Abstract: 【Background】Despite numerous attempts to elucidate the mechanism underlying other iatrogenic immunodeficiency-associated lymphoproliferative diseases (OIIA-LPDs), the mechanism remains poorly understood. Methotrexate-associated lymphoproliferative disorder (MTX-LPD) is one of the disease entities included in OIIA-LPDs, which is especially developed in rheumatoid arthritis (RA) as we demonstrated (Tokuhira et al, Leuk Lymphoma. 2012;53:616-23). Spontaneous regression of LPD after MTX withdrawal is one of the distinct characteristics of MTX-LPD, and recent articles suggest that the increase in peripheral blood (PB) lymphocytes affect this phenomenon. However, the population of lymphocytes necessary to induce MTX-LPD regression is still unknown. 【Methods】We prospectively analyzed 10 patients with RA developing confirmed LPD during MTX administration, in our institutions between January 2014 to October 2015. All patients ceased MTX after developing lymphoadenopathies. To investigate the factors involved in spontaneous regression of LPD following MTX withdrawal, we performed flowcytometric analysis of whole blood including lymphocytes from PB for CD3, 4, 8, 14, 15, 20, 56, 127, and 45RO, HLA-DR, CCR3, CCR4, CCR6, CCR7, and EBV-tetramer expression, for all patients at three time points during the clinical course of treatment; at the time of MTX cessation (w0), and 4 weeks (w4) and 12 weeks (w12) after MTX cessation. We selected 10 age-, sex-, MTX dose-, and RA duration-matched RA patients treated with MTX for more than 6 months as controls, randomly selected from among the RA patients at our institutions. The patients with MTX-LPD were divided into two groups based on the status of LPD after MTX cessation, regressive LPD (R group) and persistent LPD (P group), and clinical parameters were compared. 【Results】In the 10 RA patients with MTX-LPD, the median age was 62.4 years (40-74), and the ratio of male:female patients was 2:8. The median duration from the time of RA diagnosis to the time of MTX development was 7.2 years (1.2-20.4), and the median duration of MTX administration was 4.3 years (0.7-9.8). The pathological diagnosis of LPDs was diffuse large B cell lymphoma (DLBCL, n=5), Classical Hodgkin lymphoma (cHL, n = 3), and non-specific LPD (LPD-nos, n=2). The averages values for LDH, CRP, and sIL-2R in the serum were 220 IU/mL (176-554), 0.85 mg/dL (0.06-2.53), and 579 IU/L(206-1210), respectively. The number of patients in the R group and the P group was 7 and 3, respectively. According to LPD phenotypes, all 5 cases of DLBCL belonged to the R group, whereas 1 of 3 HL and 1 of 2 LPD-nos belonged to the R group. At w0, the median WBC count and absolute lymphocyte number in 10 MTX-LPD patients were 5810/µL (2700-11400) and 964/µL (220-2070), respectively. On the other hand, the median WBC count and absolute lymphocyte number in the 10 control patients were 5930/µL (3900-9000) and 1512/µL (755-2379), respectively. The absolute lymphocyte count in the P group was significantly higher than that in the R group. In addition, a significant increase in the lymphocyte number following MTX withdrawal was observed only in the R group, but not in the P group. There were several observations from the flowcytometric analysis. First, the proportions of effector memory CD8+ T cells (EM CD8+), EBV specific CD8+ T cells, and T helper 1 (Th1) cells were significantly decreased both in the R group and in the P group compared to the control group at w0, and this subset was significantly increased in the R group at w4 and w12 compared to the P group. Second, the proportion of CD14-CD15+ cells in the lymphocyte fraction, which includes myeloid-derived suppressor cells (MDSC), was significantly increased in LPD group at w0, and significantly decreased following MTX cessation only in the R group, but not in the P group. Third, the proportion of CD14-CD15+ cells negatively and significantly correlated with the proportion of EM CD8+T cells (r2 = 0.47, p 〈 0.0001). 【Conclusions】Although the mechanism of MTX-LPD regression following MTX withdrawal is still unknown, this study demonstrates that reconstitution with specific lymphocytes in the PB is definitely correlated with LPD-regression. The proportion of Th1 cells, EM CD 8+, EBV specific CD8+ along with CD14-CD15+ cells dramatically indicated restoration and transition following MTX cessation, suggesting their potent effect toward the progression and regression of MTX-LPD. Disclosures Tokuhira: Bristol Myers Squibb Co., Ltd: Honoraria; Pfizer Co., Ltd: Honoraria; Eizai Co., Ltd: Honoraria. Okamoto:Otsuka Pharmaceutical Co., Ltd.: Honoraria, Research Funding; Shionogi & Co., Ltd.: Research Funding; Sumitomo Dainippon Pharma Co., Ltd.: Research Funding; Kyowa Hakko Kirin Co., Ltd.: Research Funding; Eisai Co., Ltd.: Research Funding; Teijin Pharma Limited: Research Funding; Chugai Pharmaceutical Co., Ltd.: Research Funding; Toyama Chemical Co., Ltd.: Research Funding; Nippon Shinyaku Co., Ltd.: Research Funding; Bristol-Myers Squibb K.K.: Honoraria, Research Funding; Asahi Kasei Pharma Corp.: Research Funding; Astellas Pharma Inc.: Research Funding; Alexion Pharmaceuticals, Inc.: Research Funding; Pfizer Inc.: Honoraria, Research Funding; JCR Pharmaceuticals Co., Ltd.: Research Funding.

