Leung, Nelson

Abstract Abstract 3572 High dose melphalan and autologous stem cell transplantation (ASCT) is an effective treatment for light chain amyloidosis (AL) but high treatment related mortality (TRM) limits its use. Both cardiac troponin T (cTnT) and N-terminal-pro-brain natriuretic peptide (NT-pro-BNP) are sensitive predictors of high risk patients. Aim of this study is to compare two models of TRM, one using cTnT and the other BNP. Patients who underwent ASCT between 7/1996 and 7/2009 at the Mayo Clinic were analyzed retrospectively. Variables were chosen for ease of measurement and reproducibility thus echocardiographic parameters were excluded. Each additional risk factor must contribute to a significant increase in TRM. Also, since NT-pro-BNP and cTnT are already highly associated with mortality, the other risk factors must be independent predictors. Univariate analysis revealed serum albumin, B-2-microglobulin, cTnT, NT-pro-BNP, and uric acid were associated with TRM while age, sex, alkaline phosphatase, serum creatinine, CRP, proteinuria were not (Table 1). Multivariate analysis showed albumin and uric acid were independent risk factors to cTnT and NT-pro-BNP (Table 2). Of the 412 patients, 347 (8.9% TRM) could be evaluated by the cTnT, albumin and uric acid (TAU) model and 282 (9.2%) TRM could be evaluated by the NT-pro-BNP, albumin, uric acid (BAU) model. Both models showed increasing TRM with additional risk factors beyond 1 risk factor (Table 3 and 4). Albumin and uric acid were found to be significant contributors in both models. Patients with elevated cTnT alone had a 0% TRM vs 23.4% if they had 1 additional risk factor, p = 0.03. Similarly, TRM was 3.7% in patients with elevated NT-pro-BNP alone but 18.0% if they had 1 other risk factor, p = 0.05. NT-pro-BNP and cTnT could not be used in the same model because cTnT becomes non-significant. Both models were superior to visceral organ involvement which had TRM of 5.3% with 1, 9.2% with 2 and 16.7% with 3 organ involvement. Both models were also predictive of overall survival (Figure 1 and Figure 1b). Our results show that these models show that albumin and uric acid with either cTnT or NT-pro-BNP can accurately predict TRM in AL patients undergoing ASCT. Multivariate analysis shows that the NT-pro-BNP containing model may be superior since cTnT loses it significance when added to the model. Our results suggest these models could be very helpful in identifying high risk patients not suitable for ASCT. Figure 1a Overall survival evaluated by Kaplan-Meier method on 347 patients using the TAU model. Mortality increased significantly with the number of risk factors, p < 0.001. Solid line represents patients with 0 risk factors, dotted line represents patients with 1 risk factor, dashed line represents those with 2 risk factors and dotted dashed line represent those with 3 risk factors. Figure 1a. Overall survival evaluated by Kaplan-Meier method on 347 patients using the TAU model. Mortality increased significantly with the number of risk factors, p < 0.001. Solid line represents patients with 0 risk factors, dotted line represents patients with 1 risk factor, dashed line represents those with 2 risk factors and dotted dashed line represent those with 3 risk factors. Figure 1b Overall survival evaluated by Kaplan-Meier method on 282 patients using the BAU model. Overall survival was significantly longer in patients with less risk factors, p < 0.001. Solid line represents patients with 0 risk factors, dotted line represents patients with 1 risk factor, dashed line represents those with 2 risk factors and dotted dashed line represent those with 3 risk factors. Figure 1b. Overall survival evaluated by Kaplan-Meier method on 282 patients using the BAU model. Overall survival was significantly longer in patients with less risk factors, p < 0.001. Solid line represents patients with 0 risk factors, dotted line represents patients with 1 risk factor, dashed line represents those with 2 risk factors and dotted dashed line represent those with 3 risk factors. Table 1. Patient characteristics Control TRM p-value N 318 29 Age 58 (25–73) 55 (35–75) 0.54 Sex (male) 58.2% 72.4% 0.13 Albumin (g/dl) 2.7 (0.8–4.4) 2.2 (0.8–3.7) 0.02 Alkaline phosphatase (U/L) 89 (28–1014) 86 (31–1394) 0.22 b2m (mg/ml) 2.6 (1.0–35.1) 3.2 (1.3–15.2) 0.007 