ALBERT

Description: AN and MM are co-senior authors. Background. Fludarabine plus busulfan (FB) and fludarabine plus melphalan (FM) are two widely used reduced-intensity conditioning (RIC) regimens for allogeneic hematopoietic stem cell transplantation (allo-SCT). Patients and Methods. In the current survey, we compared transplantation outcomes in a cohort of 394 acute myeloid leukemia (AML) patients given grafts from HLA-identical siblings after FB (n=218; with a total busulfan dose ranging between 7.1 and 8.9 mg/kg p.o., or between 6.0 and 6.9 mg/kg i.v.) or FM (n=176; with a total melphalan dose ranging between 130 and 150 mg/m2). Patients given manipulated grafts and those given T cell depleting agents (ATG or alemtuzumab) were not included. At time of transplantation, 266 patients (68%) were in first complete remission (CR1), 69 (18%) in later CR, while 59 patients (15%) had advanced diseases. Three-hundreds and fifty-two patients (89%) received peripheral blood stem cells while the remaining 42 patients received bone marrows as stem cell source. Results. Three FB patients but no FM patients failed to engraft. Median time for reaching 500 neutrophils was 17 (1-50) days in FB patients versus 14 (9-43) days in FM patients (P

Description: Background: Allogeneic hematopoietic cell transplantation (alloHCT) is a complex therapy which can induce a multi-factorial cascade of complications, and potentially lead to patient death. The triggering event(s), sequence and severity of such complications can significantly differ between patients, but in many cases, a so-called "multi-organ failure" (MOF) is usually reported as the leading cause of death. However, a patient's clinical course can be very heterogeneous across and within cause-specific mortalities. Moreover, comorbidities present prior to alloHCT carry their own risks and represent additional confounding factors. Therefore, identification of the exact initial trigger or event leading to MOF in alloHCT patients is a critical step towards early intervention and improvement of patients' outcome. The goal of the current study was to establish and identify the exact cause of death in alloHCT patients where MOF was considered to be the main cause of death. Of note, we specifically focused on VOD/SOS because this life-threatening complication has a mortality exceeding 80% in severe cases, ending usually in MOF, and because VOD/SOS has subtle and dynamic evolution features which are not easy to capture, but could be potentially controlled by appropriate therapy (eg. defibrotide). Patients and Methods: For the purpose of this analysis, we randomly identified 241 adult patients (42% female; median age: 50 years; range 19-73) with acute leukemia (72% AML, 25% ALL, 3% other) allografted between 2010 and 2018 from a matched sibling (29%), unrelated (61%) or haploidentical donor (10%). All patients were reported to the EBMT registry to have died from MOF. Karnofsky score at time of alloHCT was 〉90 in 87% of patients. Seventy-three percent of patients underwent transplant in complete remission, and conditioning was myeloablative in 70%. Sixty patients (25%) received VOD/SOS prophylaxis treatment, mainly consisting of ursodiol and/or heparin. Patients' files were reviewed in detail in order to capture all early signs and symptoms which occurred prior to MOF, based on the classical Baltimore criteria, modified Seattle criteria, and/or the newly published EBMT criteria. These criteria included bilirubin levels, the presence of hepatomegaly or painful hepatomegaly, ascites, percentage weight gain, hemodynamic instability, and ultrasound/histologically proven VOD/SOS. Results: Using one or more of the above criteria defining VOD/SOS, we identified a total of 67 (28%) patients for whom VOD/SOS could be considered as the trigger for MOF and the leading cause of death. Interestingly, among these 67 patients, only 22 (33%) were originally reported by the centers as having developed VOD/SOS leading to MOF post-transplant. When comparing the group of 67 patients dying of VOD/SOS-related MOF and the remaining 174 patients dying of MOF not related to VOD/SOS (please see attached table), a multivariate regression analysis identified a significant increase in VOD/SOS incidence (odds ratio 3.9; 95%CI, 2.42-6.33; p

