Buadi, Francis K.

Abstract Background: Autologous stem cell transplantation is an effective therapy for patients with multiple myeloma. We and others have previously reported the influence of lenalidomide based regimens on the ability to harvest adequate number of stem cells for successful transplantation. In order to identify factors predicting for poor mobilization we studied a large group of patients who underwent an attempt at stem cell mobilization after receiving lenalidomide and dexamethasone as primary therapy for myeloma. Methods: We identified sequential patients who received lenalidomide and dexamethasone as initial therapy for their myeloma and then underwent stem cell mobilization for immediate or future stem cell transplantation. Patients who received any other regimen prior to the stem cell mobilization were excluded. Between July 2004 and May 2008, 106 patients, satisfying the above criteria were identified from the Mayo Clinic transplant database. Medical records and collection sheets were examined for the data. Results: The median (range) age at mobilization was 60 yrs (29–75); 34 (32%) were over 65 yrs and 59 (55%) were males. The median duration of lenalidomide therapy was 4 months (range; 1–13). The strategy for stem cell mobilization was GCSF alone in 92 patients (87%), cyclophosphamide (CTX) and GCSF in 11 pts and 3 pts received AMD3100 and GCSF. Among the GCSF mobilized patients, 10 pts (11%) failed to collect at least 2.5 million cells required for one transplant, including 8 patients who never achieved the minimum peripheral count threshold to initiate the collection. Two of the 11 pts undergoing primary mobilization with CTX/GCSF failed to collect any cells, while all of the 3 pts mobilized with AMD3100 were successful. Five of these pts subsequently underwent successful salvage mobilization with CTX/GCSF, 1 with AMD3100, one failed to mobilize with CTX and the rest did not repeat mobilization. Given that the total CD34 collection and the number of days of collection are influenced by the CD34 goal, we examined patient characteristics that correlated with the CD34 collections over the first 2 days. Increasing patient age and the duration of lenalidomide therapy, both correlated with decreasing 2-day CD34 collection, while the time between last dose of lenalidomide and the start of GCSF had no effect. We then performed ROC analysis to find best cut-off points that predicted the inability to collect adequate (2.5 million) CD34 cells in 3 days. Lenalidomide therapy of more than 4 months (P =0.03) and age > 63 yrs (P = 0.04) best predicted inability to achieve this endpoint. In addition, a peripheral blood CD34 count < 5/uL on day 5 after start of GCSF was highly predictive of failure to reach this endpoint. Conclusions: Inability to collect adequate stem cells with lenalidomide appears to be related to patient age and the duration of lenalidomide therapy. We recommend early stem cell collection and storage, if a delayed transplantation approach is taken. Patients receiving more than 4 cycles of therapy and those over 65 years should undergo mobilization with CTX+G-CSF, rather than G-CSF alone. Majority of the patients who fail G-CSF based collection can be mobilized using CTX and G-CSF. Early identification of failures after G-CSF administration using the peripheral CD34 counts can potentially allow salvage using strategies such as AMD3100.

