ALBERT

Description: Background: Umbilical cord blood transplantation (UCBT) is a potentially curative therapy acute leukemia (AL) patients. Transplantation benefit must be balanced against risks, such as transplant related mortality and relapse. The complex nature of hematopoietic stem cell transplantation data (HCT), rich in interactions and possibly nonlinear associations, has motivated us to apply machine learning (ML) for predictive modeling. ML is a field of artificial intelligence and is part of the data mining approach for data analysis. Our group has recently reported on a ML based prediction model for short term HCT outcomes (Shouval R et al; JCO 2015). Using a ML algorithm, the perspective of the current study was prediction of leukemia free survival (LFS) at 2 years after an UCBT, while exploring variables' importance and interactions. Patients & Methods: A cohort of 3,149 UCBT were analyzed. Inclusion criteria encompassed patients at all ages, undergoing an UCBT (single/double unit) in EBMT centers from the year 2004 to 2014, for AL, in all disease status. All conditioning and graft versus host disease prophylaxis regiments were included. A total of 24 variables were considered, including the number of total nucleated cell dose (TNC), donor and recipients HLA typing, as well as recipient, disease and transplant characteristics. The Random Survival Forest (RSF) ML algorithm was applied for model construction and data exploration. RSF is known to be adaptive to data, is able to automatically recover nonlinear effects and complex interactions among variables, and yields nonparametric prediction over test data. The analysis pipeline consisted of prediction model development, assessment of variable importance by their minimal depth from the tree trunk, and exploration of the top ranking variable with dependence plots. The latter promotes understanding of non-trivial associations between variables and outcomes. Results : The 2 years LFS was 49%, with a median follow up of 30 months. A RSF model of 1000 trees was developed, with each tree constructed on a bootstrap sample from the original cohort. A prediction error of 36.0% was calculated. The 10 most predictive variables (in ascending order) were disease status, age, TNC harvested and infused, recipient CMV serostatus, interval from diagnosis to UCBT, transplant year, previous autologous transplant, and use of anti-thymocyte globulin (ATG). Selected findings from exploration of variables-outcome relationship with dependence plots included a varying effect of TNCs in specific subpopulations. Increasing the number of infused TNCs had a positive effect on predicted LFS in patients receiving HLA mismatched (2 or more HLA mismatch) (figure) or single unit CB grafts, and patients in earlier disease status or older age. ATG administration was associated with worse LFS, whether unadjusted or adjusted to all other variables. However, there was an additional negative effect in advanced disease status patients, recipients of HLA mismatched or single CB units grafts, and older patients. Patients in 1st complete remission (CR) had higher predicted LFS as compared to those in 2nd CR. However, in patients receiving a HLA mismatched or a double CB graft, the difference in LFS between CR1 and CR2 was attenuated. Younger age had a favorable impact in early disease status, but lost its positive effect in advanced disease. Conclusions: A prediction model for LFS 2 years post UBCT was developed using the RSF ML algorithm. Variables were ranked according to their predictive contribution. Disease status, age, and TNC count were found to be the most important factors. Dependence plots revealed interactions and nonlinear associations between variables and the outcome, such as the effect of cell dose on HLA disparity. Apart from the study's clinical findings, it carries a methodological significance. A novel ML approach for prediction, variable selection and data exploration, accounting for long term time to event outcomes, has proved useful in the field of HCT. Figure 1. Variable marginal dependence coplot of predicted LFS at 2 years against TNC, conditional on HLA matching. Individual cases are marked with blue circles (alive or censored) and red `x's (event). Linear smooth (a linear extrapolation of the prediction function), with shaded 95% confidence band, indicates trends of variable dependence. Figure 1. Variable marginal dependence coplot of predicted LFS at 2 years against TNC, conditional on HLA matching. Individual cases are marked with blue circles (alive or censored) and red `x's (event). Linear smooth (a linear extrapolation of the prediction function), with shaded 95% confidence band, indicates trends of variable dependence. Disclosures Mohty: Janssen: Honoraria; Celgene: Honoraria. Sanz:JANSSEN CILAG: Honoraria, Research Funding, Speakers Bureau. Bader:Neovii: Other: Institutional grants; Medac: Other: Institutional grants; Riemser: Other: Institutional grants; Amgen: Consultancy; Novartis: Consultancy; Jazz Pharmaceuticals: Consultancy.

