Citations

Abstract Background: AML outcome in elderly patients unfit for intensive chemotherapy continues to be challenging. Hypomethylating agents (HMA) can be effective in these patients but responses are short-lived (Fenaux et al, JCO 2009; Dombret et al Blood 2015). Even though HMAs achieve superior response rates to supportive care alone, complete remissions (CR) are rarely achieved and often quickly lost when they occur. The majority of patients will either have stable disease or progress through therapy. We hereby describe the outcome of these patients at our institution after they fail HMAs. Methods: A total of 56 AML patients' data from Feb 2007 to June 2015 at Mayo Clinic were reviewed after appropriate IRB approval was obtained. All cases had their bone marrow slides reviewed at our institution. Patients were considered for our study if they received HMA as frontline therapy for their AML. Patients who received HMAs after induction chemotherapy failure or for their MDS were excluded. Response was identified based on IWG AML criteria (Cheson et al, JCO 2003). Prognostic factors were analyzed by univariate and multivariate analyses. Survival estimates were calculated using Kaplan-Meier curves and univariate and multivariate analyses was based on log-rank testing using JMP software version 10. Results: i) Characteristics at Diagnosis: We identified 56 AML patients treated with HMA as frontline therapy with median age of 76 years (range, 59-91), 68% of which were males. Cases were classified as AML with myelodysplasia related changes in 25%, and therapy related myeloid neoplasm in 2%, and not otherwise specific in 73%. Out of 56 cases, 23% were secondary to MDS, 2% ET (Essential Thrombocytopenia), 5% PV (Polycythemia Vera), 5% MF (Myelofibrosis), 4% from CMML (Chronic Myelomonocytic Leukemia); while 61% were de novo AML. Per cytogenetics data, 2% were considered favorable risk, 62% intermediate, and 36% poor (34% had complex karyotype). ii) Clinical outcome to Frontline HMA: Out of 56 patients, 15(27%) patients received azacitidine (AZA) and 41 (73%) received decitabine (DAC) with a median number of cycles of 4 (range 1-62). Median number of cycles was higher for AZA vs DAC (8 vs 4, p =0.3). CR was found in 10 (18%), CR with incomplete count recovery in 3 (5%), partial remission (PR) in 3 (5%) patients, with overall response of 28% and median response duration of 10 months (range, 1-78 months). Thirteen (81%) out of 16 responders relapsed. Therefore, only 53 patients were included in primary or secondary failure analysis. iii) Clinical outcome to Frontline HMA failure: Out of 53 patients, 12 (23%) received subsequent treatments. Out of the 12 patients, 3 (25%) received another HMA (switch), 2 (17%) subcutaneous low dose cytarabine, 4 (33%) intensive chemotherapy, while 3 (25%) received either therapy on a clinical trial or targeted therapy. None of the 12 patients who got first salvage therapy achieved remission. Five out of the 12 patients received second salvage therapy, 2 (40%) of which achieved CR. Only one patient had third salvage therapy and did not respond. No patient was able to proceed to allogenic hematopoietic cell transplantation. Median overall survival (mOS) for this group was 2 months. Median OS for patients who received subsequent salvage therapies was better than those who did not receive any subsequent therapy after failing HMA (9.5 vs 2 months, p = 0.0009). Conclusions: Elderly patients diagnosed with AML who are unfit for intensive chemotherapy have a poor prognosis. HMA can get some patients into remission but most patients will ultimately fail. Outcome for patients who have primary or secondary (relapse) failure is very poor with a median OS of 2 months. Switch of HMA did not result in any response. Response to salvage therapy continues to be low but offering salvage therapies seem to improve OS. Our study provides historical data for future novel therapies, which are sorely needed for these patients. Figure Figure. Disclosures Al-Kali: Celgene: Research Funding; Onconova Therapeutics, Inc.: Research Funding.