Abstract: Introduction: Tyrosine kinase inhibitors (TKIs) have dramatically improved the prognosis of chronic myelogenous leukemia (CML). The treatment with TKIs continues to improve the depth of response and overall survival of CML patients, but the life-long use of TKI is known to be associated with late complications such as cardiovascular events as well as heavy financial burden, and thus impairs the quality of life. To overcome these issues, many studies evaluating the possibility of TKI discontinuation have been ongoing worldwide. In order to achieve durable treatment-free remission, it is crucial to understand the dynamics of CML-leukemia initiating cells (LICs). We previously reported that CD25 was highly expressed in murine and human CML-LICs (Kobayashi CI et al., Blood, 2014). The aim of this study was to assess whether the proportion of CD25 positive cells in hematopoietic stem/progenitor cell fraction of bone marrow cells in CML patients treated with TKIs is associated with their molecular response and could serve as a novel surrogate marker to stop TKI therapy. Methods: Bone marrow samples were obtained from patients with CML in chronic phase who were diagnosed and have been treated solely with TKIs at Keio University Hospital (Tokyo, Japan). This study was approved by the institutional ethical committee and informed consent was obtained from each patient. Both quantitative and qualitative PCR of BCR-ABL was performed using bone marrow mononuclear cells (BMMNCs). The proportion of CD25 positive cells in bone marrow hematopoietic stem/progenitor cell (HSPC; CD34+CD38-) fraction was evaluated by flow cytometry using FITC-labeled anti-CD34, PE- labeled anti-CD38 and APC-labeled anti-CD25 antibodies. The response to TKIs at the time of evaluation was determined according to the previous report (Yoshida C et al., Int J Clin Oncol, 2012): complete cytogenetic remission (CCyR) defined as Philadelphia chromosome undetectable and quantitative PCR copy numbers 〉 731 among BMMNCs; major molecular remission (MMR) as quantitative PCR copy numbers ≤731, and complete molecular remission (CMR) as undetectable BCR-ABL by quantitative and qualitative PCR. Results: Bone marrow samples obtained from 95 patients were evaluated (median age 53 years old; male/female, 67/28). Analysis was performed prior to TKI exposure in nine patients and under TKI therapy including 2nd generation TKI in 64 patients (imatinib, 22; dasatinib, 33; nilotinib, 9). Remaining 22 patients were treatment free because they enrolled in a clinical trial of TKI discontinuation. The proportion of CD25 positive cells in HSPC fraction significantly decreased in patients with prior TKI exposure relative to patients at diagnosis (n=86; Mean 4.2%, SD 7.0% vs n=9; Mean 22.4%, SD 11.3%, P 〈 0.01). In addition, the proportion of CD25 positive cells in HSPC fraction significantly correlated the level of quantitative PCR (Figure, P 〈 0.0001), and MMR was also dividable from CMR by the proportion of CD25 positive cells (MMR; n=29, Mean 4.6%, SD 4.8% vs CMR; n=49, Mean 2.6%, SD 2.5%, P 〈 0.05). The expression of CD25 was still detected in the majority of patients who achieved CMR, including those who sustained CMR after the discontinuation of TKIs. Conclusion: We confirmed that the expression of CD25 in HSPC fraction of CML patients was significantly correlated with the response to TKI therapy, and may serve as an asset to select patients who are likely to achieve durable treatment-free survival. Figure Figure. Disclosures Karigane: Celgene: Honoraria. Sakurai:Celgene: Honoraria. Matsuki:Bristol-Myers Squibb: Honoraria; Celgene: Honoraria; Nippon Shinyaku: Honoraria. Kikuchi:Celgene: Honoraria; Takeda Pharmaceutical Company: Honoraria; Kyowa Hakko Kirin: Honoraria. Mitsuhashi:LSI Medience: Consultancy. Okamoto:Sumitomo Dainippon Pharma Co., Ltd.: Research Funding; Asahi Kasei Pharma Corp.: Research Funding; Astellas Pharma Inc.: Research Funding; Shionogi & Co., Ltd.: Research Funding; Nippon Shinyaku Co., Ltd.: Research Funding; Alexion Pharmaceuticals, Inc.: Research Funding; Toyama Chemical Co., Ltd.: Research Funding; Otsuka Pharmaceutical Co., Ltd.: Honoraria, Research Funding; Teijin Pharma Limited: Research Funding; Bristol-Myers Squibb K.K.: Honoraria, Research Funding; Kyowa Hakko Kirin Co., Ltd.: Research Funding; Eisai Co., Ltd.: Research Funding; Chugai Pharmaceutical Co., Ltd.: Research Funding; Pfizer Inc.: Honoraria, Research Funding; JCR Pharmaceuticals Co., Ltd.: Research Funding.