Abstract Abstract 636 Cellular vaccine may offer additional benefit as consolidative therapy after autologous stem cell transplant (ASCT) for multiple myeloma (MM). We have reported previously from two phase II studies the survival of MM patients who received APC8020 (autologous antigen presenting cell vaccine) after ASCT (Lacy, AJH '09, (84): 799; Lin, ASH annual meeting 2010, abstract 1958). To determine if there was an outcome advantage for additional consolidative vaccine therapy after ASCT, we performed this matched case control analysis comparing the survival of patients who received ASCT only (control) at Mayo Clinic to those who received APC8020 after ASCT in the same time period. Forty-three study patients accrued between July 1998 and March 2002 were eligible for matching. To reduce bias from disease risks and transplant-related mortality, control patients were required to be alive and progression free six months after transplant and have a M-spike value of < 2g/dL at 4 mos (1st study) or 9 mos (2nd study) after transplant, similar to post-ASCT patients who were enrolled in the two APC8020 studies. A total of 111 control patients were eligible for matching. Control patients were matched to study patients by age (+/− 5 yrs), ISS (International Staging System) stage, and plasma cell labeling index (PLCI, either >= 3 or < 3). A matched control was not identified for one study patient. Of the remaining 42 study patients, all had at least one matched control patient. Thirty-two patients had 2 matched controls. Cox proportional hazards analysis showed that there was no statistically significant difference in time to progression between the study and control patients (HR 1.54, 95% CI 0.92 – 2.58, p = 0.1). However the overall survival is significantly improved among the study patients (HR 0.54, 95% CI 0.30 – 0.99, p = 0.05). This analysis is the only known report of long-term follow-up of the largest group of MM patients treated with APC8020 vaccine. While many current prognostic tests were not available at the time of the initial vaccine studies, the control patients in this analysis were matched by known disease risks that are still in use today. The survival advantage in study patients remained statistically significant compared to matched controls. Interestingly, the time to progression is unchanged between study and control patients. This is similar to other reports of vaccine therapy such as in sipuleucel-T (Provenge™) in prostate cancer where an overall survival benefit is seen without change in progression-free survival. A larger, randomized phase III study is needed to confirm the encouraging survival benefit of this treatment modality in MM. Further insight into the mechanisms of action of APC8020 is needed to improve our understanding of the role of this therapy in the context of current practice of MM maintenance therapies. Disclosures: Sims: Dendreon Corporation: Employment.

Abstract Abstract 2887 Introduction: Cardiac involvement is the major cause of death in patients with immunoglobulin light chain amyloidosis (AL). Detection of cardiac involvement and risk stratification has been facilitated by cardiac biomarkers like troponin T (cTnT) and N-terminal brain natriuretic peptide (NT-proBNP). A novel high sensitivity cTnT (hs-cTnT) assay has been developed, and we evaluated its diagnostic use with three questions in mind: 1) How do the cTnT and hs-cTnT perform in the AL amyloid staging system? 2) Does higher sensitivity add significant additional value in terms of prognosticating outcomes for patients with AL amyloidosis? 3) Can the current AL amyloidosis staging system be further improved upon? Methods: Stored serum samples (-20°C) from 224 pts with AL were analyzed for concentrations of hsTnT, TnT, and NT-proBNP on the E170 Modular analyzer (Roche Diagnostics, Penzberg, Germany). 99th percentile reference limits were <0.014 and <0.010 mcg/L for hsTnT and TnT, respectively. Results: Median values for hsTnT, TnT, and NT-proBNP were 38 ng/L (range 0–075.4), 0.017 mcg/L (<0.0–0.904), and 1230 ng/L (0–32, 226), respectively. The correlation coefficient between hsTnT and TnT was 0.972. Those classified by echocardiographic parameters as having (n=143) or not having (n=81) cardiac involvement had TnT concentrations of 0.04 and 0.01 mcg/L and hsTnT levels of 52.2 and 15.6 ng/L, respectively. The direct numeric result from the hs-cTnT result CANNOT merely be substituted for a cTnT result in the Mayo AL staging system since 14% of patients would be misclassified. The performance of the receiver operation curve derived hs-cTnT cut-point of 54 ng/L is a slight improvement over the direct substitution of 35 ng/L if replacement of one assay for another is required. An alternate staging option using hs-cTnT alone—using the two thresholds14 ng/L and 54 ng/L (figure)—performs as well as either the original Mayo AL staging system or a derivative system incorporating hs-cTnT with respective relative risks of death (95%CI) of 3.6 (2.3, 5.7), 3.8 (2.5, 5.9), and 3.3 (2.2, .50). On multivariate analysis, our newly described alternate 3 level, hs-cTnT alone staging system is independent of other factors including period of diagnosis, type of therapy, and NT-proBNP value, the last of which dropped out of the model. Alternate models using NT-proBNP and cTnT were explored, but none performed better than the original staging system or the new hs-cTnT system. Conclusion: The direct numeric result from the hs-cTnT result cannot merely be substituted for a cTnT result in the Mayo amyloid staging system. Consideration could be made for AL staging systems using hs-cTnT alone and relegating the NT-proBNP for measuring cardiac response. Disclosures: Jaffe: Roche: Consultancy.

