Ex vivo CD34+ selected T-cell depletion (TCD) has been developed as a strategy to reduce the incidence of graft-versus-host disease (GVHD) after allogeneic hematopoietic stem cell transplantation (allo-HSCT). Clinical characteristics, treatment responses, and outcomes of patients developing acute (a-) and chronic (c-) GVHD after TCD allo-HSCT have not been well established. We evaluated 241 consecutive patients (median age 57 years) with acute leukemia (n = 191, 79%) or myelodysplastic syndrome (MDS) (n = 50, 21%) undergoing CD34+ selected TCD allo-HSCT without post-HCST immunosuppression in a single institution. Cumulative incidences of grade II–IV and III–IV aGVHD at 180 days were 16% (95% CI:12–21) and 5% (95% CI:3–9), respectively. The skin was the most frequent organ involved, followed by the GI tract. Patients were treated with topical corticosteroids, poorly absorbed corticosteroids (Budesonide), and/or systemic corticosteroids. The overall day 28 treatment response was high at 82%. Cumulative incidence of any cGVHD at 3 years was 5% (95% CI:3–9), with a median time of onset of 256 days (range 95–1645). The 3-year transplant-related mortality, relapse, overall survival and disease-free survival were 24% (95%CI: 18–30), 22% (95% CI:17–27), 57% (95% CI:50–64) and 54% (95% CI:47–61), respectively. The 1-year and 3-years probability of cGVHD-free/relapse-free survival (CRFS) were 65% (95% CI:59–71) and 52% (95% CI:45–59), respectively. Our findings support the use of ex vivo CD34+ selected TCD allograft as a calcineurin inhibitor-free intervention for the prevention of GVHD in patients with acute leukemia and MDS.

Graft‐vs‐host Disease (GVHD) is a frequent complication following allogeneic stem cell transplantation (allo‐HCT). The National Institute of Health consensus group established new guidelines for the evaluation of chronic GVHD. However, GVHD assessment remains challenging due to its complexity and requirement for laborious evaluation. We, therefore, established a standardized approach for the assessment of chronic GVHD in accordance with the NIH consensus criteria (NCC) guidelines. At a single institution, all allograft recipients were evaluated for GVHD within the first‐year post allo‐HCT following a three‐step workflow (real‐time assessment, consensus review, and documentation). A GVHD adjudication committee was created and a dynamic electronic GVHD data capture form was developed guiding the clinician through a comprehensive review of systems following the NCC guidelines. We found that the assessment and reporting of GVHD reached 100% compliance. The establishment of an institutional GVHD adjudication committee enabled standardized assessment of GVHD. Our workflow can be adopted by other centers to create a similar framework for dedicated GVHD evaluation.

Introduction: With recent improvements in therapy, median survival after diagnosis for patients with Multiple Myeloma (MM) now exceeds 5 years. However, for a variety of reasons most patients do not receive the most intensive therapy for MM, namely autologous stem cell transplantation (auto-SCT). In order to better understand which patients will receive the greatest benefit from auto-SCT, we developed a machine learning (ML) model to identify relevant predictors of 5-year mortality for MM patients who received auto-SCT. Previous studies have identified adverse cytogenetics (e.g., chromosome 13 deletion), elevated B2-microglobulin, elevated lactate dehydrogenase, and the receipt of more than 1 year of standard chemotherapy as prognostic risk factors. Using several ML techniques, we built a predictive model for 5-year mortality post-transplant and identified features that were most predictive. Methods: We used the de-identified nationwide Flatiron Health electronic health record (EHR)-derived database. Patients with a confirmed MM diagnosis between January 1, 2011 and May 31, 2019, and who received an auto-SCT at any date after their diagnosis, were eligible for analysis. Clinically relevant data available for these patients included demographic information, lab results, date of transplant, medications administered before transplant, lines of therapy, M-spike results, time from diagnosis to auto-SCT, and mortality data. Patients were classified based on 5-year survival after auto-SCT. Patients who underwent auto-SCT within the last 5 years and are still alive were excluded from this analysis, as were those who received an allogeneic transplant after auto-SCT. Using this feature set, we trained a variety of industry-standard ML models, including logistic regression, support vector machines, gradient boosted trees, and random forest. We used 5-fold cross validation to evaluate model performance, evaluated based on the area under the receiver operating curve (AUC). We then determined the relative importance of each feature in the logistic regression model and reviewed each feature's clinical relevance. Results: 1016 patients (588 patients who died within 5 years from their transplant date, and 428 patients who died 5 years or later following transplant) were included in the cohort. The logistic regression model was the best performing model, achieving an AUC of 0.77, accuracy of 0.70, and F1 score of 0.60 (Table). Eight of the 10 most predictive features for early mortality were presence of chromosome 1 abnormalities, higher age at diagnosis, higher serum albumin levels, higher number of visits before transplant, presence of ICD codes for comorbid conditions, and presence of administrations of pomalidomide, bortezomib, or zoledronic acid. These features were determined to be clinically meaningful and all were associated with mortality before 5 years. Receipt of fosaprepitant was also predictive of mortality before 5 years, though the clinical relationship is more challenging to explain. Presence of M-spike elevation in the 100 days before or after auto-SCT was predictive of reduced mortality. The latter finding, along with higher mortality in patients with higher serum albumin levels, was counterintuitive. Discussion: Auto-SCT procedures can have high toxicity and cost; therefore, accurate prediction of outcomes could improve understanding of the utility of auto-SCT for individual MM patients. Our study demonstrates the potential of using ML models for risk prediction in MM, though the presence of counterintuitive findings (e.g. higher albumin correlated to poorer survival) will require additional investigation. We hope that this study inspires future research into using ML techniques to support personalized clinical decision-making, help organize supportive care initiatives, and inform pre-approval decisions made by payers. Disclosures Chen: Flatiron Health, Inc: Employment; Roche: Equity Ownership. Garapati:Flatiron Health, Inc.: Employment. Wu:Flatiron Health: Employment. Ko:Flatiron Health, Inc.: Employment. Falk:Flatiron Health, Inc.: Employment; Roche: Equity Ownership. Dierov:Flatiron Health, Inc.: Employment; Roche: Equity Ownership. Stasiw:Roche: Equity Ownership; Flatiron Health, Inc.: Employment. Opong:Flatiron Health, Inc., which is an independent subsidiary of the Roche Group: Employment, Research Funding. Carson:Flatiron Health, Inc., which is an independent subsidiary of the Roche Group: Employment, Research Funding; Roche: Equity Ownership.