Description: Relapse remains a major barrier to the long-term overall survival (OS) of pediatric patients with acute lymphoblastic leukemia (ALL). Allogeneic HSCT is a potentially curative option for these patients and its mechanism relies on graft-versus-leukemia (GvL) and delivery of high-dose chemotherapy. However, the mechanism and degree of GvL in curing ALL is not fully understood. Unrelated cord blood (CB) is an attractive donor option for these patients due to rapid procurement and lenient HLA matching without increased graft-versus-host disease (GvHD). Despite less GvHD, relapse is not increased when compared to other donor sources. Therefore, through collaboration between Eurocord, PDWP-EBMT and Duke University, we performed a retrospective analysis to identify risk factors associated with relapse, and the role of GvHD (as time dependent co-variate) in preventing relapse after unrelated CB transplantation (CBT) for children with ALL. We analyzed 640 children (30 months) was associated with improved OS and LFS (Table 2). Conversely, the use of ATG was associated with lower OS. In CR2 patients, the CI of relapse was 28%. In MVA, longer duration from diagnosis to CBT (〉30 months) and receiving TBI were both associated with less relapse. Importantly, receiving a fully matched HLA CB graft was a strong risk factor for increased relapse compared to mismatched CB recipients. Acute GvHD (II-IV) in this cohort was associated with higher mortality but not with relapse. In this large retrospective collaborative study, we investigated the impact of GvHD as a marker of GvL in pediatric ALL patients after CBT. In CR1 patients, acute GvHD and use of TBI were protective of relapse. However in CR2 patients, use of TBI, longer duration from diagnosis to CBT and HLA mismatched grafts were associated with decreased incidence of relapse but not with the presence of GvHD. These results showed that the impact of GvHD as a GvL marker is more evident in CR1 patients than CR2. Whereas, the impact of HLA mismatch CB on relapse is more important in CR2 patients. Decreasing immunosuppression more rapidly for CR1 patients or avoiding HLA identical CB for CR2 patients should be further investigated. In both cohorts, TBI has shown to be an important protective factor for relapse. Disclosures Bader: Riemser: Other: Institutional grants; Neovii: Other: Institutional grants; Medac: Other: Institutional grants; Amgen: Consultancy; Novartis: Consultancy; Jazz Pharmaceuticals: Consultancy.

Description: Infant acute leukemia has poor prognosis due to both chemo-resistance and toxicity in these very young patients (pts). Allogeneic stem cell transplantation (HSCT) is indicated for all pts with infant AML and ALL beyond first complete remission (CR). While some cooperative protocols exist for infant ALL with clear criteria for HSCT in CR1 pts, there is no international consensus for HSCT in infants with CR1 AML. In this population of low body weight pts, with aggressive disease, unrelated cord blood (UCB) is an attractive stem cell source due to high cell content and prompt donor availability. We retrospectively analyzed 254 children with ALL (n=154) or AML (n=95) diagnosed within 1 year (y) of age (median age 5.7 months (range 0.03-12)). Median age at UCBT was 1.15 y (range 0.3-11). Pts received single-unit unrelated UCBT after myeloablative conditioning regimen (MAC) from 1995 to 2012 in EBMT centers. Pts were transplanted in CR1 (n=135, 54%), CR2 (n=79, 32%) or more advanced disease (n=35, 14%). Data on cytogenetic at diagnosis were available for 182 pts (ALL=111, AML= 71), 24% were intermediate risk and 76% poor risk. For pts with ALL, 82 had chromosome 11 abnormalities defined as (t(4;11)-n=47, 11q23 abnormalities not otherwise specified-n=25, and others-n=10). For pts in CR1 median time from diagnosis to UCBT was 6 months (range 2.8- 20.7 months) and median duration of remission prior to transplant was 127 days (range 15-326 ). CB units (21%) were identical to recipient, 57% had 1 HLA disparity (antigen level for HLA-A and B, allelic level for DRB1), and 22% had 2 HLA disparities. Median TNC dose at infusion was 9.4x107/kg (range 1- 18x107/kg). Most pts received busulfan (Bu) based conditioning (75%, n=186). The most frequent regimens were Bu+cyclophosphamide (Cy) (n=68), Bu+Cy+Melphalan (n=43), Bu+Cy+Fludarabine+Thiotepa (n=31). Thirty one pts received conditionings that included TBI. ATG was used in 79% of pts. GVHD prophylaxis consisted either of CSA±steroids or CSA±MMF in 72% and 11% of pts, respectively. Median follow-up was 42 (range 3-170) months. Cumulative incidence (CI) of day 60 neutrophil recovery was 89±2%, with median time of 21 (range 7-55) days. CI of engraftment was 88% for those