Abstract: Background: Bronchiolitis obliterans (BO) after allogeneic stem cell transplantation (allo-SCT) is a late-onset, life-threatening respiratory complication which significantly reduces patients’ quality of life. Despite different therapeutic protocols, BO mortality remains high and most patients die of respiratory failure or infections. In recent practice, the source of stem cells and the conditioning regimen for allo-SCT have become more varied, but their influence on the incidence of BO is not established. Here, we retrospectively analyzed incidence of and risk factors for BO in allo-SCT protocols. Patients and methods: Between Jan 1994 and June 2005, 2692 patients underwent allo-SCT in 14 facilities of the Kanto Study Group for Cell Therapy (KSGCT) in Japan, and 2154 surviving at least 100 days after transplantation were evaluated in this study. Clinical diagnosis of BO was made by pulmonary function tests (PFT) revealing a forced expiratory volume for 1 second (FEV1) less than 70% and FEV1/forced vital capacity less than 80% of the predicted value, along with typical changes on high-resolution computed tomography. Results: BO developed in 57 patients with a cumulative incidence at 5 years post transplant of 2.6%. The Kaplan Meier estimate of median time interval from transplant to diagnosis of BO was 335 days (83–907). The cumulative incidence of BO at 5 years was 1.62% (12/691) in bone marrow transplants from related donors (R-BMT), 3.83% (16/424) in peripheral blood stem cell transplantation from related donors (R-PBSCT), 2.91% (24/808) in BMT from unrelated donors (UR-BMT), and 2.65% (5/199) in unrelated cord blood transplantation (CBT). The incidence of BO after R-PBSCT was significantly higher than after any other type of allo-SCT (p=0.02). At BO diagnosis, the mean value of FEV1% decreased to 52.1% from 82.2% pre-transplant. 94% of patients had already developed chronic GVHD before the onset of BO. Risk factors for BO by univariate analysis were R-PBSCT (p=0.019) and preceding chronic GVHD (p=0.000). Twenty eight patients died after developing BO, 21 of respiratory failure. Only one patient died of relapse of primary disease. Overall 5 yr-survival of patients with BO from the time of diagnosis was 46.5%, significantly less than for those without (76.2% from day 335, p=0.000) by semi-landmark analysis. Conclusions: The incidence of BO in CBT recipients was higher than R-BMT recipients and not significantly different with UR-BMT recipients. R-PBSCT recipients who have already developed chronic GVHD have a higher risk for developing BO and need extensive care and repeated PFT examinations.

Abstract: Fractionated cyclophosphamide, vincristine, doxorubicin and dexamethasone (hyper-CVAD) alternating with high-dose methotrexate and cytarabine (MA) therapy is a widely used combination chemotherapy for lymphoid malignancies. Its use in combination with Ribuximab (R) has been shown to be safe and effective for Burkitt's lymphoma (BL) and mantle cell lymphoma (MCL). Acute toxicities of the treatment are mostly hematologic, leading to cases with neutropenic fever and infection. However, long-term toxicities in terms of patients' immune function have not been fully elucidated. Purpose To evaluate the long-term immune function after R-hyper-CVAD/MA therapy, and to compared it with those after R-CHOP therapy. Method Patients who have received R-hyper-CVAD/MA therapy or R-CHOP therapy for more than 6 cycles for B-cell lymphomas (diffuse large B cell lymphoma (DLBCL), BL and MCL) between January 2007 and December 2012 were identified. The patients' immune functions were assessed in regards to white blood cell (WBC) count, lymphocyte count, lymphocyte surface marker analysis and immunoglobulin (Ig) measurement. Result Altogether, 21 patients underwent evaluation. 11 patients received R-CHOP therapy, and 10 patients received R-hyper-CVAD/MA therapy. The mean age of the patients was 57 and 53.3 years-old, respectively. All patients in the R-CHOP group had DLBCL, whereas patients who received R-hyper-CVAD/MA were either treated for BL in 4, MCL in 1 and DLBCL in 5 patients. The median time from the initiation of treatment to the assessment of immune function was 1060 (range 434 – 2475) and 947.5 (range 448 – 1966) days, respectively. In terms of immune function, IgG level, the proportion of CD4, CD8 positive (+) cells, CD4/8 ratio were significantly different between those who received R-CHOP therapy and R-hyper-CVAD/MA therapy. The median IgG level was 910 mg/dL in those who received R-CHOP therapy as opposed to 484 mg/dL in those who received R-hyper-CVAD/MA therapy. 4 patients in the R-CHOP group presented with IgG level below normal range, whereas 9 patients in the R-hyper-CVAD/MA group had an IgG level below normal range (p = 0.011). The median proportion of CD4 and CD8 (+) cells were 32.7%, 24.8% in the R-CHOP group and 23.35%, 41.35% in the R-hyper-CVAD/MA group, leading to the median CD4/8 ratio of 1.22 and 0.5, respectively. The total WBC count, lymphocyte count, proportion of CD19 (+) cells, IgM or IgA levels were within normal limit, and did not significantly differ between groups. Of the patients in the R-hyper-CVAD/MA group, 3 patients required regular immunoglobulin replacement, and 1 patient had 2 episodes of hospitalization within 2 years of last treatment, one due to viral meningitis and another due to pneumonia. None of the patients in the R-CHOP group presented with infection episodes that required hospitalization. Conclusion Our results suggest that patients with aggressive B-cell lymphoma treated with R-hyper-CVAD/MA therapy remain at higher risk of developing hypo-gammaglobulinemia, even after 5 years out of treatment. Despite the recovery of B-lymphocytes, the proportion of CD4 (+) cells remained low in those who received R-hyper-CVAD/MA therapy, leading to an increased proportion of CD8 (+) cells and decreased CD4/8 ratio. Patients who have received R-hyper-CVAD/MA therapy should be closely monitored for their immune function for a longer period due to prolonged hypogammaglobulinemia. Disclosures: Okamoto: Novartis, BMS, GSK, Kirin, Chugai, Alexion: Honoraria, Research Funding.