Abstract Background Carfilzomib is a proteasome inhibitor that irreversibly binds its target and has been validated as both relapse and upfront therapies for multiple myeloma. To optimize pre transplant response, we combined carfilzomib with the widely accessible backbone of cyclophosphamide-thalidomide-dexamethasone (CTD). We previously reported results of the Phase I component of the trial (in which no MTD was reached); we now report results for the MTD and fully accrued expansion cohorts at the MTD of the CYCLONE trial NCT01057225. Methods Newly diagnosed myeloma patients intended for stem cell transplant were eligible. All patients were treated on a 28 day cycle with carfilzomib IV Days 1,2,8,9,15,16 (see Table below for dosing per cohort) along with cyclophosphamide 300 mg/m2 PO Days 1,8,15, thalidomide 100 mg PO Days 1-28 and dexamethasone 40 mg PO Days 1,8,15,22. The initial phase I/ II regimen is shown below – as no DLTs were observed, we increased dosing of carfilzomib to 36 and then 45mg/m2 and have now nearly fully accrued to the Phase II dose level +1 (27mg/m2) and expansion MTD (36mg/m2). Treatment was for 4 cycles with planned SCT post induction. The primary endpoint of the trial is the proportion of patients who have ≥very good partial response (VGPR) to treatment. All patients received herpes zoster prophylaxis and ASA daily. Results A total of 54 patients have been accrued to the trial, 6 in the initial Phase 1, 21 in the initial Phase II (CFZ 27mg/m2), and 27 at CFZ 36-45mg/m2. Median age was 63 (range 27-77) and 52% were male. ISS Stage was advanced (II-III) in 57% The overall partial response rate or better following 4 cycles of CYCLONE at dose level 0 or greater is 91%: CR 18%, VGPR 58%, PR 16%, MR 2%, SD 4%. One patient withdrew from the trial after cycle 1. At dose level 2, 3/7 patients experienced dose-limiting toxicity in cycle 1 including: grade 3 alanine aminotransferase increase (possibly related); grade 3 infusion reaction (probably related); and grade 4 heart failure with grade 3 dyspnea, atrial fibrillation, and fatigue (all possibly related). No dose-limiting toxicity was observed in 6 patients at dose level 1 and thus DL1 was deemed the MTD. Across dose levels, adverse events are fully evaluable in 48 patients. AEs of grade 3 or higher at least possibly related to CYCLONE occurred in 26 (54%) – 35% non hematological and 33% hematological. Most commonly reported non hematological toxicities (all grades) included fatigue (63%), constipation (46%), elevated creatinine (27%), hyperglycemia (27%), lethargy (25%) peripheral sensory neuropathy (25% - all grade 1), and somnolence (21%); however, grade 3/4 toxicities occurring in >5% were uncommon: hypertension (8%), thromboembolic event (6%) and hyperglycemia (6%). Three cases of pneumonia required hospitalization. There were no cardiac events seen in greater than 5% of patients. Grade 3/4 hematological toxicities included lymphopenia (23%) and neutropenia (17%). All patients advancing to SCT successfully collected stem cells. One patient died on study from pneumonia. 90% completed at least 4 cycles. With median follow-up of 16 months (range 0-64), 1-year PFS was 90% and 1-year OS was 98%. Conclusion The 4 drug CYCLONE regimen is highly efficacious with an overall response rate after only 4 cycles of 91% (76% ≥VGPR) at the dosing level of carfilzomib IV 20/27-36 mg/m2 in newly diagnosed myeloma. Toxicities are manageable, with only grade 1 neuropathy and minimal cardiac or pulmonary toxicity. CYCLONE is a combination that results in rapid and deep responses with minimal neuropathy prior to stem cell transplant. Comparative studies of this regimen with longer duration of therapy (as induction or consolidation) at the defined MTD should be considered. Disclosures: Mikhael: Celgene: Research Funding; Onyx: Research Funding; Sanofi: Research Funding. Reeder:Millenium: Research Funding; Celgene: Research Funding; Novartis: Research Funding. Costa:Otsuka: Research Funding. Stewart:Celgene: Honoraria; Millenium: Honoraria, Research Funding; Onyx: Research Funding.