Abstract Abstract 5226 Background Chronic myelomonocytic leukemia (CMML) is a clonal stem cell disorder with overlapping features between myelodysplastic syndromes and myeloproliferative neoplasms. Allogeneic stem-cell transplant (SCT) is considered to be a potentially curative option (Eur J Haematol. 2013; 90:355); with karyotype and comorbidity index (CI) being independent prognostic factors. The role of newer CMML prognostic models remains to be elucidated. We carried out this study to analyze the predictive value of the Mayo model (Leukemia 2013; 27:1504) in assessing transplant outcomes for patients with CMML. Methods After due IRB approval, 25 patients with WHO-defined CMML that underwent allo-SCT at Mayo Clinic from 1992 through 2013 were identified. All patients underwent bone marrow examination and cytogenetic evaluation at diagnosis. Clinical features including status at transplant, graft-versus-host disease (GVHD) prophylaxis, donor-recipient HLA-matching, donor-recipient blood types and CMV status were documented. Patients were evaluated for development of GVHD, disease relapse, remission status, and death from all causes. We evaluated the prognostic relevance of clinical and laboratory parameters including those previously identified by the MDAPS (Blood 2002;99:840), Spanish cytogenetic stratification (Haematologica 2011;96:375), and the recently described Mayo model Results Among 25 study patients, 14 (56%) were males and 15 (60%) had WHO defined CMML-1 (6-symptomatic/transfusion dependent, 3-monosomy 7). The median age at transplant was 51 years (range, 18-66 years). Graft sources included: 19 (76%) peripheral blood, 5 (20%) bone marrow, and 1 (4%) double umbilical cord blood. Ten (40%) patients received a reduced intensity conditioning. At last follow up, 15 (60%) deaths were documented while the remainders are alive and disease free. There were six (25%) post-transplant relapses all resulting in mortality. The 5-year OS was 42% and the 5-year non relapse mortality was 35%. Based on the Mayo model, 15 (60%) patients received a high-risk prognostication, 6 (24%) were intermediate and 3 (12%) were low. In an univariate analysis that included: demographics, blood counts at diagnosis and transplant, WHO classification (p=0.19), Spanish karyotypic stratification (p=0.67), prognostication according to the MDAPS (p=0.35) and Mayo model, disease status at transplant, SCT comorbidity index (p=0.06), graft sources, transplant conditioning regimens (p=0.08), development of acute (p=0.64) and chronic GVHD (p=0.06); thrombocytopenia at diagnosis (p=0.04), disease status at transplant (p=0.02), and a high risk prognostication using the Mayo model (p=0.01) were statistically significant. On a multivariable analysis only high risk prognostication by the Mayo model (p=0.02) retained its negative prognostic impact. In an univariate analysis for relapse-free survival, risk stratification by the Spanish cytogenetic system (p=0.04) and the Mayo model were prognostic (p=0.01), with the high risk prognostication by the Mayo model retaining its negative prognostic impact (p=0.01). Conclusions Allogeneic SCT remains a viable treatment option for patients with CMML. The Mayo model serves as a valuable tool, helping with the identification of high risk CMML patients. These patients seem to benefit from allo-SCT in comparison to non-transplant therapeutic options. Given the smaller sample size, validation in a larger patient cohort is needed. Disclosures: No relevant conflicts of interest to declare.

Abstract Background Cytogenetic abnormalities have significant prognostic role in AML at the time of diagnosis. Several prognostic groupings have been recommended including the NCCN and European LeukemiaNet classification. It is still unknown whether cytogenetic abnormalities have any prognostic role at the time of relapse for predicting clinical outcome. In addition, clonal evolution (CE) has been seen in many AML cases, but its prognostic value remains to be identified. Aim To study cytogenetic abnormalities and clonal evolution impact on clinical outcome for patients diagnosed with rAML Method A retrospective, single institution study through chart review of all cases diagnosed with rAML at Mayo Clinic Rochester between 2003 to 2011 was performed. Pts who achieved complete remission and relapsed at our institution were included. Pts with PML-RARA and BCR-ABL rearrangements were excluded from this study. Pts with available cytogenetics data, both at presentation and relapse, were included. Response to treatment was defined as per international working group (IWG) criteria (Cheson JCO 2003). Appropriate IRB approval was obtained in accordance with Helsinki declaration. Comparison between groups’ medians was done using Wilcoxon test, while