receiving TNC9.4x107/kg, p=0.006. At day 100, 90% of pts had full donor chimerism (data available for 73% of pts who engrafted). Of the 28 pts who had graft failure, 11 died within 30 days after UCBT, 7 had autologous recovery, and 10 received a subsequent HSCT (3 autologous, 7 allogeneic). Of the 7 pts with a second allo-HSCT, 4 are alive at a median of 2 years. CI of day 100 acute GVHD (grade II-IV) was 39±3%, median time of onset was 18 days (range 6-100). Among 96 pts who experienced grade II-IV acute GVHD, 58 had grade II, 24 had grade III, and 14 had grade IV. CI of 1 y chronic GvHD was 10±3%. CI of 1y TRM was 25±3%. CI of TRM was higher for infants transplanted for ALL than for AML (31% and 15% , respectively, p= 0.004. One hundred and fifteen pts died. The causes of death were relapse (n=55), infection (n=30), GvHD (n=9), or other transplant-related events (n=21). CI of 4y relapse was 26±4%. It was 28% for pts transplanted for ALL and 21% for AML, p=0.32. CI of 4y relapse was 19%, 31% and 50% for pts transplanted in CR1, CR2 or more advanced disease, p=0.007. Age at diagnosis less than 3 months was associated with a trend for higher CI of relapse at 4y (36% versus 22%, p=0.06). Leukemia-free-survival (LFS) at 4y was 49±3%; it was 40% and 69% for those transplanted for ALL and AML, respectively p=0.001. According to diagnosis and disease status, 4y LFS for ALL was: 50% for pts in CR1, 36% for CR2 and 0% for advanced disease, (p=0.001). For AML it was: 82% for pts in CR1, 58% for CR2 and 11% for advanced disease, (p=0.001). LFS was better for pts transplanted with a short delay from diagnosis to transplant (within 6 months of diagnosis, 60% versus 45%, p=0.01). For pts with chromosome 11 abnormalities, 4y LFS was 45%. In multivariable model, diagnosis of ALL (HR 1.8, p=0.004), advanced disease status at UCBT (HR 2.4, p

Description: Bone marrow failure (BMF) syndrome is a group of rare hereditary or idiopathic disorders occurring at all ages. For BMF patients under 40 years of age and having an HLA-identical related donor, allogeneic bone marrow transplantation (HSCT) is indicated. The use of cord blood transplantation (CBT) from an HLA-identical sibling donor is of interest in non-malignant disease due to the low risk of graft-versus-host disease (GVHD) associated with this type of stem cell source and the absence of risk to the donor. We analyzed outcomes of 122 children and young adults with inherited or acquired BMF syndrome, who received a CBT from an HLA-identical related donor. Patients were transplanted in EBMT centers between 1988 and 2014 and reported to Eurocord and EBMT. Ninety-six patients had an inherited and 26 patients had an acquired BMF. The specific diagnosis were: Fanconi anemia (n=48), Diamond Blackfan anemia (n=25), Amegakariocytic thrombocytopenia (n=6), Kostmann syndrome (n=4), Dyskeratosis congenital (n=3), Schwachman-Diamond syndrome (n=2), and unclassified inherited BMF (n=8) for the inherited group and severe aplastic anemia (n=25), pure red cell aplasia (n=1) for the acquired group. Eighty-nine patients received a single cord blood (CB) unit, whereas 33 patients received a combination of CB and bone marrow (BM) cells from the same sibling donor. The main reason for adding BM was due to the low cell dose of the single CB unit alone. Of patients receiving CB+BM, 6 had an acquired and 24 had an inherited BMF. Median age at CBT was 6.7 (range 1-16) years (5.6 years for patients with acquired and 6.9 years for those with inherited BMF). Median interval between diagnosis and CBT was 21 (range 2-114) months for the acquired and 55 (range 2-160) months for the inherited group, respectively. Fifty-nine patients (53%) received a reduced intensity conditioning (RIC) and 52 (47%) were treated with myeloablative regimens (MAC) (data available for 111 of 122 patients). Fludarabine-based regimens were used in 54 patients (44%). For patients receiving a RIC the most common regimen was cyclophosphamide and fludarabine (56%), and for MAC, cyclophosphamide and busulfan (46%). Total body irradiation was used in 8 patients (both in RIC and MAC settings). GVHD prophylaxis consisted mainly of cyclosporine alone (CSA) in 53 patients (43%), CSA + micophenolate mofetil in 11 patients (9%) and CSA + methotrexate in 23 patients (19%). The median number of total nucleated cells (TNC) infused was 6.2x107/Kg (1-24.2x107/Kg) for patients receiving a single CB unit and 26.8x107/Kg (5.3-41x107/Kg) for those receiving CB+BM (median TNC was 3x107/Kg in CB, and 22x107/Kg in BM graft). Median follow up is 5 (range 0.6-25) years. The cumulative incidence (CI) of neutrophil recovery at day 60 was 91%+9%, and the median time to engraftment was 21 (range 7-100) days. The CI of platelets