NT-pro-BNP (pg/ml) 564 (12–35000) 1661 (21–35001) 0.03 CRP (mg/L) 0.3 (0.01–16.6) 0.3 (0.05–3.2) 0.34 cTnT (ng/ml) 0.01 (0–1.0) 0.03 (0.01–0.22) <0.001 Scr (mg/dl) 1.1 (0.4–12) 1.2 (0.8–2.3) 0.09 Proteinuria (g/d) 3.4 (0.01–35.4) 5.1 (0.1–20.5) 0.10 Uric acid (mg/dl) 6.2 (1–14.8) 7.5 (3.7–13.2) 0.02 Organ Involvement 1 94.7% 5.3% 0.04 2 90.7% 9.3% 3 83.3% 16.7% Table 2. Multivariate analysis using proportional hazard model cutoffs >Risk Ratio >p-value Model TAU cTnT 0.04 ng/ml 2.78 0.01 Albumin 3.3 g/dl 3.11 0.03 Uric acid 8.5 mg/dl 2.69 0.02 Model BAU NT-pro-BNP 1270 pg/ml 3.77 0.003 Albumin 3.3 g/dl 3.53 0.04 Uric acid 8.5 mg/dl 3.11 0.03 Table 3. Model of cTnT, albumin and uric acid TAU Risk Factors n TRM p-value Risk Ratio 0 69 2.9% 1 205 5.9% 0.33 2 61 16.4% 0.009 2.98 3 12 41.7% 0.05 8.52 Table 4. Model of NT-pro-BNP, albumin and uric acid BAU Risk Factors >n >TRM >p-value >Risk Ratio 0 28 0% 1 161 5.6% 0.20 2 82 14.6% 0.02 2.7 3 11 45.5% 0.01 10.8 Disclosures: Kumar: Celgene: Consultancy, Research Funding; Millennium: Research Funding; Merck: Consultancy, Research Funding; Novartis: Research Funding; Genzyme: Consultancy, Research Funding; Cephalon: Research Funding. Dispenzieri:Celgene: Honoraria, Research Funding; Binding Site: Honoraria. Lacy:Celgene: Research Funding.

Abstract Background: Cytogenetic evaluation, especially using fluorescence in situ hybridization (FISH), at the time of diagnosis is essential for initial risk stratification and the employment of risk-adapted treatment strategies in multiple myeloma. Little is known about the occurrence and prognostic significance of cytogenetic evolution during follow up. Methods: We studied 433 patients who were diagnosed with multiple myeloma between January 2000 and December 2011 and had at least two FISH evaluations at Mayo Clinic Rochester, including the diagnostic specimen. Bone marrow aspirates were evaluated for deletions, monosomies, trisomies, and tetrasomies using chromosome- or centromere-specific FISH probes. IGH rearrangements were evaluated using an IGH break-apart probe and up to five potential partners (FGFR3, CCND1, CCND3, MAF, and MAFB). Cytogenetic evolution was defined as a new deletion, monosomy, trisomy, tetrasomy, or translocation during follow up. Multivariable-adjusted logistic regression models were used to assess the associations between the parameters of interest and the presence of cytogenetic evolution in follow-up specimens. Multivariable-adjusted Cox proportional hazards models were used to assess the effect of cytogenetic evolution on overall survival. All models were adjusted for sex, age, the presence of high-risk FISH abnormalities, and the number of abnormalities at the time of diagnosis. Likelihood ratio tests were used to assess the goodness of fit of nested models. The χ2 or Fisher's exact test was used to assess the distribution of cytogenetic abnormalities in subgroups. Results: The median age at diagnosis was 60 years (32 - 82), 264 (61%) of the patients were male. The median overall survival for the entire cohort was 7.0 years (6.2 - 7.8). At the time of diagnosis, 150 (35%) and 57 (13%) of the 433 patients presented with a hyperdiploid karyotype and cytogenetic high-risk abnormalities, respectively. Independent of each other, the presence of a translocation at the time of diagnosis was associated with decreased odds of cytogenetic evolution during follow up (OR 0.39, 95% CI 0.24 - 0.63, p < 0.001) while the presence of at least one trisomy or tetrasomy at the time of diagnosis was associated with increased odds (OR 2.53, 95% CI 1.37 - 4.70, p = 0.003). A greater proportion of patients presenting with a hyperdiploid karyotype experienced cytogenetic evolution during follow up. Those patients more frequently evolved additional trisomies and tetrasomies, while translocations were more common in those presenting with a non-hyperdiploid karyotype (Table 1). The development of additional abnormalities during the three years following diagnosis (compared to no new abnormalities) was associated with increased subsequent mortality in those who survived at least three years (HR 3.22, 95% CI 1.82 - 5.68, p < 0.001). Including the time between first and last cytogenetic evaluation as a covariate did not significantly change the parameter estimates or improve model fit (p = 0.727). Conclusions: Demographics, risk profile, and overall survival of this cohort reflect the fact that patients had to survive long enough to undergo repeated cytogenetic evaluation. Hyperdiploid and non-hyperdiploid genotypes were associated with distinct behavior regarding cytogenetic evolution during follow up. The identification of cytogenetic evolution was an adverse prognostic factor in those who survived at least three years after diagnosis. These findings emphasize the importance of the dynamics of the underlying clonal disease process for accurate risk assessment and suggest that selected subgroups of patients may benefit from risk stratification during follow up. Table 1. Cytogenetic evolution during follow up in 433 patients with multiple myeloma stratified by karyotype at the time of diagnosis Hyperdiploid (n = 150) Non-hyperdiploid (n = 283) p New abnormality 76 (51%) 108 (38%) 0.012 New monosomy 8 (5%) 24 (8%) 0.243 New trisomy 48 (32%) 55 (19%) 0.004 New tetrasomy 37 (25%) 27 (10%) < 0.001 New deletion 17 (11%) 27 (10%) 0.557 New translocation 1 (1%) 11 (4%) 0.065 Most common new abnormality [type (percent of type in each group)] Monosomy mono(13) (75%) mono(13) (62%) Trisomy tri(11) (22%) tri(3) (22%) Tetrasomy tetra(15) (48%) tetra(15) (30%) Deletion del(17p) (79%) del(17p) (63%) Translocation t(11;14) (100%) t(11;14) (36%) Data are given as count (percent) unless denoted otherwise. Disclosures Binder: American Society of Hematology: Research Funding. Kumar:AbbVie: Research Funding; Onyx: Research Funding; Sanofi: Research Funding; Celgene, Millenium, Sanofi, Skyline, BMS, Onyx, Noxxon,: Other: Consultant, no compensation,; Skyline, Noxxon: Honoraria; Millenium/Takeda: Research Funding; Janssen: Research Funding; Celgene: Research Funding.

Abstract Abstract 3557 Introduction: Autologous stem cell transplantation for patients with amyloidosis results in high reported hematologic and organ response rates compared with conventional chemotherapy. Patients: Four-hundred and thirty-four patients have undergone transplantation between March 8, 1996, and April 13, 2010 Table 1. Clinical parameters seen in patients with amyloid are given in Table 2. The overall day-100 mortality seen in this group of patients is 44 of 434 patients (10%). Results: The most important determinant of outcome is stage defined by BNP and troponin levels. Figure 1 demonstrates the survival for patients for all 3 cardiac stages. Troponin <.035; NT-proBNP <332. Stage 1 both low, Stage 3 both elevated. The survival of all patients based on whether they achieved a complete response, a partial response (>50%), or no response was analyzed. Median survival has not been reached for the complete response group, is 107 months for the partial response group, and is 32 months for the no response group (p<0.0001). Figure 2, divides the patients at the median dFLC level of 13.5 mg/dL demonstrating that for patients with a higher level of free light chain, the median survival is 87.6 months and has not been reached for those with a lower level of free light chain. A proportional hazards model for variables that impact on survival was done in all patients, and the only relevant predictor of outcome was stage (p<0.0001). When the same analysis was performed using a landmark of 6 months, the only predictor of outcome was stage (p=0.0005). When best response was incorporated into the landmark model, it was the strongest predictor of survival. The amyloid stage predicted survival (p= 0.0005), and the best response to therapy (p <0.0001). When the same analysis was performed in the entire group of 434, stage and best response each remained significant (p<0.0001). Discussion: There is a high response rate associated with high-dose therapy that was not observed in the melphalan and prednisone era. For eligible patients who can be transplanted safely, high-dose melphalan is an effective therapy for many patients with amyloidosis. Best response and stage at diagnosis are the best predictors of overall survival. Disclosures: Lacy: Celgene: Research Funding. Dispenzieri:Celgene: Honoraria, Research Funding; Binding Site: Honoraria. Kumar:Celgene: Consultancy, Research Funding; Millennium: Research Funding; Merck: Consultancy, Research Funding; Novartis: Research Funding; Genzyme: Consultancy, Research Funding; Cephalon: Research Funding.