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: Background. We previously reported data from 103 pts receiving PBSC from HLA-matched unrelated donors after conditioning with 2 Gy total body irradiation (TBI) plus 90 mg/m2 fludarabine. Postgrafting immunosuppression included MMF (administered from day 0 until day +40 with taper through day +96), and CSP (given from day -3 to day +100, with taper through day 180) (historical group). The incidences of grade III–IV acute and chronic extensive GVHD were 14% and 48%, respectively. Several studies have suggested that CSP could prevent activation-induced cell death of T-cells, and thus potentially delay the eradication of donor-versus-host alloreactive T-cells, preventing tolerance induction. Conversely, antimetabolite inhibitors, such as MMF, could delete alloreactive T-cells by induction of activation-induced cell death and apoptosis, and thus might favor tolerance induction Pts and Methods. Here, we investigated whether postgrafting immunosuppression with prolonged MMF (given at 15 mg/kg orally thrice a day from day 0 to day +30, then at 15 mg/kg orally twice a day until day 150, and then tapered at day 150 and discontinued at day 180) and truncated CSP (abruptly discontinued on day 80), would promote tolerance induction and reduce the incidence of GVHD (current protocol, n=71) after unrelated HCT with nonmyeloablative conditioning. Results. Sustained donor engraftment was achieved in 68 pts (96%) in the current protocol, versus in 98 pts (95%) in the historical group. Grades II, III and IV acute GVHD were seen in 36 (50.7%), 11 (15.5%) and 7 (9.9%) pts, respectively, in the current protocol, versus 39 (37.9%), 11 (10.7%) and 4 (3.9%) pts, respectively, in the historical group. The incidences of grade II–IV and grade III–IV acute GVHD were 75% and 23% in the current protocol, versus 52% (P=.05) and 14% (P=.14) in the historical group. The 1-yr incidence of chronic GVHD was 46% in the current protocol, versus 48% in the historical group. Finally, the 1-yr probabilities of relapse, nonrelapse mortality and progression-free survival were 24%, 36% and 40%, respectively, in the current protocol, versus 26% (P=.9), 18% (P=.005) and 56% (P=.04) in the historical group. The increased incidence of nonrelapse mortality in the current protocol was not due only to the increased incidence of grade II–IV acute GVHD, since nonrelapse mortality remained significantly higher in the current protocol than in the historical group after adjusting for occurrence of grade II and grade III–IV acute GVHD (P=.04). Evaluation of pretransplant comorbidities is ongoing. Conclusions. Postgrafting immunosuppression with prolonged MMF and truncated CSP failed to decrease the incidence of GVHD after nonmyeloablative conditioning with URD. Ongoing efforts are directed at reducing the risk of acute GVHD after nonmyeloablative conditioning for unrelated donors by replacing tacrolimus for CSP, and by adding sirolimus to MMF and CSP.

Description: Poor graft function (PGF) is a rare but serious complication of hematopoietic cell transplantation (HCT). Boost of CD34+stem cells could be a therapeutic option but is not always feasible nor successfull. Due to their hematopoietic support properties and immune regulatory effects, multipotent mesenchymal stromal cells (MSC) could be considered a good candidate to help restore bone marrow (BM) niches homeostasis and facilitate hematopoiesis after HCT. Here, we conducted a multicenter prospective study to evaluate the efficacy of MSC perfusion to improve PGF after HCT. Thirty patients (median age 51y, range 11 to 70y) received a single IV administration of third-party donor BM-derived MSC (1-2 x 10 exp6/kg body weight) for PGF after allogeneic HCT (median 151 days after HCT, range 62- 430). PGF was defined as (1) at least one cytopenia (Hb 〈 80 g/L, absolute neutrophil count [ANC] 0.5 x 10exp9/L and transfusion independence). Corresponding response rates increased at day 180, with 69.2% (95% CI, 51.5 - 86.9%) OR and 61.5% (95% CI, 42.8 - 80.2%) CR, respectively. Overall survival at 1 year post-MSC was 70% (95% CI, 53.6 - 86.3%), with all but one of the d90-responders being alive (1 death to relapse of hematological malignancy). No severe adverse event related to MSC administration was reported during the 1-year follow-up. In conclusion , perfusion of BM-derived MSC from a third party donor appeared to be safe and to improve PGF after allogeneic HCT, although spontaneous amelioration cannot be excluded. Further (ideally comparative) studies are warranted to confirme our results. Disclosures Selleslag: Celyad: Other: Clinical trial research (no honoraria recieved); Novartis: Consultancy, Honoraria, Speakers Bureau.