engraftment at day 180 was 89+3% with a median time of 35 (range 7-138) days. Thirteen patients experienced primary graft failure (PFG) (12 patients after a single CB and one after a CB+BM graft). For those patients with available information, treatment of graft failure was second BM-HSCT (n=6), and growth factors (n=1). Of the 13 patients with graft failure, 11 died within in a median time of 1.9 (0.3 to 6) months after CBT. Two patients are alive after the second HSCT. The 100-day CI of grade II-IV acute GVHD was 11+3% (15 patients: grade II, n=8; grade III, n=6; grade VI, n=1), and the 1-year CI of chronic GVHD was 12+4%. The 1-year CI of transplant-related mortality was 13+3%, 17 patients died: 11 due to PGF (infections (n=6) hemorrhagic syndrome (n=1) and other causes (n=4)) and the remaining 6 of acute respiratory distress syndrome (n=1), and infections (n=5). Five-year overall survival (OS) for the whole population was 87+3%. OS was 80+8% for patients with acquired BMF, 87+3% for Fanconi anemia patients, 95+4% for those with Diamond Blackfan, and 86+7% for other inherited BMFs. In pediatric and young adult patients with either inherited or acquired BMF syndrome, CBT from an HLA-identical sibling donor is associated with excellent long-term outcomes with particular low rates of both acute and chronic GVHD, this latter observation being of particular importance for patients with non-malignant disorders. In case of inherited BMF, collecting cord blood unit at the birth of a new sibling is strongly recommended. Disclosures Dufour: Pfizer: Consultancy.

Description: Key Points UCBT is a suitable option for children with JMML, being able to cure a relevant proportion of patients. Because disease recurrence remains the major cause of treatment failure after UCBT, strategies aimed at reducing relapse are desirable.

Description: Fanconi anemia (FA) is a DNA repair syndrome characterized by bone marrow failure, congenital abnormalities and cancer predisposition. Based on previous experimental results showing the in vivo proliferative advantage of gene corrected FA patients' hematopoietic stem cells (HSCs; Rio, Navarro et al. Blood 2017) a gene therapy trial in non-conditioned FA-A patients was initiated in 2016. Six patients have been treated to-date using fresh and cryopreserved CD34+ cells mobilized to peripheral blood with G-CSF and plerixafor, and transduced with the PGK-FANCA.Wpre* lentiviral vector. Cell doses infused in four patients with a follow-up of at least 12 months varied from 0.6 to 1.4 million CD34+ cells/kg. Transduction efficacies of these samples, determined as vector copies per cell, ranged from 0.17 to 0.53 copies/cell. Despite the absence of patients' conditioning, a marked in vivo expansion of gene-corrected cells was observed in all hematopoietic cell lineages analyzed in BM and PB. Significantly, up to 44% of corrected cells were determined in total PB cells at the most recent follow-up visit (24 month) in the first treated patient. Insertion site analyses in PB cells showed an oligoclonal pattern of hematopoietic reconstitution, and revealed engraftment of multipotent corrected HSCs and no evidence of insertion-site mediated clonal expansion. Functional studies showed significant increases in the resistance of BM progenitors to mitomycin C in all treated patients. Additionally, patients with higher levels of corrected cells also showed significant increases in the chromosomal stability of T cells exposed to diepoxybutane. Finally, analyses discriminating the presence of corrected and uncorrected PB cells in these patients showed marked increases in the total number of corrected leukocytes, contrasting to progressive decreases of uncorrected cells. Our studies demonstrate for the first time that lentiviral-mediated gene therapy results in progressive engraftment and phenotypic correction of HSCs in non-conditioned FA patients, suggesting that this gene therapy approach may constitute a low-toxicity option for the treatment and prevention of BMF in patients with FA. Disclosures Bueren: Rocket Pharmaceuticals Inc: Consultancy, Equity Ownership, Patents & Royalties, Research Funding. Navarro:Rocket Pharmaceuticals Inc: Equity Ownership, Patents & Royalties, Research Funding. Segovia:Rocket Pharmaceuticals Inc: Consultancy, Equity Ownership, Patents & Royalties, Research Funding. Casado:Rocket Pharmaceuticals Inc: Patents & Royalties. Schwartz:Rocket Pharmaceuticals: Employment, Equity Ownership. Schmidt:GeneWerk GmbH: Employment; German Cancer Research Center: Employment; bluebird bio: Consultancy. Rio:Rocket Pharmaceuticals Inc: Equity Ownership, Patents & Royalties, Research Funding. Sevilla:Rocket Pharmaceuticals Inc: Honoraria, Patents & Royalties.