Abstract 5079 Introduction: IgD monoclonal proteins are rare. They are not seen as a MGUS and are present in 1% to 2% of patients with myeloma. In light chain amyloidosis (AL), IgD monoclonal proteins are rare. When an IgD protein is found, amyloidosis is often omitted from the differential diagnosis. An IgD protein in amyloidosis has been reported in single cases but never as a patient series. We report IgD AL in 53. Patients & Methods: Clinical and demographic data for patients were retrieved from the patient records. Factors of interest were compared between patients who did and did not have an IgD protein. Results: Among 3,955 patients with AL amyloidosis seen, 53 patients (1.3%) had a serum IgD monoclonal protein. (Table 1) A serum monoclonal protein peak was visible on SPE in only 14, and only 5 had an M spike greater than 1 g/dL. On immunofixation of the serum, the IgD light chain was k in 11, λ in 35, and uncertain in 6; 1 patient had a biclonal D λ and G k protein. A urine monoclonal protein was detected in 43 of 51 patients; urinary immunofixation detected a λ light chain in 33 and a k light chain in 10. Biopsy of tissues showed amyloid deposits in the bone marrow in 47% and in the fat aspirate in 73%. Six patients (age 47–63 years) underwent autologous SCT. Four had renal & two had cardiac AL. All 6 had hematologic PR, 4 CR, and 4 had organ response. One patient had relapse of disease and is now on dialysis, and another had relapse and is alive with salvage chemotherapy. One patient had disease relapse and died of progressive GI amyloidosis at 22 months. The other 5 are alive at a median of 68 months (range, 7.5–83.5 months). We compared the 53 patients with IgD amyloidosis with 144 patients with non-IgD amyloidosis. (Table 2) Findings that were significantly different between groups included a lower frequency of renal amyloidosis (P=.005) and a lower prevalence of cardiac amyloidosis (P=.047). There was a higher serum albumin level (P=.04) related to the lower level of proteinuria. No difference in survival was seen between the groups. (Figure). Variables that might affect survival—liver size, performance status, septal thickness, serum creatinine level, and β2-microglobulin level—were not different between groups. Conclusions: IgD AL patients have a lower frequency of renal involvement and possibly also of cardiac involvement. The overall survival of these patients does not appear to be different from that of patients who have AL associated with another monoclonal protein. IgD monoclonal proteins are so closely linked to the diagnosis of multiple myeloma in the mind of a clinician that the possibility of amyloidosis may be overlooked. Disclosures: No relevant conflicts of interest to declare.