Description: Background. We investigated the impact of graft composition on GVHD in 1520 allogeneic transplant recipients given PBSC. Methods. Graft composition was determined by FACS analyses and expressed as cells/kg of recipient weight. Cell types were defined as follows: CD3+ T cell (CD3+CD56−), CD4+ T cell (CD3+CD4+), CD8+ T cell (CD3+CD8+), natural killer (NK) cell (CD3−CD56+), NK/T cell (CD3+CD56+), B cell (CD20+), monocyte (CD14+), and CD34+ cells (CD34+CD14−). Cell doses were modeled as continuous linear variables or by quartiles. Cox (for chronic) and logistic (for acute) regression models were adjusted for conditioning regimen intensity, pt age, female donor to male recipient vs. other gender combination, and disease risk. Results. The median age among all pts was 52 (45 among myeloablative, 57 among nonmyeloablative); 27% of pts were male transplanted from a female donor; 66% of pts had high-risk disease; 49% were transplanted from an alternative donor (47% unrelated, 2% mismatched sibling); 51% received a nonmyeloablative conditioning. Table 1 shows graft composition and GVHD incidences according to donor type and conditioning intensity. There were no statistically significant associations between cell dose and acute GVHD, although there was a suggestion that the association of NK/T cell was dependent on conditioning. In particular, increasing NK/T cells were associated with an increased probability of grades 2–4 GVHD among myeloablative pts (p=.03) but not among nonmyeloablative pts. There were no statistically significant associations between cell dose and chronic GVHD, but several cell types showed evidence of an interaction between donor and cell dose. In particular, increasing B cells and CD3+, CD4+, and CD8+ T cells were associated with an increased risk of chronic GVHD among pts with an HLA-identical donor (p=.006, p=.0003, p=.0002, p=.02, respectively) while the associations among alternative donors were in the opposite direction and were not statistically significant (p=.68, p=.26, p=.44, and p=.15, respectively). Lack of correlation in the alternative donor group may reflect differences in T-cell responses to the greater minor histocompatability divergence, or changes in T-cell functionality associated with the dose of G-CSF (higher in sibling donors) used for mobilization. Increasing CD34+ cells was not statistically associated with chronic GVHD in either myeloablative (negative association, p=.10) or nonmyeloablative (positive association, p=.11) pts. Conclusions. Higher numbers of transplanted T and B cells were associated with an increased risk of extensive chronic GVHD in pts given grafts from HLA-identical siblings. Increasing CD34+ cells were not associated with a statistically significant increase in chronic GVHD. Further analyses looking at other clinical outcomes such as survival are forthcoming. Table 1. Graft composition and GVHD according to conditioning intensity and donor type Nonmyeloablative conditioning and HLA-id sibling Myeloablative conditioning and HLA-id sibling Nonmyeloablative conditioning and alternative donor Myeloablative conditioning and alternative donor Median # of T cell transplanted (× 108/Kg) 3.4 2.5 2.5 2.8 Median # of CD34+ cell transplanted (× 106/Kg) 8.6 7.5 6.9 7.5 Incidence of grade II–IV acute GVHD (%) 47 66 61 81 Incidence of extensive chronic GVHD (%) 40 54 41 49

Description: Charlson Comorbidity Index (CCI) is a weighted index that assigns scores to comorbidities and has been applied to predict mortality for patients (pts) with various medical conditions including solid tumors. Recently, CCI was used to predict the risk of non-relapse mortality (NRM) and overall survival (OS) for pts with hematologic malignancies given ablative or nonablative HCT1. The aim of this study is to a) better define previously identified comorbidities, b) investigate additional HCT-related comorbidities, and c) establish comorbidity scores suitable for HCT. We retrospectively reviewed comorbidities and laboratory data of 1058 pts given HCT at our center between 1997–2003 after ablative (n=763) or nonablative (n=295) conditioning (Table 1). The following refinements of the original CCI were made: pulmonary functions tests (PFTs) were used for pulmonary, liver function tests for hepatic, and ejection fraction ≤50% added to cardiac comorbidities. Also, new comorbidities were evaluated including bleeding, recent infections under treatment, depression, and obesity. Pts were randomly divided into a training (n=710) and a testing set (n=349). In the training set, the unadjusted HRs for 2-year NRM were calculated for each comorbidity. These values were then adjusted for other comorbidities, disease risk, conditioning type, and pt age. The adjusted HRs were employed as weights for individual comorbidities (Table 2). The refined pulmonary and hepatic comorbidity definitions were useful in identifying pts with increased risk of NRM. In addition, new comorbidities such as infection, bleeding, depression, and obesity appeared to be predictive. Some comorbidities had higher scores than the original CCI due to higher predictive power in the training set. The modified score was then validated in the testing set. The 2-year rates of NRM for scores of 0, 1–2, and ≥3 were 13%, 24%, and 40%, respectively, and of OS were 71%, 56%, and 37%, respectively. Applying the scores to nonablative and ablative pts, respectively, NRM of 9 vs 12% (p=0.04) were seen for score 0, 17 vs 27% (p=0.002) for scores 1–2, and 39 vs 42% (p=0.18) for scores ≥3. This modified HCT-specific comorbidity index provided a simple, readily applicable and valid method of estimating the risk of NRM and OS among pts given ablative or nonablative HCT. We suggest using this index in assessing risks for NRM and OS before allogeneic HCT. Sorror M et al. Comparing morbidity and mortality of HLA-matched unrelated donor hematopoietic cell transplantation after nonmyeloablative and myeloablative conditioning: influence of pretransplant comorbidities. Blood. Online April 27, 2004; DOI 10.1182/blood-2004-02-0545 Table 1: pts Nonablative(n=295), % Ablative(n=763), % P TBI-total body irradiation, FLU-fludarabine, BU-busulfan, and CY-cyclophosphamide Conditioning TBI 23 0 TBI/FLU 77 0 BU/CY 0 60 CY/TBI 0 40 Age, median (range) years 54 (1–73) 42 (1–66)