survival estimates were calculated using Kaplan-Meier curves using JMP V9. Results Out of 192 pts, 74 pts matched our inclusion criteria. The median age at diagnosis is 60 years (range 18-81), 53% were males. At relapse, median white blood cell (WBC) was 3.2 x109, hemoglobin 10.4 g/dL, platelet 59 x109, peripheral blood (PB) blast 2% (range, 0-96), bone marrow (BM) blasts 23% (range, 1-95) and cellularity 60%. Relapse site was mainly in the BM (95%). Flt3 mutation was found in 12/43 (28%), while NPM1 mutation in 2/7 (29%). Upon treatment with salvage therapy, 27 (36%) pts achieved complete remission (CR) and 16 (22%) pts partial remission (PR) for a total overall response (OR) of 58%. Twenty one (28%) pts were able to proceed to allogeneic stem cell transplant (SCT) after achieving remission. Median overall survival (mOS) was 198 days. Pts had a longer mOS if they proceeded to SCT (386 days vs 121 days, p =0.0003). Cytogenetic analysis (CG) were diploid in 32 (43%) of 74 pts at relapse. Five pts had favorable cytogenetics (3 pts with t(8,21) and 2 pts with inv (16)). CG was divided into favorable, intermediate and poor prognosis as per NCCN classification in 5 (7%), 45 (61%), 24 (32%), respectively. As per European LeukemiaNet, there were 4 CG groups divided into: favorable, Intermediate-1, intermediate-2, and adverse in 5 (7%), 33 (45%), 12 (16%), 24 (32%), respectively. Upon comparison of CG at diagnosis and relapse; clonal evolution (a change in CG due to gain or loss of chromosomal abnormalities) was found in 42/74 (57%) of pts. CG reverted back to diploid in 8/42 (19%) or gained additional CG abnormalities at relapse in 32 (76%) pts. CR2 was achieved in 80% of favorable CG (NCCN grouping), 40% (intermediate), and 21% in poor risk (p=0.01). Using NCCN CG grouping mOS was: 435 days (favorable), 266 (intermediate), and 104 (poor) (p value = 0.01). However, using European LeukemiaNet grouping mOS was: 435 days (favorable), 174 (intermeidtae-1), 363 (intermediate-2), and 104 (poor) (p=0.004). CE did not affect mOS (p =0.7). On multivariate analysis, response to salvage therapy (p<0.0001), NCCN CG grouping (p =0.03), wbc (p=0.045), and PB blasts (p=0.01) were significant but not age, BM blast, platelet and CE. Conclusion CG were diploid in 43% of relapsed pts. CE (gain or loss of cytogenetic abnormalities) was seen in 57% of CG at relapse compared to diagnosis but did not affect mOS. NCCN (and not European LeukemiaNet) cytogenetic grouping did affect clinical outcome through both achieving second CR and mOS (both on univariate and multivariate analysis). Our results confirm the prognostic value of CG on relapse, and need to be confirmed by larger studies. Disclosures: No relevant conflicts of interest to declare.

Abstract Background Tunneled central venous catheters have traditionally been used for administration of chemotherapy (chemo) for acute leukemia treatment. Peripherally inserted central catheters (PICC) are increasingly being used, but there is an increased risk of PICC-associated thrombosis (PAT). There is limited information on management of PAT in this subgroup of patients (pt). Methods In this retrospective cohort study we investigated the primary management of PICC related upper extremity venous thrombosis (TB) in pt with hematologic malignancy undergoing chemo at Mayo Clinic Rochester. Eligible pt were 18 years and older, had documentation of PICC placement at our institution, initiation of chemo within a week of PICC placement, and the TB was objectively documented by compression venous ultrasound (CVU). Superficial venous TB (SVT) was defined as TB of subcutaneous veins (basilic and cephalic) and deep venous TB (DVT) defined as TB of brachial, axillary, subclavian, or innominate superior vena cava and internal jugular veins. We retrospectively stratified initial management (IM) into 1) PICC removal alone (PR), 2) PR and therapeutic anticoagulation (AC), 3) AC alone, and 4) conservative management (CM) consisting of symptomatic care and observation. We also investigated factors influencing IM of PAT. Long term outcomes including incidence of pulmonary embolism (PE), post phlebitic syndrome and recurrence of TB were noted. Results Between 2006 and 2012, 190 pt with AML (n=160), MDS or MDS/MF (n=10), or ALL (n=20) met our study criteria. Overall, 40/190 (21.6%) developed PAT: AML n=38, ALL n=1, MDS/MF n=1. SVT occurred in 16/40 (40%) pt and DVT occurred in 24/40 (60%) pt. IM is shown in Table 1. One pt with SVT received AC for 6 weeks as well as PR. In this pt, IM included AC because of proximity of TB to deep veins. 