Abstract Background: Light chain Amyloidosis (AL) is characterized by deposition of light chain derived amyloid fibrils in major organs and/or soft tissue. An observational study on natural history and outcome of localized immunoglobulin light-chain amyloidosis without vital organ (liver, heart, kidney, peripheral and autonomic nervous system) involvement has shown extremely low rate of progression (1%) to systemic amyloidosis at a median follow-up of 74.4 months (Mahmood S et al. The Lancet Haematology; 2015; 6:e241-e250). There is, however, limited evidence in published literature on natural history of AL amyloidosis confined to fat and/or bone marrow biopsy, without involvement of vital organs or other soft tissues. Methods: We retrospectively identified patients with AL amyloidosis limited to fat and/or bone marrow aspirate in a single-institution database. Patients were evaluated for progression to systemic amyloidosis. Statistical analysis was done using JMP 10.0.0 (SAS Institute Inc.). Results: We identified 117 patients, with a median age of 70 years, who had light chain amyloidosis detected in abdominal fat aspirate and/or bone marrow biopsy, without systemic involvement. Amyloid was seen in fat only in 39%, marrow only in 54% and in both sites in 7%. The median follow up was 45.6 months (95% CI-38.1-57.7) from detection of amyloid. Of these, 64% were alive at the time of analysis. Among 117 patients, 65 were treated for a diagnosis of another plasma cell disorder made prior to or concurrent with detection of amyloid. The remaining 52 patients only had isolated fat or marrow amyloid. Among 65 patients with another diagnosis of plasma cell disorder requiring treatment, 3 progressed to systemic amyloidosis, one each with cardiac, renal and lymph node (LN) involvement detected at 32, 42 and 65 months respectively from the detection of amyloid. The first 2 patients had underlying multiple myeloma, and the third patient with LN involvement had underlying Waldenström macroglobulinemia. Among 52 patients without another diagnosis of a plasma cell disorder requiring treatment, at a median follow up of 32 months, no progression to systemic amyloidosis was observed. Median overall survival (OS) in 117 patients from detection of amyloid, using Kaplan-Meier survival estimates, was 60.2 months (95% CI-48.9-146.1). Conclusion: Our study highlights the fact that isolated amyloidosis detected in fat and/or bone marrow aspirate, in the absence of another plasma cell dyscrasias that require therapy, is unlikely to progress to systemic amyloidosis. Watchful waiting might be considered in such patients after a thorough evaluation to rule out systemic involvement. Disclosures Gertz: Smith Kline: Honoraria; Novartis: Honoraria; Onyx: Honoraria; millenium: Consultancy, Honoraria; Celgene: Honoraria. Kumar:Celgene: Consultancy, Research Funding; Abbvie: Research Funding; Millenium: Consultancy, Research Funding; Novartis: Research Funding; Onyx: Consultancy, Research Funding; Janssen: Consultancy, Research Funding; BMS: Consultancy, Research Funding.

Abstract Background: Amyloidosis is a multisystem disease with extracellular deposition of pathological insoluble beta-fibrillar proteins. Involvement of the heart is seen in more than one-half of the patients with systemic AL amyloidosis and it is the most important determinant of clinical outcome. Cardiac amyloidosis is a restrictive infiltrative cardiomyopathy in which, despite declines in stroke volume that accompany disease progression, the ejection fraction (EF) often remains preserved even in advanced stages of the disease. King et al. proposed a novel index of myocardial function, the myocardial contraction fraction (MCF), defined as the ratio of stroke volume (SV) to myocardial volume (MV). MCF is a measure of myocardial shortening, which differentiated myocardial performance in patients with similar degrees of hypertrophy and in a recent small cohort of AL amyloidosis patients appeared superior to left ventricular ejection fraction (EF) in predicting overall survival (OS). It was our goal to assess the prognostic role of MCF in a large cohort of patients with AL amyloidosis in the context of other prognostic variables. Methods: Patients seen between 4/1/1999 and 2/1/2015 were eligible for this retrospective study if they had an ECHO at the Mayo Clinic, Rochester, MN within 30 days of their AL amyloidosis diagnosis with measurements of left ventricular chamber size and wall thickness needed to calculate MCF, EF, and the presence or absence of pericardial effusion. To capture the full cohort, modeling was done first excluding Mayo (2012) staging and global averaged left ventricular longitudinal peak systolic strain (LV strain) since limited numbers of patients had these studies, 342 and 294, respectively. Thresholds of continuous variables were chosen based on receiver operator curves targeting death at 1-year. Cox proportional hazards analysis was used to identify factors that were prognostic for OS. Statistical analyses were done using JMP 9.0 (SAS, Cary, NC). Results: Among the 722 patients satisfying entry criteria,median age was 64 years (range 32-94) and 66% were male. The best cutoff for MCF was 21% (sensitivity: 74%, specificity: 60%; AUC=0.699) and distinguished two groups with different OS (median 53 vs. 9 months, P<0.0001; Figure). On univariate analysis the baseline ECHO variables predicting OS were interventricular septum thickness (IVS) >12 mm (RR 1.8, P=0.0002), EF <60% (RR 1.7), MCF ≤21% (RR 2.1), and presence of pericardial effusion (RR 2.0), all with P<0.0001. On multivariate MCF ≤21% (RR 1.6, P <0.0001), EF <60% (RR 1.4, P=0.0007) and the presence of pericardial effusion (RR 1.3, P=0.002) were