Description: We retrospectively analyzed the efficacy of unrelated HCT after nonmyeloablative conditioning with fludarabine 3 x 30 mg/m2 and 2 Gy total body irradiation as treatment for CML. Postgrafting immunosuppression consisted of mycophenolate mofetil and cyclosporine. Data from 21 CML pts in CP1 (n=12), AP (n=5), CP2 (n=3), or BC (n=1) were analyzed. Median pt age was 54 (range, 33–66) years. Median time from diagnosis to HCT was 26 (range, 5–121) months. Two pts (in BC and in AP) had previously failed myeloablative allogeneic HCT. Four pts received marrow as stem cell source, and 17 G-PBMC. Two pts each had single HLA class 1-allele mismatches with their donor, the remaining were matched for 10/10 HLA antigens. One pt died before donor engraftment was evaluable. Donor engraftment at day 28 (defined as ≥ 5% donor T-cell chimerism) was observed in 18 of the 20 evaluable pts (90%), and was sustained in 11 of those (55%). All graft rejections were nonfatal and followed by autologous hematopoietic reconstitution. Day 28 T-cell chimerism was assessed in 19 pts. Eight of 9 pts with day 28 donor T-cell chimerism levels ≤ 40 rejected their grafts, compared with 1 of 10 pts with day 28 T-cell chimerism levels 〉 40%. The two marrow recipients who had not received chemotherapy prior to the HCT rejected their grafts. Eight of 11 (70%) patients with sustained engraftment (including 4 of 5 pts in CP1, 3 of 4 pts in AP, 1 of 2 pts CP2) achieved sustained complete cytogenetic remissions (CCR) (Table 1). In addition, another patient who had progressed to blast crisis after HCT achieved CCR and BCR/ABL molecular negativity by Q-PCR, following a short course of farnestyl transferase inhibitor and development of chronic GVHD. Analysis of minimal residual disease by Q-PCR in 6 pts with sustained CCR showed disappearance of BCR/ABL transcripts 84 to 524 days after HCT, consistent with graft-versus-tumor effects. Grade II-IV acute GVHD was seen in 11 pts and extensive chronic GVHD in 8. One pt transplanted in CP2 died of progressive disease (PD), and two pts transplanted in CP2 (n=1) and in AP (n=1) died of nonrelapse mortality (NRM), while in CCR. Three of 9 pts with graft rejection died of progressive disease, 5 were alive in CP, and 1 in AP at the time of the analysis. One hundred-day and two-year nonrelapse mortalities for all pts were 0% and 12%, respectively. Two year probability of overall survivals were 100% for pts transplanted in CP1, and 0% for those transplanted in more advanced stage. In summary, unrelated donor HCT following nonmyeloablative conditioning with fludarabine and low-dose TBI was feasible with low nonrelapse mortality. Sustained CCR were achieved in patients with stable engraftment, but graft rejection occurred in 45% of pts. Further efforts are being directed at reducing the risk of graft rejection by exclusive use of G-PBMC and increasing the intensity of pretransplant immunosuppression. Outcome Status at HCT Graft rejection Outcome (days after HCT) CP1 (n=12) No (n=5) Alive in CCR (1205+, 1126+, 1123+, 867*+, 89+) Yes (n=7) Alive in CP (1188+, 960+, 946+, 698+, 533+) Alive in AP (356+) Death of PD (806) AP (n=5) No (n=4) Alive in CCR (267+, 115+) Death of NRM (478) Death of PD (241) Yes (n=1) Death of PD (271) CP2 (n=3) No (n=2) Death of NRM (164) Death of PD (233) Yes (n=1) Death of PD (540) BC (n=1) NE Died of PD (21)