Abstract: Background: Previous studies have shown that the absolute lymphocyte count (ALC) in peripheral blood at diagnosis may be an independent prognostic factor of IPI for patients with diffuse large B-cell lymphoma (DLBCL). In the rituximab era, the U.S. National Comprehensive Cancer Network International Prognostic Index (NCCN-IPI) was developed to improve the risk stratification of DLBCL in comparison to the existing IPI. Therefore, the aim of this study was to clarify the impact of ALC at diagnosis on event free survival (EFS) and overall survival (OS) on analysis performed with factors included in NCCN-IPI. Patients and methods: We retrospectively reviewed the ALC of 413 patients with newly diagnosed DLBCL treated with R-CHOP at our hospital between January 2005 and March 2013. Primary central nervous system lymphoma patients were excluded from this study. ALC was determined in all patients from complete blood count with differential white blood count at the time of diagnosis, and prior to therapy administration. EFS and OS were estimated according to the Kaplan-Meier method. Multivariate analysis was performed with the proportional hazard Cox model. Results: The median ALC was 1.2x10E9/L (range, 0.06-9.0). We set an ALC cut-point at 1.0x10E9/L based on previous studies. The median follow-up duration was 40 months. Baseline characteristics according to ALC ( 〈 1.0x10E9/L[n=145] and 〉 1.0x10E9/L[n=268]) are summarized in Table1. Patients with ALC 〈 1.0x10E9/L had a significantly poorer EFS and OS than patients with ALC 〉 1.0x10E9/L (5-year EFS, 37.0% versus 68.9%, p 〈 0.001; 5-year OS, 46.3% versus 80.0%, p 〈 0.001). On multivariate analysis performed with factors included in IPI and NCCN-IPI, ALC remained an independent predictor of EFS (IPI: hazard ratio [HR] 1.95; 95% confidence interval [CI] 1.43-2.68; p 〈 0.001, NCCN-IPI: HR 1.94; 95%CI 1.42-2.65; p 〈 0.001) and OS (IPI: HR 2.35; 95%CI 1.61-3.42; p 〈 0.001, NCCN-IPI: HR 2.29; 95%CI 1.57-3.33; p 〈 0.001) (Table2). Importantly, within the poor R-IPI group, ALC distinguished patients with different 5-year EFS (24.4% versus 50.4%, p 〈 0.001) and OS (35.7% versus 65.7%, p 〈 0.001). For the high NCCN-IPI group also, ALC distinguished patients with different 5-year EFS (14.8% versus 39.8%, p 〈 0.01) and OS (17.5% versus 54.5%, p 〈 0.001) (Figure1). Conclusions: According to our results, ALC 〈 1.0x10E9/L is an adverse prognostic factor and independent of IPI and NCCN-IPI. ALC might be more successful in identifying high-risk patients in which IPI and NCCN-IPI analysis was unrevealing. Our results suggest that other therapeutic strategies may be more effective in high-risk patients with DLBCL. Disclosures No relevant conflicts of interest to declare.