independent predictors of OS. Additional modeling was done with the subset of 342 patients who were assessable for Mayo (2012) staging, (i.e. measurements of immunoglobulin free light chain, NT-proBNP, and troponin T). On univariate analysis, patients with a Mayo (2012) stage ≥ 3 had a RR 2.1, p<0.0001. On multivariate, only MCF ≤21% (RR 1.5, P =0.01), Mayo (2012) stage ≥ 3 (RR 1.5, P=0.01) and EF<60% (RR 1.7, P=0.0006) were independent prognostic determinants. Finally, we addressed LV-strain among the 294 patients in whom the study was performed. There was a significant correlation between MCF and LV strain (rho =-0.86, P<0.0001). On univariate, LV-strain less negative than -13 generated a RR of 2.3, p<0.0001. In contrast to the model containing MCF, on LV-strain multivariate, EF was no longer significant, but LV-strain less negative than -13 (RR 2.0, P=0.008) and Mayo (2012) stage ≥ 3 (RR 2.0, P=0.008) predicted independently OS. Conclusions: MCF ≤21% identified a subgroup of patients with a high mortality risk, and it was independent of the Mayo (2012) staging but highly correlated with LV strain. An advantage of MCF, which is a novel volumetric measurement of the left ventricular chamber, is that, unlike LV-strain, MCF is simple to calculate from routine ECHO measures. Figure 1. Overall survival of 722 patients according to MCF<21%. Figure 1. Overall survival of 722 patients according to MCF<21%. Disclosures Kumar: AbbVie: Research Funding; Celgene: Research Funding; Janssen: Research Funding; Skyline, Noxxon: Honoraria; Celgene, Millenium, Sanofi, Skyline, BMS, Onyx, Noxxon,: Other: Consultant, no compensation,; Sanofi: Research Funding; Millenium/Takeda: Research Funding; Onyx: Research Funding.

Abstract Background Induction with novel agents, including proteasome inhibitors and immunomodulators, prior to autologous stem cell transplantation (ASCT) is the standard of care frontline treatment of transplant-eligible patients with multiple myeloma (MM). Common novel agent-based induction regimens include three drug combinations with a backbone of bortezomib (V), thalidomide (T) and/or lenalidomide (R). There is limited information on the comparative efficacy of novel regimens used as part of initial therapy prior to an ASCT in terms of overall survival (OS). To address this question, we retrospectively analyzed 1096 consecutive patients undergoing ASCT for MM at Mayo Clinic. Methods Study patients included those who underwent first transplant within 12 months of diagnosis, did not receive more than one induction regimen, did not relapse prior to transplant and did not receive treatment for smoldering MM in the past. Baseline characteristics, response rates and survival data were extracted from an electronic medical record and statistical analysis was done using JMP 10.0.0 (SAS Institute Inc.). Choice of induction regimen was almost exclusively made by the referring physician. Results Median age at diagnosis was 60.3 years (Interquartile range 53.9-65.9). ISS staging at diagnosis was available for 49.7% of patients, of which, 34.2%, 39.3% and 26.5% of patients had ISS stage 1, 2 and 3 disease, respectively. Estimated median follow up was 65.6 months (95% CI 58.3-73.3). Baseline characteristics of patients by initial treatment have been summarized in table 1. Major subgroups by induction therapy included patients receiving Cy-Bor-d (n=193, 17.6%), Vd (n=64, 5.83%), Rd (n=253, 23.1%), VRd (n=126, 34.6%), Td (n=157, 14.3%) and VAD or dexamethasone alone as the non-novel agent comparator group (n=229, 20.9%). Median overall survival (OS) was significantly different between novel agent-based and conventional regimens (Log-rank test; p=0.0029), which were 102.7 months (95% CI, 95.2-112.2) and 78.8 months (95% CI, 67.8-92.6) respectively. Median OS with each regimen is shown in table 1. Among novel agent-based regimens, there was no significant difference in median OS, except Td, which was inferior to Rd (HR 1.62; 95% CI 1.17-2.24). Conclusion We cannot demonstrate with overall survival as an endpoint that the doublets of Vd and Rd are inferior to VRd or Cy-bor-d. There were insufficient patients treated with VTd to be included. Our study shows that the choice of novel induction regimen prior to ASCT does not affect survival (with the exception of Td being inferior to Rd), reflecting the wide array of effective salvage options. However, further prospective randomized clinical trials comparing these regimens are required to validate these findings. Table 1. Baseline characteristics and overall survival. Baseline characteristics and survival Cy-Bor-d (n=193) Vd (n=64) Rd (n=253) VRd (n=126) Td (n=157) VAD or Dex (n=229) p-value Median age 61.9 62.15 60.8 60.8 59.3 58.5 0.0041 Sex (percent males) 56.48 57.81 56.92 61.91 57.96 59.82 0.9271 ISS at diagnosis 1:24.82% 2:40.69% 3:34.48% 1:29.73% 2:24.32% 3:45.94% 1:40% 2:43.78% 3:16.22% 1:41.76% 2:34.06% 3:24.18% 1:28% 2:36% 3:36% 1:20% 2:60% 3:20% 0.0005 Median follow-up (95% CI) 20.3 (17.1-22.7) 49.5 (44.7-55.7) 59 (54.5-67) 26.9 (22.8-30.7) 126.7 (120.2-132.9) 143.4 (132.5-152.6) <0.0001 Median OS Not reached (Estimated 5 year OS rate 75%) 97.2 months (95% CI, 66.6-NR) 112 months (95% CI, 97.4-124.3) Not reached (Estimated 5-year OS rate 77%) 81.1 months (95% CI, 59.1-99.1) 78.8 months (95% CI, 67.8-92.6) 0.0019 Figure 1. Figure 1. Disclosures Kumar: Onyx: Consultancy, Research Funding; Novartis: Research Funding; Abbvie: Research Funding; Celgene: Consultancy, Research Funding; Millenium: Consultancy, Research Funding; BMS: Consultancy, Research Funding; Janssen: Consultancy, Research Funding. Gertz:Smith Kline: Honoraria; Novartis: Honoraria; Onyx: Honoraria; millenium: Consultancy, Honoraria; Celgene: Honoraria.