7 pt had a follow up (f/u) CVU within 3 months demonstrating: TB progression (prog, n=1, IM with PR alone but with prog AC was started); TB stability (stab, n=2, both IM with PR alone) and TB resolution/improvement (R/I, n= 4, 2 IM with PR and 2 IM with CM) In the 5 pt with CM, 2 pt had f/u CVU demonstrating improvement, one pt had charted clinical improvement , and 2 pt had no charted clinical f/u. Of DVT pt on AC (n=14), 5 pt completed at least 2 months of AC. Reasons for early cessation (n=9) of AC included thrombocytopenia (tcp, n=4), hematoma (n=1), hematuria (n=1), subarachnoid hemorrhage (n=1), and unknown (n=1). One pt died 2 days after presence of DVT was noted. Two pt initially treated with PR alone developed a second TB with placement of a new PICC on contralateral arm and AC was subsequently initiated (duration: 3 months (n=1), and 8 days (n=1)). 14 pt with DVT had f/u CVU within 3 months which demonstrated prog (n=2), stab (n= 8), or R/I ( n=4) of TB (Table 2). The 2 pt with TB prog subsequently received longer term AC guided by platelet count. In 5 pt with PICC not initially removed, 3 pt had PR after completion of chemo. (1-4 days), one pt had PR at discharge (27 days from TB), and one pt had PICC in place on discharge. One pt with DVT experienced a PE within one year f/u. This pt had initially completed 3 days of AC which was stopped because of tcp. Conclusion In this cohort of patients with PAT, the most common IM consisted of PR, which appeared to result in a low recurrence rate among pt with SVT. In pt with DVT, the role of AC remains to be defined given the abbreviated course of AC in most pt. Though, the lack of CVU followup in all pt limits conclusions, there appeared to be a low rate of symptomatic prog. Initiation of AC was largely guided by platelet count and degree of occlusive symptoms. Thrombocytopenia was also the most commonly cited reason for discontinuing AC. Close clinical follow up aided by the use of CVU was integral to ensure that PAT did not lead to further adverse events. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 2910 Poster Board II-886 Background: Polycythemia vera (PV) and primary myelofibrosis (PMF) are stem cell-derived myeloproliferative neoplasms characterized clinically by increased red cell volume and bone marrow fibrosis, respectively. Phlebotomy is the cornerstone of treatment in PV whereas hydroxyurea has been the traditional drug of choice in both PV and PMF. Over the last 20 years, several new treatment approaches have been introduced and promoted (e.g. interferon-alpha, anagrelide, thalidomide, allogeneic stem cell transplantation) but their benefit in terms of survival has not been confirmed in controlled studies. Methods: Study patients were recruited form the Mayo Clinic database for myeloproliferative neoplasms. Diagnosis was based on 2001 WHO criteria. Follow-up information was updated in July, 2009. In order to assess changes in survival over time, patients were stratified into two groups by year-of-diagnosis: 1970-1990 and 1991-2005. Survival analysis was performed by the Kaplan-Meier method and comparisons made by the log-rank test. Cox regression model was used for multivariable analysis. Results: 652 patients were studied, including 314 with PMF (median age 57 years, range 18-88; 193 males) and 338 with PV (median age 63 years, range 18-93; 183 males). Overall median survivals were 6.5 years for PMF and 12.2 years for PV (p<0.0001); ten- and twenty-year survival rates were 37% and 13% in PMF and 60% and 18% in PV (Figure). Both the International Prognostic Scoring System for PMF (IPSS; Cervantes et al. Blood 2009;113:2895) and the age/leukocyte count-based PSS for PV (Gangat et al. BJH 2007;138:354) were validated in the current series of patients. Considering all patients, there was no significant difference in survival when either PMF (p=0.91) or PV (p=0.73) patients were stratified according to year-of-diagnosis (Figure); median survivals were 4.7 and 6.6 years in PMF patients diagnosed 1970-1990 (n=54) and 1991-2005 (n=260), respectively; the corresponding median survival figures for PV were 11.8 (n=170) and 12.2 (n=168) years. These results did not change when adjusted for age and sex and, in PV, for age/leukocyte count-based PSS. In PMF, however, borderline significance (p=0.05) was noted when IPSS scores were introduced into the Cox model with further analysis showing improved survival in Int-2/high risk patients with PMF diagnosed after 1990 (p=0.01). Conclusion: The current study provides mature survival data in PMF and PV using large cohorts of consecutive patients seen at a single institution. Although median survival is significantly better in PV compared to PMF, their survival curves converge after 20 years of follow-up, underscoring the poor long-term prognosis in PV patients. The study also suggests improved survival for high/intermediate-2 risk PMF in recent years. Disclosures: Off Label Use: hydroxyurea use in myelofibrosis.