Abstract: Abstract 1993 Background: Recent advance in allogeneic hematopoietic stem cell transplantation (HSCT) has led to an increasing use of transplantation for older, having comorbidities, and more heavily pretreated patients. Thus, the evaluation of comorbidity is of increasing importance in order to estimate precisely benefits and risks of transplant procedure. The HCT-CI scoring system has been proposed, and has been confirmed as a useful tool to predict the probability of post-transplant non-relapse mortality (NRM) and overall survival (OS) while taking into account any pre-transplant comorbidity. However, HCT-CI has not been fully validated by prospective studies in multicenter setting, and its predictive value remains to be elucidated in Japanese population. Objectives: The aim of this cohort study is to prospectively evaluate whether HCT-CI score could be predictive for transplant outcomes Japanese patients in multicenter transplant setting. Patients and Methods: 244 consecutive patients (median age 45.3 years; range 16 to 74) with a variety of hematological disorders who had received myeloablative (MA, n= 167) or reduced-intensity (RIST, n= 77) conditioning followed by the first HSCT were enrolled in this study from 2007 to 2009 in 14 facilities of the KSGCT in Japan. The stem cell sources included bone marrow (BM, n=156), peripheral blood (PB, n=30) and cord blood (CB, n=58). The commonest diagnosis was myeloid malignancies (AML/MDS; 58.2 %) followed by ALL (18.4%), and NHL (10.7%). Pre-transplant co-morbidities were scored according to original HCT-CI scoring system in all patients. Patients were followed for 2 years or until relapse/death before two years after transplantation. Results: The most frequent comorbidity observed was mild & severe pulmonary comorbidities (51.9%), followed by mild hepatic comorbidity (14.6%) and active infections (9.3%). The other comorbidities observed in this cohort included psychiatric problems (4.8%), diabetes (3.6%), moderate to severe hepatic (2.8%), and solid tumors (2.8%). The OS and NRM of all patients were 59.2% and 20.4% at 2 years, respectively. The proportion of patients with HCT-CI score 0, 1–2, 3≤ was 32%, 30%, 38%, respectively. The original HCT-CI score failed to predict for OS and NRM (Fig 1). Furthermore, it was not correlated with KS score at 2Y and specific causes of death. We then applied flexible HCT-CI risk scoring system (re-stratified as low-risk; 0–3, and high-risk; 4≤). The distribution of flexible HCT-CI score did not differ significantly in terms of age at the time of transplant (51≤ vs. ≤50), sources of stem cells (BM vs. PB vs. CB), and conditioning (MA vs. RIST). The better predictive capacity for NRM and OS at 2Y of the flexible HCT-CI than original HCT-CI was confirmed (p=0.003) (fig 2). Multivariate analysis identified age (≥50) (p=0.010, HR=1.80), PS (≥90%) (p=0.010, HR=2.57), donor type (HLA-mismatched/unrelated donor) (p=0.020, HR=1.82), and flexible HCT-CI score (4≤) (p=0.019,HR=1.82) as significant predictors of worse OS at 2Y. However, flexible HCT-CI score did not remain as significant predictor of NRM at 2Y while age (p=0.017, HR=2.17), PS (p=0.0001, HR =4.60)and donor type (p =0.019, HR=2.80) remained as significant predictors. Conclusion: Our multicenter prospective study suggests that the modified 2-group scoring HCT-CI, not the original scoring HCT-CI, is a useful predictor of 2 -year OS but not NRM in Japanese. Reconsidering the scoring weight of comorbidity and incorporation of other risk factors (age, PS and donor type) may further improve the predictive capacity of flexible HCT-CI scoring system. Disclosures: Nannya: Otsuka Seiyaku Koujyou: Research Funding.

Abstract: Background: Tyrosine kinase inhibitor (TKI) remains to be the mainstay of treatment for patients with chronic myelogenous leukemia (CML). Second generation TKIs have been shown to successfully treat patients who are resistant or intolerant to imatinib, as well as induce faster and deeper molecular response when used as first line therapy. Since the initial report of the French STIM study (Mahon FX, et al. Lancet Oncol 2010; 11: 1029-35) that showed successful discontinuation of imatinib therapy in approximately 50% of patients who have sustained complete molecular response (CMR), several groups have confirmed that a subset of patients can discontinue TKI for a long period. While factors associated with successful discontinuation are not yet well defined, it also remains an open question whether patients who have failed the initial attempt of discontinuation will have to remain on life-long treatment with TKIs. Patients and Methods: Patients who have been treated at Keio University Hospital for CML, who had been in confirmed stable CMR for over 2 years at the time of study enrollment, and who had no history of accelerated phase/ blastic phase while on treatment with TKI, were eligible to enroll in the study. Patients were monitored monthly for the first 6 months after discontinuation, every 2 months until 12 months, and every 2 to 3 months thereafter. Treatment with a TKI was initiated if the peripheral blood quantitative PCR (TMA method) value exceeded 100 copies. Once the patient was restarted on TKI therapy, and regained sustained CMR for over 2 years, they were allowed to reenter the study and discontinue treatment, upon patients' choice. Results: Sixty-seven patients have been enrolled in the study, of which 53 patients who have been observed for over 2 years since first TKI (imatinib 48; dasatinib 1; nilotinib 4) discontinuation were analyzed. The median age of the patients was 54 (range 28-83) years. Thirty-seven (69.8%) patients were male. In terms of baseline characteristics, 18 (34.0%) had been treated with interferon prior to TKI use, and 41 (77.4%) were CMV positive. The Sokal risk score was low in 34 (64.2%), intermediate in 11 (20.8%) and