Abstract Background: Thrombotic microangiopathy (TMA) is a hematologic syndrome associated with endothelial injury, microvascular thrombosis, and hemolytic anemia resulting in end organ damage. TMA is rarely associated with multiple myeloma (MM) and is usually reported in the setting of allogeneic hematopoietic stem cell transplants (AHSCT), with associated graft versus host disease and calcineurin inhibitor use. We present 11 cases of TMA in MM patients not associated with AHSCT. Methods: Cases were collected from multiple centers in patients with confirmed diagnosis of MM. TMA was diagnosed by the presence of thrombocytopenia and evidence of microangiopathic hemolytic anemia. Results: Eleven patients with MM were identified to have developed TMA during the course of their disease. Two patients presented with TMA at the time of MM diagnosis. A third had been off treatment for over a year prior to presenting with a TMA which correlated with a significant increase in her free light chains. The other cases occurred during active treatment. All remaining patients received proteasome inhibitors, with three patients on bortezomib-based regimens and five patients on carfilzomib. One patient on a bortezomib-based regimen had been stable prior to developing a TMA in the setting of sepsis. In the remaining patients, the TMA developed rapidly after initiating the proteasome inhibitor. In all cases, the TMA improved after drug withdrawal. One patient was later rechallenged with bortezomib and again developed TMA. Discussion: A review of the literature reveals only nine cases of TMA with MM. Of these, 4 patients had TMA as the initial presentation of MM, and 5 patients were on treatment, most often bortezomib-based regimens. Our series patients match those in the literature closely and demonstrate that TMAs can occur in MM outside of the context of AHSCT. In roughly one third of patients, TMAs were the initial presentation of MM and improved with treatment of the MM. In others, the use of proteasome inhibitors appears to be associated with development of TMA. In these cases, the temporal correlation, resolution of TMA after medication withdrawal, and the instance of recurrence of TMA with rechallenge of bortezomib, support a causal relationship between the medication and the TMA. Thus, a high clinical suspicion for TMA must exist for patients on proteasome inhibitors who develop worsening anemia and thrombocytopenia. Medication should be withdrawn if TMA is diagnosed. Disclosures Vij: Onyx: Consultancy, Honoraria, Research Funding, Speakers Bureau; Millennium: Honoraria, Speakers Bureau; Celgene: Consultancy, Honoraria, Research Funding, Speakers Bureau; BMS: Consultancy; Takeda: Consultancy, Research Funding; Novartis: Consultancy; Sanofi: Consultancy; Janssen: Consultancy; Merck: Consultancy.