Description: Background: Inotuzumab ozogamicin (InO) is an anti-CD22 antibody linked to calicheamicin, approved for relapsed/refractory (R/R) CD22+ B-cell precursor adult acute lymphoblastic leukemia (BCP-ALL) at starting dose of 1.8 mg/m2/course. A phase I study in pediatric patients (pts) is being performed and sponsored by Erasmus MC in collaboration with Pfizer, registered in the Dutch Trial Registry (NTR57360). Study design: Children aged 1-18 years with R/R CD22+ BCP-ALL (ALL) were included after IRB approval and informed consent was obtained. Key inclusion criteria consisted of M2/M3 marrow and adequate liver and kidney function. The study aimed to determine the recommended phase 2 dose (RP2D) of single agent InO; secondary objectives included safety, response and pharmacokinetics (PK). Dose escalation followed the Rolling-6 design, with dose-limiting toxicity (DLT) evaluation during course 1 and defined as delay in hematological (hem) recovery after D42 in responding pts and grade (gr) ≥3 non-hem toxicities 〉48 hours (or 〉7 days for liver tests). DL1 was 1.4 mg/m2 (0.6-0.4-0.4 mg/m2). Overall response rate (ORR) was defined as CR, CRp (ANC 〉0.5x10.9/l but PLT ≤50x10.9/l) and CRi (ANC ≤0.5x10.9/l and/or PLT ≤50,000/µL). Central minimal residual disease (MRD) analysis by flow cytometry and RQ-PCR was considered negative if

Description: Donor lymphocyte infusions (DLI) are a therapeutic approach broadly used in relapse post Allogeneic Hematopoietic Cell Transplant (Allo-HCT). In CML relapse, DLI have proven very effective, but in other hematological malignancies its effectiveness has been reported rather erratic and poor; hence the need of a better understanding of factors influencing outcomes. Questions regarding the CD3 dose, the use of G-CSF mobilized DLI or the potential toxicity associated to the use of DLI from mismatched donors remain open. We report a cohort of 66 patients who, within a cell therapy program (Banc de Sang i Teixits, Barcelona), consecutively underwent allo-HCT and received DLI for relapse. Disease diagnosis was as follows: Lymphoproliferative malignancy (LPD) 29 patients AML and high-risk MDS 24 patients and ALL 13 patients (CML patients were not included). Median age at allo-HCT was 48 years (5-68). The combination female-donor and male-patient was 21%. 27% allo-HCT were T-cell depleted and 29% received a myeloablative conditioning regimen. Immunosupression was CsA-based in 83% of the allo-HCT. Transplants from HLA fully match donors were 54 (81%) and 19% were from mismatch donors. Indication for DLI was morphological relapse in 61 patients (92%) and disease detected by flow cytometer or at a molecular level in 5 patients (8%). 32 (48%) DLI were obtained at stem cell collection day (G-CSF mobilized) and cryopreserved, whereas 34 (52%) DLI were obtained by lymphapheresis. The median follow-up was 198 days (9-4246). The estimated 1-year OS was 60%. A total of 100 DLI were infused, with a median of 1.5 DLI/patient and a median time from allo-HCT to DLI of 303 days (70-5153). Median CD3+ total dose was 2x107 CD3+/Kg (first DLI median dose was 1x107 CD3+/Kg). The time interval from allo-HCT to DLI ≥10 months (Log-Rank 3.64, p=0.056) was associated to better survival. In line with this, there was a trend in patients relapsing ≥ 9 months post allo-HCT (median relapse date) for better survival (Log-Rank 3.33, p=0.068). Twenty-six patients (42%) developed GvHD post-DLI. In 17 patients overall grade was 2-4, which was not associated to poorer outcome. Lately, 8 patients developed extensive chronic GvHD. The development of chronic GvHD was associated to a better survival (Log-Rank 6.07, p=0.014). A total dose ≥1x107CD3+/Kg was associated to a better survival (Log-Rank 4.78, p=0.029) compared to total lower doses. The achievement of complete remission post DLI was also associated to better OS (Log-Rank 4.54, p=0.33). Ten patients died due to non-relapse mortality causes and twenty-seven due to disease progression. Of interest, we found a trend on outcomes when using non G-CSF mobilized DLI compared to-G-GCF mobilized DLI (Log-Rank 2.65, p= 0.104). However, the administration of a pre DLI debulky therapy was not associated with better outcomes. Variables with p