high in 4 (7.5%) patients. Among the 53 patients, 45 (84.9%) were checked for the existence of BIM deletion, among which 7 (13.2%) patients were positive. The median time on TKI treatment was 98 (range 32-147) months and the median duration of CMR was 38 (range 24-106) months. The median follow-up of the patients at the time of this analysis since study enrollment was 61 (range 26-66) months. Treatment was restarted in 28 (45%) patients (imatinib 7; dasatinib 20; nilotinib 1). While this occurred within the first 6 months of treatment discontinuation in most patients, 6 patients were started on treatment beyond 12 months of drug-free survival (DFS) (at month 14, 20, 23, 36, 36, and 52, respectively). Five patients presented with a fluctuating copy number early after TKI discontinuation, whereas 1 patient only became positive for bcr-abl after 30 months of treatment discontinuation. The estimated 24-months DFS was 52.8% (95% confidence interval (CI) 39.5-65.8%) (Fig 1). All patients have restored CMR at least at one occasion after recommencing TKIs. No single factor was significantly associated with success of discontinuation. Among the patients who had sustained CMR for over 24 months after re-initiation of TKI, 10 patients elected to challenge discontinuation of TKI for the second time. All patients were on dasatinib at the time of discontinuation. The median age of these patients was 58.5 (range 31-75) years. The median time on TKI prior to second discontinuation was 33 (range 26-45) months and the median duration of CMR after treatment re-initiation was 26.5 (range 25-44) months. All but one patient were restarted on treatment at the time of the analysis (median observation 26 (range 13-35) months), leading to a DFS of 20% (95% CI 5.0-54.1%) at 12 and 24 months (Fig 1). Conclusion: Long-term observation of the outcome of TKI discontinuation in CML patients who had sustained CMR for over 2 years showed cases of late relapses as well as small chance of success on the second attempt of TKI discontinuation even with the use of second generation TKIs. While the result of first discontinuation was similar to previous reports, attempt of second discontinuation was less successful compared to the French group, despite changing the drug of use from imatinib to dasatinib. Figure 1 Figure 1. Disclosures Matsuki: Nippon Shinyaku: Honoraria; Celgene: Honoraria; Bristol-Myers Squibb: Honoraria. Sakurai:Celgene: Honoraria. Karigane:Celgene: Honoraria. Kikuchi:Celgene: Honoraria; Takeda Pharmaceutical Company: Honoraria; Kyowa Hakko Kirin: Honoraria. Yokoyama:BMS: Research Funding. Okamoto:Eisai Co., Ltd.: Research Funding; Chugai Pharmaceutical Co., Ltd.: Research Funding; Nippon Shinyaku Co., Ltd.: Research Funding; Alexion Pharmaceuticals, Inc.: Research Funding; Astellas Pharma Inc.: Research Funding; Toyama Chemical Co., Ltd.: Research Funding; Otsuka Pharmaceutical Co., Ltd.: Honoraria, Research Funding; Kyowa Hakko Kirin Co., Ltd.: Research Funding; Shionogi & Co., Ltd.: Research Funding; Bristol-Myers Squibb K.K.: Honoraria, Research Funding; Asahi Kasei Pharma Corp.: Research Funding; Sumitomo Dainippon Pharma Co., Ltd.: Research Funding; Teijin Pharma Limited: Research Funding; Pfizer Inc.: Honoraria, Research Funding; JCR Pharmaceuticals Co., Ltd.: Research Funding.

Abstract: Background: Among tyrosine kinase inhibitors (TKIs), imatinib mesylate was the first TKI successfully used for the treatment of chronic myelogenous leukemia (CML) in chronic phase (CP), and a majority of patients still remains on its long-term treatment. Although imatinib has been well tolerated in clinical practice, and the side effect profile has usually been mild to moderate, there are limited data available regarding the long-term TKIs treatment on kidney function and associated complications such as anemia. This study aimed to evaluate the effect of long-term imatinib treatment on estimated glomerular filtration rate (eGFR) and hemoglobin (Hb) level in patients with CML in CP. Patients and methods: By using the institutional database, patients with CML in CP who had been on imatinib as the first and the only TKI for over 5 years at Keio University Hospital (Tokyo, Japan) were selected and retrospectively analyzed. Estimated GFR was calculated by the Modification of Diet in Renal Disease equation for Japanese defined by the Japanese Society of Nephrology. All statistical analyses were performed with EZR, which is a graphical user interface for R. Results: Eighty-two patients were evaluable. The median age at initiating imatinib was 49.5 years (range, 19-76). The median duration of imatinib treatment was 105 months (range, 60-170). During this study period, imatinib was given at a dose of 400mg/day. The dose reduction was indicted in 8 patients but not because of progressing renal impairment. The mean eGFR was 77 ml/min/1.73m2 (range, 38-120), and the value was below 60 ml/min/1.73m2 in 12 patients before initiating imatinib. The mean value significantly decreased to 62 ml/min/1.73m2 (range, 34-98) over the 5 years after imatinib treatment (P 〈 0.001), and the values reached 〈 60 ml/min/1.73m2 in 43 of 82 patients (P 〈 0.001). In an univariate analysis of patients excluding 12 patients with below 60 ml/min/1.73m2, older age and lower eGFR value at the initiation of imatinib were associated with later development of chronic kidney disease ( 〈 60 ml/min/1.73m2) (P 〈 0.001 and 0.002, respectively). Mean Hb level at 5 years after starting imatinib was significantly lower as compared with that before (12.9+1.7 g/dl vs. 12.4+1.3 g/dl, P 〈 0.01). The declining rate of eGFR was negatively correlated with those of Hb levels (correlation coefficient -0.249, P 〈 0.05). In 20 patients with low Hb level (9.7+1.2 g/dl) and renal dysfunction, median serum erythropoietin (EPO) level was 31.9 mIU/ml (range, 9.1-119). Furthermore, 11 patients with