Abstract Background Multiple myeloma (MM) is the second most common hematologic malignancy in the U.S. and is characterized by increased immunoglobulin production and infiltration of the bone marrow by malignant plasma cells. Myeloma cell growth is supported by both the elements present in the bone marrow microenvironment as well as deregulation of internal cellular systems associated with proliferation and apoptosis. Defective programmed cell death by BCL2 or MCL1 upregulation is observed in >80% of myeloma cases and is associated with an aggressive clinical course. Remarkably, there is no approved drug with the ability to target BCL2 or MCL1. Various strategies to mitigate the effects of increased BCL2 functionality, via small molecule inhibitors or BCL2-specific antisense oligonucleotides have been previously examined in MM. These studies revealed that 1) apart from BCL2, additional anti-apoptotic members belonging to the BCL2 family (i.e. MCL1, BCL-xL) play integral roles in maintenance of the myeloma cell phenotype, 2) drugs, which target only BCL2 offer limited therapeutic advantage and 3) given that survival factors engaged in MM are multifactorial, perhaps isolated targeting of the BCL2 pathways may not be sufficient. We therefore hypothesized that the clinical development of drugs that target various antiapoptotic BCL2 family members will require a partnership with established anti-MM regimens whose activity can potentially be further enhanced with correction of the apoptotic response system. Aim Since lenalidomide and dexamethasone (LD) is an established therapy for plasma cell cancers, we investigated if the pan-BCL2 inhibitor AT-101 (BH3 mimetic), which has high binding specificity for BCL2, MCL1 and BCL-xL, can be an effective therapeutic partner to enhance anti-MM effects of LD. Methods Human MM (KMS11, U266, OPM2) and WM (BCWM.1) cell lines along with their corresponding bortezomib resistant (BR) clones, (KMS11/BR, U266/BR, OPM2/BR and BCWM.1/BR) which were developed in our laboratory were used in this study. For gene expression profiling, the Illumina HumanHT-12 v3 whole-genome gene expression array and Nanostring nCounter mRNA quantification assays (NanoString, Seattle, WA) were utilized. Statistical analysis was conducted using R-based packages and the MeV software (TIGR). Apoptosis was measured by annexin-v/PI staining, and mitochondrial membrane permeability (MOMP) was assessed using TMRM followed by flow cytometry. Protein profiles were ascertained by western blot. Results Gene expression and immunoblot analysis of six plasma cell cancer models showed upregulation of various BCL2 family members, notably MCL1 and BCL-xL. In a dose-dependent manner, AT-101 was able to downregulate BCL2 and MCL1 and induced apoptotic cell death in MM and WM cells in a dose dependent manner. Tumor cell death was associated with caspase and PARP-1 cleavage accompanied by an increase in MOMP. This cytotoxic effect and BCL2 downregulation were further potentiated when AT-101 was combined with lenalidomide/dexamethasone (LDA). Nanostring nCounter mRNA quantification and IPA analysis revealed differential changes in the CCNA, FRZB, FYN, IRF1, PTPN11 genes in LDA treated cells. IPA canonical pathway analysis demonstrated the p53 signaling and the cyclins and cell cycle regulation pathways to be the most significantly activated by LDA therapy. Summary In summary, we describe for the first time the cellular and molecular events associated with the use of AT-101 in combination with lenalidomide/dexamethasone in preclinical models of plasma cell malignancy. This study lays the rationale for engaging and handicapping the intrinsic apoptotic system through the pan-BCL2 targeting capabilities of AT-101; thus enhancing the anti-MM effects of lenalidomide/dexamethasone. Disclosures: Foran: Celgene: Research Funding.

Abstract Abstract 2868 Introduction: Improvement in renal parameters is fundamental for the assessment of organ response to treatment in immunoglobulin light chain amyloidosis (AL). Renal response, defined as a >50% reduction in proteinuria with <25% decrease in renal function, is associated with superior overall survival (OS); however, a systematic study of the interplay of proteinuria, creatinine, and free light chain has not been performed. The intent of this study is to better understand the components of the renal response criteria. Methods: Patients who underwent autologous stem cell transplantation from 1995 to 2010 were eligible for this retrospective analysis if they had > 1 year of follow-up and a baseline 24 hour urine protein > 1 g/d. Those who were dialysis dependent or died before 1 year were excluded as not evaluable for renal response. Results: Of the 435 patients screened, 152 met criteria. On univariate analysis, >50% proteinuria reduction (p < 0.001), serum free light chain (sFLC) reduction by >50% (p = 0.002), <25% increase in serum creatinine (Scr) (p = 0.0003), and a 50% proteinuria reduction within 15 months of treatment (p = 0.04) were associated with better OS. Multivariate analysis using a proportional hazard model showed that sFLC reduction by >50% (0.006), <25% increase in serum creatinine (Scr) (p = 0.0005) and >50% proteinuria reduction (p = 0.0006) were independently associated with OS. In addition, OS was significantly better among patients achieving >85% reduction in proteinuria than those with < 85% but >50%, p = 0.03 (Figure 1). Patients with >75% reduction in sFLC were more likely to achieve >50% proteinuria reduction (p = 0.009). Most importantly, among those with > 25% increase in Scr, only those patients who had <85% proteinuria reduction had a worse OS (Figure 2a, p = 0.006). There was no difference in OS among those with >85% proteinuria reduction and a rise in Scr of > 25% (Figure 2b, p = 0.77). Conclusion: Renal response is an excellent prognostic marker in patients with AL. This is the first study to show the depth of renal response correlates with OS. Patients who had the highest levels of proteinuria reduction (> 85%) had a significantly better OS than those with >50% but <85%. We also found that the negative impact of >25% decrease in renal function was only pertinent in those who did not have >85% reduction in proteinuria. These findings should help better refine the definition of renal response criteria in AL patients. Disclosures: Leung: Binding Site: Honoraria.