Abstract Abstract 2797 Poster Board II-773 Background: Primary systemic amyloidosis (AL) is characterized by deposition of light chain derived amyloid fibrils in multiple organs leading to varying degree of dysfunction. Cardiac involvement is a common feature of AL, and when significant is highly predictive of poor outcome. The currently used prognostic classification is based on cardiac biomarkers troponin-T (cTnT) and N-terminal pro B-type Natriuretic peptide (NT-ProBNP) and is effective at predicting outcome in patients with AL. However, this is driven primarily by the degree of cardiac involvement and does not take into account the degree of plasma cell burden and its impact on the disease course. It is possible that some of the heterogeneity in outcome can be explained by the plasma cell characteristics rather than the degree of the end organ damage. Methods: We examined the baseline clinical and laboratory data from 2119 patients with AL who were seen at our institution with in 90 days of their diagnosis. The data were collected from a prospectively maintained data base, and additional testing was done for free light chain levels (FLC) on stored sera from patients seen before the routine introduction of FLC testing. Cox proportional hazards analysis was performed to estimate the prognostic values of different variables. Results: We first examined the impact of plasma cell clone related characteristics namely, difference between involved and uninvolved FLC (FLC-Diff), bone marrow plasma cell (BMPC) %, plasma cell labeling index (PCLI), beta 2 microglobulin (B2M), and presence of circulating plasma cells, on overall survival (OS) from diagnosis. All variables were dichotomized into high and low based on their median value (FLC-diff: 20 mg/dL, BMPC: 10%, PCLI as a continuous variable, B2M: 3 mg/dL and circulating cells: absent or present). While all were found to be prognostic for OS in univariate analysis; in a multivariate analysis incorporating a stepwise selection only B2M and FLC-diff were significant. We then used these two variables along with the cTnT and NT-ProBNP that are used in the current model to develop a staging system. The median values for all four variables were used for developing the scores. Patients were assigned a score of 1 for presence of each characteristic (FLC difference > 20 mg/dL, troponin-T > 0.02, NT-ProBNP > 1000, and B2M > 3) or 0 if the value was at or below the cutoff. The scores were added to obtain a composite prognostic score that grouped the 370 patients (who had all the variables available for analysis) into 5 groups with very different OS (4.4, 9.3, 24.4, 43, and not reached for stages 4, 3, 2, 1, and 0 respectively; Figure A). However, since the NT-ProBNP did not have independent impact on survival in a multivariable model incorporating all the four variables, we also examined a system that only included FLC-diff, cTnT and B2M. This system again allowed grouping of patients (n=450 with all three variables available) into four distinct groups with divergent outcome with OS of 4.6, 10.5, 36.8, and not reached for stages 3, 2, 1, and 0 respectively; (Figure B). Conclusion: Incorporation of plasma cell related measurements into the existing staging system using cardiac biomarkers allows better risk stratification of patients with AL amyloidosis. The system using FLC, B2M and troponin has the advantage of easily available laboratory values and can be widely adopted. Addition of NT-ProBNP into the system allows better refinement of the intermediate risk groups. This system will allow upfront risk stratification and development of risk-adapted therapies for patients with AL. Disclosures: Gertz: Genzyme: Research Funding.

Abstract Abstract 1892 Background: Primary or light chain amyloidosis (AL) is a clonal plasma cell disorder characterized by a relatively low plasma cell burden and multi-organ deposition of immunoglobulin light chain derived amyloid fibrils. The survival of patients with amyloidosis is quite variable with median survival of 12–18 months in different series, and largely dependent on the number and severity of organs involved. In addition to new drugs, the combination of melphalan and dexamethasone is an effective regimen for AL and risk adapted approaches to SCT has decreased treatment related mortality. It is not clear if the recent progress in risk stratification and treatment approaches have translated into improved survival for these patients. Methods: We studied two separate cohorts of patients; the first cohort consisting of 2118 patients with AL seen at Mayo Clinic over a 40-year period between November 1967 and August 2006 was used to examine the trends in overall survival from diagnosis during this time period. The second cohort of 443 patients seen between September 2006 and August 2009 were examined only for the changes in early mortality and for validation of a model for predicting early mortality developed using the first cohort of patients. Results: The estimated median follow up for the entire group was 7.8 years (95% CI; 7.1, 8.5). The patients were divided into 4 cohorts based on date of diagnosis; 1966–1976 (n=121), 1977–1986 (n=343), 1987–1996 (n=636) and 1997–2006 (n=1017). The median OS from diagnosis for the four cohorts were 0.9, 1.2, 1.2 and 1.5 years respectively, P < 0.001 (Figure 1). More importantly there was steady improvement in the long-term survival among these patients with 4-year survival estimates of 16%, 21%, 24% and 33% respectively. Next, we specifically looked at the survival trends within the last cohort (1996-2006), dividing the patients into three time periods; 1996–1999 (n=263), 2000–2003 (n=291) and 2003–2006 (n=463). While the OS from diagnosis was relatively unchanged between the first and second time periods (28 vs. 30% 4 year survival), there was significant improvement in the last three year period (42% survival at 4 years); P =0.02. Given the high early mortality in this disease, we specifically examined the one-year mortality among the entire patient cohort and found no clear improvement in terms of one-year mortality. Using logistic regression analysis, we identified cTnT > 0.01, NT-ProBNP > 4200 and Uric Acid > 8.0 mg/dL to be the best predictors of death within one year of diagnosis. The probability of death within one year for those with none, 1, 2, and 3 risk factors was 19, 38, 61, and 78% respectively, P < 0.01 (Figure 2). This model was further validated among patients seen during the most recent three years, September 2006 and August 2009. Conclusion: Survival in AL amyloidosis has improved over time, particularly in the recent years and likely reflects the recent improvements in therapy. However, early mortality remains a problem and baseline factors can be used to identify patients at risk for early mortality, who then can be targeted for risk adapted therapy approaches in clinical trials. Disclosures: Kumar: Celgene: Consultancy, Research Funding; Millennium: Research Funding; Merck: Consultancy, Research Funding; Novartis: Research Funding; Genzyme: Consultancy, Research Funding; Cephalon: Research Funding. Off Label Use: Lenalidomide for treatment of newly diagnosed myeloma. Dispenzieri:Celgene: Honoraria, Research Funding; Binding Site: Honoraria.