eGFR 〈 60 ml/min/1.73m2 achieving a durable molecular remission took part in an institutional TKI discontinuation trial. At 1 year after discontinuing imatinib, their mean eGFR values significantly improved (50.0 + 6.5 to 56.0 + 10.2 ml/min/1.73m2, P 〈 0.05) as well as Hb level (12.0 + 1.7 to 14.0 + 1.6 g/dl, P 〈 0.01). Conclusion: Our findings indicated that long-term use of imatinib is frequently associated with reversible but continuous decline in eGFR level, which could lead to anemia partly due to inadequate production of EPO. Although the degree of nephrotoxicity is usually mild, close monitoring of renal function is recommended particularly in older patients with pre-existing renal dysfunction. Disclosures Sakurai: Celgene: Honoraria. Karigane:Celgene: Honoraria. Matsuki:Bristol-Myers Squibb: Honoraria; Celgene: Honoraria; Nippon Shinyaku: Honoraria. Kikuchi:Celgene: Honoraria; Takeda Pharmaceutical Company: Honoraria; Kyowa Hakko Kirin: Honoraria. Okamoto:Nippon Shinyaku Co., Ltd.: Research Funding; Alexion Pharmaceuticals, Inc.: Research Funding; Otsuka Pharmaceutical Co., Ltd.: Honoraria, Research Funding; Asahi Kasei Pharma Corp.: Research Funding; Teijin Pharma Limited: Research Funding; Kyowa Hakko Kirin Co., Ltd.: Research Funding; Shionogi & Co., Ltd.: Research Funding; Astellas Pharma Inc.: Research Funding; Toyama Chemical Co., Ltd.: Research Funding; Eisai Co., Ltd.: Research Funding; Sumitomo Dainippon Pharma Co., Ltd.: Research Funding; Chugai Pharmaceutical Co., Ltd.: Research Funding; Bristol-Myers Squibb K.K.: Honoraria, Research Funding; Pfizer Inc.: Honoraria, Research Funding; JCR Pharmaceuticals Co., Ltd.: Research Funding.

Abstract: Abstract 2035 Background: Recently, we described that adult AML patients with granulocytic sarcoma (GS) at diagnosis presented unique characteristics including younger age, higher WBC counts, and higher frequency of French-British-American M4 and M5 morphology than those without GS. Furthermore, GS adversely affected the relapse rate and disease free survival. However, the appropriate therapeutic strategy, especially indication of allogeneic hematopoieteic stem cell transplantation (allo-HSCT), has remained unclear. Here, we present a large-scale retrospective analysis of 503 adult AML patients who underwent allo-HSCT that compared the clinical characteristics and treatment outcomes of patients with GS to those without GS. Patients and Methods: This study included 503 consecutive adult AML patients (median age 44 (15–73), male/female; 301/202), excluding patients with acute promyelocytic leukemia, who received allo-HSCT for the first time between January 2000 and December 2008 at 9 institutions participating in the Kanto Study Group for Cell Therapy (KSGCT). GS group was consisted of patients in whom GS occurred any time in the clinical course until transplantation, such as at diagnosis, relapse, and disease progression. The c2-test was used for comparison of binary variables. The Mann-Whitney U test was used for comparison of continuous variables. overall survival (OS) and leukemia free survival (LFS) were estimated by the Kaplan-Meier method, and compared using the log-rank test. The Cox proportional hazards regression model was used for multivariate analysis of prognostic factors. P 〈 0.05 was considered as statistically significance. Results: Of the 503 patients, 44 patients (8.7%) had GS before transplantation. All factors including patients' characteristics and transplantation procedure were not significantly different in each group, but the patients in GS group were significantly younger (GS: median 34 years; range, 17–60 years vs. non-GS: median 46 years; range, 15–73 years; p=0.002). Treatment outcomes, GS vs. non-GS: The cumulative incidences of grade II-‡W acute GVHD, chronic GVHD, and NRM at 1 year were not significantly different in each group. The 5-year OS rate did not differ significantly between the GS and non-GS groups (47% vs. 44%, respectively; p=0.621). The 5-year LFS rate also did not differ significantly (41% vs. 38%, respectively; p=0.646). Since patients in GS group were significantly younger as mentioned above, we performed a subgroup analysis on 336 patients aged 50 or younger; 37 patients in GS group and 298 in non-GS group. Age was not significantly different between the two groups. In this subgroup, too, there was no significant difference in both the 5-year OS rate and the 5-year LFS rate. Treatment outcomes in GS group: Patients in CR at allo-HSCT achieved the significantly superior 5-year OS rate and LFS rate compared to those in non-CR, (OS: 75% vs. 24%, respectively; p=0.002), (LFS: 60% vs. 24%, respectively; p=0.046). In contrast, the prognosis of 10 patients who performed allo-HSCT with accompanying GS was dismal; 9 patients experienced a relapse within 6 months. 5-year OS rate and LFS of 8 patients, who received prior local irradiation for GS, did not significantly differ from those who did not. Both acute and chronic GVHD did not significantly affect the OS and LFS rates, respectively. Twenty-four patients (55%) experienced a relapse after allo-HSCT. Regarding the sites of relapse, only GS, only bone marrow, and both bone marrow and GS were involved in 1 patient (4%), 13 (54%), and 10 (42%), respectively. In 10 patients (91%) out of 11 with GS at relapse, GS occurred at the same site between before and after allo-HSCT. Multivariate analysis identified age and disease status as independent prognostic factors for OS. Conclusion: Patients with GS, who treated with allo-HSCT, could achieve the comparable prognosis to those without GS despite poor prognosis of GS. Notably, an excellent treatment outcome was brought about in patients received allo-HSCT in CR. This study suggests that allo-HSCT during an appropriated period may overcome this unfavorable disease. Disclosures: No relevant conflicts of interest to declare.