Abstract Introduction L-asparaginase therapy results in systemic depletion of antithrombin (AT) and fibrinogen (FG), resulting in thrombotic (TE) and hemorrhagic events (HE). Prophylactic correction of coagulation disturbances have been inconclusive with regard to cost effectiveness, and specifically cost effectiveness of antithrombin concentrate (ATC) replacement has not been evaluated. (Abbott et al, 2009, Mitchell et al, 2003). Currently, AT repletion is recommended for levels less than 60%. (Stock et al, 2012) Several treatment protocols with distinctly different L-asparaginase (LAS) or pegaspargase (PEG) doses complicate the frequency and intensity of ATC replacement. We carried out this study to evaluate the cost effectiveness of ATC replacement, based on our institutional practice. Methods This single center, retrospective, observational study was approved by the Institutional Review Board and conducted at Mayo Clinic, Rochester, MN. Consecutive adult patients with a confirmed diagnosis of ALL receiving LAS or PEG between January 2000 and October 2012 were evaluated. Patients with hereditary or acquired thrombophilias or pre-existing TE or HE, were excluded. Per institutional guidelines, ATC was administered if levels were less than 70%. The Centers for Medicare & Medicaid (CMS) single drug historical pricing guide was utilized for 2000-2004. ATC cost beginning January 1, 2005, was derived from the CMS average sales price, reported quarterly. To adjust for price differences due to regimen utilization trends, each dose was also calculated at the final 2012 cost of 3.05/unit. ATC repletion practices and estimated expenditures were compared between regimens. Results Fifty four patients were treated with LAS or PEG containing regimens. Of these, 47 (87%) received ATC. Twenty-eight patients received LASP, 22 received PEGASP and 4 received both. Distribution across protocols was; CALGB 9111 (n=20, Larson et al, 1998), E2993 (n=13, Goldstone et al, 2008), C10403 (n=9, NLM: NCT00558519), CCG 1941 (n=7, Gaynon et al, 2006) and Augmented HyperCVAD (n=5, Faderl et al, 2005). PEG and LAS dosing schemes are outlined in table 1 . LAS doses were converted to PEG upon specific protocol revision or at discontinuation of LAS availability in 2012. Regimens specified a dose cap with the exception of C10403. A total of 399 LAS and PEG doses resulted in 378 instances of ATC repletion. The reimbursement price of ATC in 2000 was estimated at 1.50/unit U.S. dollars (USD) and 3.04/unit USD by the fourth quarter of 2012. The total estimated average sales price was 2,471,745 USD over the study period. C10403 resulted in the highest median number of PEG doses per patient (n = 5) and a median of 17 ATC doses per dose of PEG. This resulted in C10403 being the most expensive regimen with the median expense for ATC replacement being 118,064 USD (chart 1). This conclusion was also seen after comparing different regimens using the most recent 2012 cost estimates. The overall incidence of TE or HE was 17% and equally distributed between regimens (please see accompanying abstract). Conclusions Repletion of ATC in patients receiving asparaginase containing chemotherapy regimens is very expensive and continues to add to costs at many centers, without clear benefit, especially with regards to TE and HE. A prospective, randomized, multicenter clinical trial looking to answer this question is the need of the hour. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 2895 Poster Board II-871 Background: Recent studies have reported a significant association between JAK2V617F acquisition and a specific JAK2 germline haplotype; rs12343867 (located at intron 14 of the JAK2 gene) is one of several SNPs that tag this haplotype. We performed rs12343867 SNP analysis in a consecutive cohort of patients with primary myelofibrosis (PMF) in order to study the clinical and laboratory correlates of the particular JAK2 haplotype and its effect on survival. Methods: Study patients were recruited form the Mayo Clinic database for PMF. Molecular studies were performed on DNA extracted from stored bone marrow. rs12343867 SNP genotyping was performed by a commercially available Taqman assay. Quantitative JAK2V617F analysis was done according to previously published methods. Standard statistical methods were used to test significance of associations between SNP genotype and other variables. Survival analysis was performed by the Kaplan-Meier method and comparisons made by the log-rank test. Cox regression model was used for multivariable analysis. Results: 132 patients with PMF were studied (median age 62, range 28-82; 82 males). Risk distribution according to the International Prognostic Scoring System (IPSS) for PMF were low in 39 (29%) patients, intermediate-1 in 41 (31%), intermediate-2 in 25 (20%) and high in 27 (20%). 