Abstract Abstract 3101 The POEMS (polyneuropathy, organomegaly, endocrinopathy, M-protein, skin changes) syndrome is a rare paraneoplastic disorder secondary to a clonal plasma cell process. The most experience with successful outcomes in POEMS with disseminated disease has been with high dose chemotherapy followed by autologous stem cell transplantation (ASCT). However, the risk of relapse of POEMS following an ASCT or the outcome of the patients who relapse is not known. We therefore performed a retrospective review of POEMS patients who underwent ASCT at our center and subsequently relapsed or progressed. Over the past 12 years, 58 patients with POEMS underwent autologous stem cell transplantation at the Mayo Clinic, Rochester between Mar 18, 1999 and Jun 22, 2011. With a median follow up of 40 months (1–147 months), 12 patients have relapsed. The median time to relapse was 41 months. The 1, 2, and 5 year PFS rates were 98%, 94%, and 75%, respectively (Figure 1), and the 1, 2, and 5 year OS rates were 98%, 98%, and 94%. Of the 12 patients with documented relapse, 7 had baseline PET/CT scans all of which had FDG avid lesions. By 1 year, the FDG avidity had not resolved in 4 of these patients who then went on to either develop additional lesions or increasing SUV of existing lesions. Of the 12 patients with documented relapse, 8 had baseline VEGF measurements, and all but 1 had normalization by 1 year. The first sign of relapse was a clinical deterioration in 2, a rising M-protein in 3, and a rising plasma VEGF level in 5 patients. All of these patients had new or worsening of existing lesions on PET/CT scan (n= 9) or x-ray imaging (n= 3). The symptoms of clinical relapse included pain at the site of the lytic lesion and generalized pain with swelling. Of the patients who relapsed, 5 were treated with radiation therapy, 4 received lenalidomide or thalidomide, 2 had alkylator based therapy and 1 had bortezomib based therapy. One patient was treated with surgical resection of his hepatic plasmacytoma and 2 patients have been observed without needing definitive treatment as of yet. One patient also received anti-VEGF treatment with bevacizumab unsuccessfully after failing cyclophosphamide therapy. In this large cohort of POEMS patients treated with high dose chemotherapy and autologous stem cell transplantation in a single center, a relapse rate of 21% was noted. Not every patient who relapsed has been deemed to need further treatment, and most patients could be salvaged with chemotherapy or radiation. Only one patient died from relapsed POEMS following stem cell transplantation. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 806FN2 POEMS syndrome is defined by the presence of peripheral neuropathy (P), monoclonal plasma cell disorder (M), organomegaly (O), endocrinopathy (E) and skin changes (S). The majority of patients have sclerotic bone lesions, although the underlying mechanisms are still unclear. Increased circulating vascular endothelial growth factor (VEGF) is another feature of the disease. However, there is almost no information on the role of other angiogenic molecules in POEMS biology. The aim of the study was to evaluate bone metabolism and angiogenic cytokines in POEMS and compare the results with multiple myeloma (MM) and osteosclerotic patients of other etiology. We retrospectively studied 61 POEMS patients (40M/21F, median age 52 years) who were diagnosed between 1996 and 2010 and were treated and followed in Mayo Clinic Rochester (MN, USA), Hôpital Saint-Louis, Paris (France) and in Alexandra Hospital, Athens (Greece). We evaluated the following indices of bone remodeling