed LeY-T cells from patient 01, 02 and 04 secreted high levels of IFN-g (805.5 ± 198.9 pg/ml) and low levels of IL-2 (3.14 ± 0.65 pg/ml) prior to adoptive transfer. In contrast, post adoptive transfer, co-culture with OVCAR-3 cells, and MHC class I and II blocking antibody, induced these PB LeY-T cells to secrete IL-4 (24.13 ± 12.47pg/ml) and IL-10 (9.35 ± 0.44 pg/ml); no IFN-g or IL-2 was detected. This Th2 polarization of CAR-T cells was shown in PB samples taken at days 5 to 28 (patient 01), day 28 to 8 months (patient 02) and day 1 to day 28 (patient 04) post-adoptive transfer (Figure 2). In addition, when Th1/Th2 cytokines were measured in patient 02 PB and BM plasma, Th1 cytokines were detected immediately post-adoptive transfer (IFN-g=39 pg/ml, IL-2=8.6 pg/ml at peak); however these decreased to basal levels by day 3 (IFN-g=1.8 pg/ml) and day 2 (IL-2=2.6 pg/ml). The same trend in serum cytokines was observed for patient 01. Therefore, our studies reveal that, after adoptive transfer, LeY CAR-T cells persist in vivo, maintain their expression of the CAR, but show rapid polarization from a Th1 to a Th2 phenotype. Disclosures: No relevant conflicts of interest to declare.

Description: Introduction FDG-PET imaging is a powerful predictor of both positive and negative outcome of initial therapy for DLBCL. Existing data on the utility of PET to predict outcome of ASCT have included a range of histological subtypes, limiting its direct applicability to the setting of DLBCL. Methods Data from 41 consecutive patients who had PET scans immediately prior to undergoing ASCT for treatment of first relapse of de novo DLBCL were analysed. We compared the clinical characteristics, event free survival (EFS) and overall survival (OS) of those with residual FDG avid disease (n=21) against those without FDG avid disease (n=20). Results The median age was 53(18–71). Patient characteristics are shown in the table. At 3 years, the OS was 67(±8)%, and the EFS was the 62(±8)%. EFS for PET positive patients was 38(±12)% vs 79(±9%) for PET negative patients (p=0.04). OS was 37(±17%) vs 81(±10%) (p=0.08) respectively. Conclusion Excellent EFS and OS can be anticipated from ASCT when undertaken in those who have achieved a PET negative remission following salvage chemotherapy. Conversely, failure to achieve a PET negative remission after salvage chemotherapy in relapsed DLBCL predicts for a poorer, but not irretrievable outcome after ASCT, suggesting that a proportion of PET positive patients can be successfully salvaged by either ASCT or post-ASCT salvage therapies. Patient Characteristics PET Positive (n=21) PET negative (n=20) PET negative (n=20) Female 28% 40% Median Age (Range) 53 (28–65) 53(81–71) Salvage Regimen ICE ±R 66% ICE ±R 80% Conditioning Regimen BEAM 47%, SBCNU 23% BuMel 14% other 16% BEAM 25%, SBCNU 25% , BuMel 25% other 25% Rituximab with salvage 71% 95% Rituximab in initial therapy 33% 25% RT with salvage or ASCT 43% 25%