77 (58%) patients were JAK2V617F positive; median mutant allele burden was 26% (range 1-85) and 21 patients displayed > 50% mutant allele burden. The rs12343867 genotype distributions were C/C 22%, C/T 44% and T/T 34% . The corresponding figures in JAK2V617F positive/negative cases were 31%/9%, 38%/53%, 31%/38% (p=0.01). Among the 21 patients with > 50% JAK2V617F allele burden, 14 (67%) displayed the C/C allele, 2 C/T and 5 T/T. The specific SNP genotype was not significantly affected by age (p=0.33), sex (p=0.46), leukocyte count (p=0.39), platelet count (p=0.09), or IPSS risk category (p=0.63). Patients were followed for a median of 54 months and during this period 73 (55%) deaths were documented. The T/T SNP genotype was significantly associated with shortened survival, compared to either C/C or C/T (p=0.001; Figure). Multivariable analysis showed this association to be independent of IPSS, JAK2V617F mutational status, age or sex. The adverse prognostic effect of the T/T genotype was also apparent when JAK2V617F negative (p=0.01) or positive (p=0.06) cases were analyzed separately. Conversely, JAK2 mutational status or allele burden did not affect survival in patients with T/T or C/C allele. Conclusion: The current study illustrates the non-random haplotype distribution of JAK2V617F in patients with PMF. More importantly, a non-JAK2 haplotype, tagged by the rs12343867 T allele, was associated with inferior survival that is not accounted for by IPSS or JAK2V617F mutational status. These findings, together with previous observations regarding shortened survival associated with low JAK2V617F allele burden, suggest the presence of molecular events in PMF that are more aggressive than JAK2V617F and not necessarily linked to the JAK2 haplotype. Disclosures: No relevant conflicts of interest to declare.

Abstract Background Chronic myelomonocytic leukemia (CMML) is a clonal hematologic disorder that was classified by the World Health Organization (WHO) as a myelodysplastic/ myeloproliferative overlap disease. Cytogenetic abnormalities have a significant prognostic role in many hematologic neoplasms, but their prognostic value in CMML has been debatable. Recently, monosomal karyotype (MK) has been reported to be a marker of poor prognosis in patients with myelodysplastic syndromes (MDS) and primary myelofibrosis, but its value in CMML is unknown. Aim To study MK effect on clinical outcome for patients diagnosed with CMML Method A retrospective study of all cases diagnosed with CMML at Mayo Clinic Rochester between 1994 to 2011 was performed. Only pts with complete cytogenetic analysis at presentation to our institution were included. MK was defined as the presence of ≥ 2 autosomal monosomies or one autosomal monosomy with at least one structural abnormality (Breems et al, JCO 2008). CK was defined as the presence of at least 3 chromosomal abnormalities. Appropriate IRB approval was obtained in accordance with Helsinki declaration. Comparison between groups’ medians was done using Wilcoxon test, while survival estimates were calculated using Kaplan-Meier curves using JMP V9. Results A total of 262 pts diagnosed with CMML had available cytogenetic data at diagnosis. Median age was 72 years, 176 (67%) were male. Median hemoglobin 10.5 g/dL, white blood cells (wbc) 12 x109/L, platelet 89 x109/L, peripheral blood (PB) blast 0, and bone marrow (BM) blast 4%. CMML2 was seen in 9% while 47% were proliferative (wbc >13). Leukemic transformation was documented in 34 pts (13%). Median overall survival was 513 days. Cytogenetic (CG) analysis was diploid in 167 pts (64%). Trisomy 8 was the most frequent cytogenetic abnormality at 8% (22), followed by complex karyotype (CK) 5% (14), then -7 at 4% (10) and MK 3% (7, six of which were also CK). Comparing pts with diploid CG to other categories indicates: to abnormal CG pts had lower wbc (0.001), PB blasts (p<0.0001), and BM blasts (p=0.0001); to CK pts had lower PB blasts (p=0.003) and higher platelets (p=0.03); to -7 pts had lower wbc (p=0.005), PB blast (p=0.0004), BM blasts (p=0.03) and higher platelets (p=0.03); to +8 pts had lower PB blast (p=0.02) and BM blasts (p= 0.01); no difference was noted when compared to MK. Median OS was statistically significantly worse in