and angiogenesis in all patients at diagnosis and in 22 of them one month after front-line treatment: i) osteoclast regulators [sRANKL and osteoprotegerin (OPG)]; ii) osteoblast inhibitor dickkopf-1 (Dkk-1); iii) bone resorption marker CTX; iv) bone formation markers [bone-specific alkaline phosphatase (bALP) and osteocalcin]; and v) angiogenic cytokines [VEGF, angiogenin (ang), angiopoietin (angp)-1 and -2]. These molecules were also measured in 60 newly diagnosed, untreated MM patients (34M/26F, median age: 54 years) and 44 healthy controls. Bone markers were also evaluated in 24 patients with HbS/beta-thalassemia (10M/14F, median age: 43 years) who presented with osteosclerosis, defined as osteosclerotic bone lesions in plain radiography and high T-score of lumbar spine DXA (median: +3.6, range: +2 to +7.9). All POEMS patients presented with P and M, while 90% had E, 83% had S, 61% had O and 32% had papilloedema. One or more documented sclerotic bone lesions in plain X-rays were observed in 78% of patients (diffuse lesions in 44%), while 43% of patients had also an osteolytic component. Seven patients had Castleman disease. At diagnosis, compared to controls, POEMS patients had increased levels of bALP (mean ± SD: 48.3 ± 22.9 IU/L vs. 21.6 ± 8.2 IU/L; p<0.001), CTX (0.97 ± 0.94 ng/ml vs. 0.33 ± 0.24 ng/ml; p<0.001), Dkk-1 (p<0.001), sRANKL (p<0.001) and sRANKL/OPG ratio (p<0.001). Patients with diffuse osteosclerosis had increased levels of Dkk-1 compared to others (p = 0.05), suggesting a balance effect to increased formation activity. Patients with POEMS had increased levels of bALP and decreased levels of CTX compared to MM patients (p<0.001 and p = 0.03, respectively), while they had increased levels of CTX and sRANKL/OPG ratio compared to HbS/beta-thal patients (p = 0.001 for both comparisons). There were no other differences regarding bone markers among studied groups. Regarding angiogenic cytokines, at diagnosis, POEMS patients had elevated VEGF (1232 ± 565 pg/ml vs. 195 ± 211 pg/ml; p<0.001), ang (266 ± 82 ng/ml vs. 190 ± 45; p<0.001) and angp-2 (3.5 ± 2.2 ng/ml vs. 1.2 ± 0.6 ng/ml; p<0.001) and decreased angp-1 (p =0.05) and angp-1/angp-2 ratio (reflects increased angiogenic capacity; 9.1 ± 8.8 vs. 44.8 ± 76.1; p<0.001) compared to normal controls. We found no strong correlations between the levels of VEGF, ang and angps. When compared to MM patients, POEMS patients had increased circulating VEGF (p = 0.001) without any difference regarding the other angiogenic cytokines. Response to therapy was accompanied by a dramatic reduction of VEGF only (551 ± 541 pg/ml compared to baseline 1800 ± 1510 pg/ml; p = 0.003). Receiver-operating characteristic (ROC) curve analysis showed that high levels of angiopoietin-2 (≥ 2.3 ng/ml) and high bALP (≥ 38 IU/l) each have a 95% specificity for patients with POEMS syndrome. These results suggest that patients with POEMS syndrome have high bone turnover with increased bone formation as compared to healthy subjects, to MM patients, who have reduced bone formation, and to osteosclerotic HbS/beta-thal patients, who have diminished bone resorption. Importantly, our analysis supports the use of angp-2, bALP along with VEGF as minor criteria for the diagnosis of POEMS. Successful treatment reduced VEGF, which seems to be a suitable marker for the follow-up of POEMS patients. Furthermore the imbalance of angp-1/angp-2 pathway may indicate possible targets for the development of novel agents against POEMS. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 3095 Background: Although HDM is the standard conditioning for pts with MM undergoing ASCT, it is not sufficient for cure. In an effort to improve upon this standard, we added escalating doses of Y-90 Zevalin (a murine IgG1 kappa monoclonal antibody Ibritumomab conjugated to Y-90) to HDM in a standard 3+3 phase I trial design. Y-90 is a beta-emitting radioisotope with a path length of 5 mm allowing it to target CD20+ cells and bystander cells within 5 mm. Methods: Eligibility included being a candidate for ASCT. Consenting patients (pts) received: day -22, rituximab 250 mg/m2 with In2B8 for scanning; day -14, rituximab 250 mg/m2 with escalating doses of Y90-Zevalin; day -2 and -1, melphalan 100 mg/m2/day; day 0, at least 2 × 10(6) CD34/kg stem cells; and GM-CSF until ANC engraftment. The 6 dose levels (DL) were (in Gy): 10; 12; 14; 16; 18; and 20. Dose limiting toxicity (DLT) was defined as: sustained pulmonary or liver toxicity (grade3; 96 hours duration or VOD that does not resove by day 30); delayed engraftment (ANC < 500 at day 21 or platelet transfusion dependence at day 35); and non-hematologic grade (gr) 4 toxicity excluding gastrointestinal or constitutional. Responses were according to IMWG criteria. Accrual was between May 2005 and June 2011. Data was frozen as of August 2, 2011. Results: Thirty pts have been accrued: DL1, 3; DL2, 3; DL3, 3; DL4, 6, DL5, 6; DL6, 6. The median age of patients was 59 (range 32, 73), and 70% were male. 60% had ASCT as primary therapy. The other patients were transplanted in chemosensitive relapse (n=3), chemoresistant relapse (n=7), and untested relapse (n=2). Median time from diagnosis to protocol therapy was 10.4 months (range 3.5, 94.3). Five pts had prior radiation therapy, and 11 had prior ASCT. Maximum tolerated dose (MTD) of Zevalin with fixed dose melphalan 200 mg/m2 was DL 5 (18 Gy) which amounted to 168 mCi (124, 182) administered. No DLTs were observed at DL 1, 2, 3, or 5. There was 1 DLT at DL 4, comprising hepatic failure (gr 5) CMV viremia (gr 3), and delayed engraftment, in a 73 yo patient with chemoresistant relapsed disease, and 2 DLTs at DL 6. These DL 6 DLTs included: 1) fatal bowel necrosis (day 12 post ASCT) in a 67 yo patient with relapsed refractory MM and history of CAD, ischemic colitis, DM; and 2) fatal biopsy proven venoocclusive disease (gr 5), which began day 91 post ASCT in a 62 yo patient with primary refractory MM, peri-mobilization S. aureus bacteremia with septic embolic, and line associated venous thrombosis. Overall, gr 3 or higher AEs were observed in 100% of pts. Non-heme gr 4 AEs were seen in: DL1, 1; DL2, 2; DL3, 0; DL4, 1; DL5, 0; DL6, 4. Median time to ANC >=500 and plts>=50 was 11 (range 8, 17) and 15 days (range 11, 89). Confirmed response rates from immediately pre-protocol to best confirmed response were: CR, 5; VGPR, 5; continued VGPR, 4, and continued PR, 7. If one includes best confirmed response from baseline for those patients using transplant as primary consolidation, 4 PRs are upstaged to VGPRs. Median follow-up for all pts is 24.5 months (0.8, 64.8), and 10 pts have died. The median PFS for pts proceeding to early and delayed ASCT was 30.4 months (95%CI 4.1-NR) and 10.3 months (95%CI 2.8–21.7), respectively. The median OS for pts proceeding to early and delayed ASCT was not attained and 11.4 months (95%CI 2.8-NA), respectively. Conclusions: For patients with multiple myeloma, the MTD of Zevalin with fixed high-dose melphalan is 18 Gy. The combination of a VGPR plus continued VGPR rate of 60%, a median PFS of 13.1 months, and a 73% 1 year OS rate make this a regimen worthy of further study. Disclosures: Kumar: Merck: Consultancy, Membership on an entity's Board of Directors or advisory committees.