Abstract: Introduction: Allogeneic hematopoietic cell transplantation (HCT) has curative potential for a variety of hematologic malignancies. However, the success of HCT is heavily dependent on the disease and remission status. Armand et al. recently proposed a disease risk index (DRI) to assess the risk for allogeneic HCT based on the disease and remission status, and could be used to risk stratification on survival (Armand et al Blood 2012), but evaluation of its clinical significance is limited. HCT-comorbidity index (HCT-CI) was developed as a measure of pretransplant organ dysfunctions (Sorror et al Blood 2005), which has been shown to be associated with nonrelapse mortality (NRM). In order to clarify any association between pretransplant factors including DRI and HCT-CI, and respective overall survival (OS), NRM and relapse rate, we retrospectively reviewed the data of patients who underwent allogeneic HCT at our department. Patients and Methods: A total of 305 patients with hematologic malignancies who underwent initial allogeneic HCT at our department between January 2000 and April 2013 were included. After excluding patients with insufficient HCT data, we included a total of 244 patients (138 male, 106 female) with a median age of 49.5 (range 15-69) years. A total of 133 patients received myeloablative conditioning and 111 received reduced-intensity regimens. Stem cell sources were bone marrow (n=177), peripheral blood (n=32), combined peripheral blood and bone marrow (n=1), and cord blood (n=34). A total of 146 patients received tacrolimus-based regimens and 98 patients received cyclosporine-based regimens for GVHD prophylaxis. The DRI has four risk-based categories (low, intermediate, high, and very high) and the HCT-CI has three categories in order of ascending risk (0, 1-2 and ≥3). OS was calculated with the Kaplan-Meier method, compared among groups with the log-rank test, and multivariable Cox regression analyses were used to evaluate factors associated with OS. The cumulative incidence of NRM and relapse were calculated while treating relapse and death without relapse, respectively, as competing events, and competing risk regression analyses were used to evaluate risk factors associated with NRM and relapse. Results: The median follow-up for survivors was 4.7 years (range 0.1-14.2 years). Pretransplant disease risks in the DRI low, intermediate, high, and very high risk groups were 8%, 60%, 25%, and 7%, and 4-year OS in the same groups were 74%, 64%, 35%, and 12%, respectively (Figure 1, p 〈 0.001). Four-year OS among patients with a HCT-CI of 0, 1-2, and ≥3 were 63%, 52%, and 41%, respectively. Multivariable analysis showed a significant association with OS for DRI (high risk hazard ratio [HR] 2.62, p 〈 0.001; very high risk HR 5.26, p 〈 0.001 versus intermediate risk), HCT-CI (HCT-CI ≥3 HR 1.64, p=0.022 versus HCT-CI 0-2), 2-4 performance status (HR 3.10, p 〈 0.001), and donor-recipient ABO minor-mismatch (HR 2.00, p=0.005 versus ABO match). The cumulative incidence of 4-year NRM was 25%, and NRM was significantly associated with HCT-CI (HCT-CI ≥3 HR 2.27, p=0.003 versus HCT-CI 0-2) and 2-4 performance status (HR 3.89, p 〈 0.001). The cumulative incidence of 4-year relapse was 25%, and relapse was significantly associated with DRI (high risk HR 2.26, p=0.012; very high risk HR 8.11, p 〈 0.001 versus intermediate risk). Finally, we reclassified all patients into four risk groups incorporating DRI and HCT-CI: DRI low-intermediate plus HCT-CI 0-2 (group I), DRI low-intermediate plus HCT-CI ≥3, or DRI high plus HCT-CI 0-2 (group II), DRI high plus HCT-CI ≥3 (group III), and DRI very high (group IV). Four-year OS among patients with group I, II, III, and IV were 68%, 47%, 25%, and 12%, respectively (Figure 2, p 〈 0.001). Conclusions: Our results suggest that risk stratification with DRI and HCT-CI for the prognosis of relapse and NRM may be useful for patients undergoing allogeneic HCT. Larger and prospective studies are warranted to more precisely validate these findings. Figure 1 Figure 1. Figure 2 Figure 2. Disclosures No relevant conflicts of interest to declare.