Abstract Abstract 2574 Rasburicase is a recombinant urate oxidase approved for the prevention of tumor lysis syndrome (TLS). In adults, the recommended dose is 0.2 mg/kg intravenously daily up to 5 days. For a 75 kg person, the cost is over $5100. Given the vast reduction of uric acid seen with this dose, it may be possible to use a reduce dose. The aim of this study is to evaluate the effectiveness of 0.1 mg/kg dose vs the recommended dose in the prevention of TLS. Between October of 2003 and February of 2009, all patients with a hematologic or oncologic diagnosis who received rasburicase were reviewed. Rasburicase had been dosed at either 0.1 mg/kg (low dose) or 0.2 mg/kg (standard dose) due to variations in practice among the practitioners. Rasburicase was given on the first day of chemotherapy and repeated as necessary. Renal insufficiency was defined by 0.5 mg/dl rise in serum creatinine (Scr) within 7 days of chemotherapy initiation. The need for dialysis was recorded separately. Labs were performed pretreatment and usually daily after chemotherapy until electrolytes stabilized. A multivariate analysis using pretreatment labs was performed to determine risk factors for renal insufficiency or failure. During the study period, 125 patients received rasburicase. Fifty-five patients received low dose rasburicase and 70 received the standard dose. The diagnoses are listed in Table 1. Lactate dehydrogenase (LDH) was similar between the two groups (604 U/l in low dose vs 665 U/l in standard dose, p = 0.69). Additional doses were required in 14.5% of the low dose patients vs 21.4% of the standard dose patients, p = 0.36. Renal insufficiency was noted only in the standard dose patients 7.4% vs 0%, p = 0.04. Rates of dialysis were similar between low dose patients 7.3% vs 17.4% in the standard dose patients, p = 0.10. Dialysis was often initiated for hyperkalemia and hyperphosphatemia. Only 12.5% were started for elevation in Scr. Uric acid reduction was higher in the standard dose patients (90.4 ± 15.8%) as compared with low dose patients (82.0 ± 19.0%, p = 0.01). Pretreatment Scr, uric acid and phosphorus levels were found to be significant variable for dialysis but only uric acid was found to be significant in the multivariate analysis. Pretreatment uric acid was a risk factor for dialysis (p < 0.001) but not for renal insufficiency (p = 0.24). Our results suggest that low dose rasburicase may be as effective as 0.2 mg/kg. While high dose rasburicase was more effective at reducing uric acid, it did not reduce the need for dialysis because dialysis was often needed for hyperkalemia and hyperphosphatemia which rasburicase did not affect. Despite bias, our results suggest low dose rasburicase may be a viable alternative and definitely support further study into its use. The potential savings could be enormous given that it would be $2550 per patient. Table 1. Baseline diagnosis of 125 patients with TLS Diagnosis Low dose Standard dose dose Diffuse large B-cell lymphoma 13 17 Chronic lymphocytic leukemia 3 12 Multiple myeloma 7 5 Non-Hodgkin's lymphoma 4 7 Acute myeloid leukemia 4 7 Burkitt's lymphoma 3 4 T-cell lymphoma 4 3 B-cell lymphoma 5 1 Solid Tumor 1 5 Myelofibrosis 3 2 Mantle cell lymphoma 3 1 Amyloidosis 0 2 Blast crisis in acute leukemia 1 1 B-cell lymphoproliferative disorder 0 1 Chronic myeloproliferative disorder 0 1 Follicular mixed lymphoma 0 1 Hodgkin's lymphoma 1 0 undifferentiated 1 0 Lymphoreticular myeloid malignancy 1 0 Post-transplant lymphoproliferative disorder 0 1 Splenic marginal cell lymphoma 1 0 Waldenstrom macroglobulinemia 1 0 Total 55 70 Disclosure: No relevant conflicts of interest to declare.