MK+ vs MK- (24 vs 527 days, p= 0.002), but not in other comparisons: -7 vs others (250 vs 527 days, p=0.2), CK+ vs CK- (256 vs 527 days, p=0.05), diploid vs others (570 vs 365 days, p=0.1), +8 vs others (312 vs 527 days, p=0.1). Pts with MK+ only or MK+CK+ did worse than CK+ only or other groups (4, 63, 304, 527 days, respectively, p<0.0001). On a multivariate analysis, MK+ (in addition to platelet, BM blast, hemoglobin, and wbc) did have an impact on OS (p=0.0004), while CK+, -7, +8, diploid CG, age, PB blast did not. Conclusion Cytogenetic abnormalities were not frequent findings in pts diagnosed with CMML (36%), but did affect wbc, PB and BM blasts. The most common cytogenetic abnormality was +8 while MK was present at 3% (less than published data in MDS). Only MK predicted statistically significant shorter mOS between all other cytogenetic categories on both univariate and multivariate analysis. This finding needs to be validated by larger cohorts of pts due to its rare occurrence in CMML. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 2893 Poster Board II-869 Background: A specific JAK2 germline haplotype has been associated with the occurrence of JAK2V617F in myeloproliferative neoplasms (MPN). Several SNPs, including rs12343867, tag this haplotype. We performed rs12343867 SNP analysis in a consecutive cohort of patients with essential thrombocythemia (ET) in order to study the clinical and JAK2V617F allele burden correlates of the particular SNP genotype. Methods: Study patients were recruited form the Mayo Clinic database for ET. Molecular studies were performed on DNA extracted from stored bone marrow. rs12343867 SNP genotyping was performed by a commercially available Taqman assay. Quantitative JAK2V617F analysis was done according to previously published methods. Standard statistical methods were used to test significance of associations between SNP genotype and other variables. Survival analysis was performed by the Kaplan-Meier method and comparisons made by the log-rank test. Cox regression model was used for multivariable analysis. Results: 226 patients with ET were studied (median age 57, range 14-91; 145 females). Median hemoglobin, leukocyte and platelet counts at initial diagnosis were 13.8 g/dL, 9.7 × 109/L and 1000 × 109/L, respectively. 118 patients (52%) were JAK2V617F positive and among them, 90 (76%) displayed JAK2V617F allele burden of < 10% (low allele burden group) and the median (range) mutant allele burden in the remaining 28 patients (higher allele burden group) was 12% (12-68). As expected, JAK2V617F-positive cases, compared to their mutation-negative counterparts, were older (p=0.006) and displayed higher hemoglobin (p=0.0001) and leukocyte (p<0.0001) counts and lower platelet (0.03) count. The rs12343867 genotype distributions were C/C 15%, C/T 53% and T/T 33% (expected frequency in the HapMap-CEU listed control population is ∼ 6%, 46% and 48%, respectively). The corresponding figures in JAK2V617F positive/negative cases were 17%%/11%, 53%/53%, 30%/36% (p=0.29). Unlike the case with JAK2V617F mutational status, rs12343867 SNP genotype did not correlate with age (p=0.69), hemoglobin level (p=0.47), leukocyte count (p=0.25) or platelet count (p=0.33). Patients were followed for a median of 7.4 years and during this period 56 (25%) deaths were documented. rs12343867 SNP genotype did not correlate with survival (Figure). Similarly, survival did not correlate with the presence or absence of JAK2V617F. However, patients with higher JAK2V617F allele burden displayed a shorter survival with borderline significance, during multivariable analysis that included age as a covariate. Conclusion: The current study shows the JAK2 haplotype enrichment in a large series of ET patients. However, the particular phenomenon was not significantly different between JAK2V617F positive and negative cases, which suggests specific haplotype association with the disease rather than the mutation. Consistent with this contention, rs12343867 genotype did not correlate with clinical or laboratory variables that typically associate with JAK2V617F mutational status. Furthermore, unlike the case with primary myelofibrosis, rs12343867 SNP genotype did not correlate with survival, an observation that is not surprising considering the lower level of complexity in molecular pathogenesis in ET compared to PMF. Disclosures: Pardanani: TargeGen: Research Funding; Cytopia: Research Funding. Off Label Use: Data from ongoing clinical trial will be presented. Tefferi:TargeGen: Research Funding.

Abstract Background: Treatment-related AML (t-AML) accounts for 10 to 20% of all AML cases and carries an especially poor prognosis (Kayser et al. 2011, Godley et al. 2008). Patients diagnosed with t-AML are likely to have abnormal cytogenetic profiles with chromosome changes that are predictive of an aggressive malignancy, poor response to therapy, and decreased overall survival(Smith et al. 2003). The simultaneous use of multiple chemotherapeutic agents of different classes makes it increasingly difficult to predict risk for developing t-AML and determining disease course. An updated analysis of predictive factors for t-AML is needed so clinicians can more accurately inform patients of their prognosis. The aim of this study was to classify t-AML according to primary malignancy, previous chemotherapy exposure, and cytogenetic profile. Methods: A retrospective chart review of patients that were diagnosed with AML at Mayo Clinic from 7/1/1990 to 5/13/2015 was performed following IRB approval. AML patients found to have a previous malignancy treated with chemotherapy were classified as t-AML. Chemotherapeutics were classified as alkylating agents, antimetabolites, anti-tubulin agents, and topoisomerase II inhibitors. Patients diagnosed with a myelodysplastic or myeloproliferative disorder prior to development of AML were excluded from this study. Previous chemotherapy exposures, duration of chemotherapy exposure, complete blood count, chromosome abnormalities, and survival data were collected for t-AML cases. Cytogenetic changes were classified as favorable, intermediate, and adverse according to the system used by Kayser et al. 2011. JMP 10.0 was used for statistical analysis. Results: Out of 584 patients, 64 patients (11%) had a primary malignancy that was treated with chemotherapy prior to being diagnosed with AML. The most common primary malignancies were breast cancer (31%), non-Hodgkin lymphoma (27%), colorectal cancer (8%), and Hodgkin lymphoma (8%). Laboratory findings showed median hemoglobin 9.6 g/dL (4.7-13.8), median white blood cells 3.2 x109 (0.6-126), median platelets 50x109 (3-320), median peripheral blood blasts of 8% (0-91), and median bone marrow blasts 38% (1-94). 95% of patients diagnosed with t-AML had been previously treated with an alkylating agent. Additional exposure to an anti-metabolite trended towards a more adverse cytogenetic profile (χ2=5.0, p=0.08) but there was not a statistically significant decrease in overall survival (KM analysis, p=0.31). The median overall survival for patients diagnosed with t-AML was 10.2 months compared to 19.2 months for patients with de-novo AML (KM analysis, p=0.04). Adverse cytogenetic profiles were associated with decreased survival (KM analysis, p <0.0001). However, there was no difference in overall survival between patients with t-AML that had intermediate cytogenetics and those with de-novo AML (KM analysis, p=0.36). None of the chemotherapy classes other than antimetabolites were associated with poor cytogenetics or survival when combined with an alkylating agent. Conclusion: Over half of all patients classified as having t-AML in this study received prior chemotherapy for breast cancer or non-Hodgkin lymphoma. Cytogenetic classification of t-AML into favorable, intermediate and adverse groups is useful in predicting disease course. Interestingly, t-AML patients with intermediate risk cytogenetics had similar overall survival to patients with de-novo AML. This suggests that the poor outcomes observed in patients with t-AML is predominantly due to the subset with adverse cytogenetics. Thus, cytogenetic analysis remains the best indicator of overall survival regardless of chemotherapy exposure. Additional work is needed to delineate the risk associated with the aforementioned chemotherapy classes. Disclosures Al-Kali: Celgene: Research Funding.