overall survival (OS) compared to unrelated donors (UD) in patients with high-risk acute leukemia (AL) and residual disease prior to HCT. Patients and Methods: We present a retrospective casecontrolled study of first alloHCT for high-risk AL by the Spanish HCT Group. Ninety-four recipients of HCT with single CB units plus CD34+ selected cells from third-party donors (Haplo-Cord, HC), where compared (1:2) with 188 recipients of UD HCT matched for age, gender, WHO diagnosis, disease status at HCT, time from diagnosis to HCT, prior auto-HCT, TBI use in conditioning and year of HCT. HC cases included 57 men (61%), median age 34 years (16-64), 51 AML and 43 ALL, 49 in first CR, 16 in CR2 and 29 more advanced, including 23 with detectable disease. Six had a prior autologous HCT. Median time from diagnosis to alloHCT was 8.3 months (2-66). Results: With a median follow-up of 85 months for survivors, 113 patients were alive, 44 in the HC group (47%) and 79 controls (42%; n.s.). NRM at 6 years (y) was comparable in both groups: HC 33.0% (95CI: 24.7-44.0) vs controls 34.4% (95CI: 28.1-42.0; n.s.). Overall, HC had a statistical trend towards a better OS than controls at 6 y (47.7%, 95CI: 37.6-57.8, vs 37.0%, 95CI: 29.9-44.0; P = .079), which may associate with their cumulative incidence of relapse (24.5%, 95CI: 17.2-34.9, vs 30.7%, 95CI: 24.7-38.2; P = .135). Next, we performed the analysis in patient subsets stratified according to their risk of relapse. In patients with low risk of relapse transplanted in CR1 or CR2 there were no differences in outcomes between HC and matched UD controls. However, patients that underwent alloHCT in more advanced disease (AD, CR3 or later, partial remission or refractory disease), had significantly better outcomes following HC HCT. OS rate at 6 y in patients with AD in the HC group was 31,0% (CI 14.2-47.9) vs 13.0% (CI 3.3-22.8) in UD group (P = .046). PFS rate at 6 y in patients with AD disease in the HC group was 24.1% (CI 8.6-39.7) vs 13.0% in UD group (P = .046). Relapse rate at 6 y in patients with AD in the HC group was 37.9% ) vs 47.8% ) in the UD group (P = .069). Conclusions: With the increase availability and use of alternative donors for alloHCT, data to inform donor choice are needed. It is unlikely that any single type of alternative donor will be the best choice for all patients lacking a matched related donor. Our data with HC HCT, in line with recent findings by Milano et al, contribute to the evidence to suggest that unrelated CB might be a preferable donor choice for patients with AL and a high-risk of relapse.

Background: GVHD contributes significantly to transplantrelated morbidity and mortality after allo-HSCT. While clinical characteristics, therapeutic approaches and outcomes of patients developing acute (a-) and chronic (c-) GVHD after unmodified grafts for allo-HSCT have been widely analyzed, there is paucity of studies having specifically addressed these issues in the CD34+ selected ex-vivo TCD setting. Methods: We evaluated all consecutive patients with acute leukemia in complete remission or myelodysplastic syndrome (MDS) with 5% blasts undergoing CD34+ selected TCD allo-HSCT from peripheral blood mobilized progenitors in a single center from 01/2008 to 06/2014. IBMTR and NIH consensus criteria were used for diagnosis and grading of aGVHD and cGVHD, respectively. Several outcomes were evaluated including the recently definedcGVHD-free/ relapsefree survival (CRFS). Results: A total of 255 patients with a median age of 56 years (range 20-73) were included (67% AML, 13% ALL, 20% MDS). Seventy-seven patients presented with grade I-IV aGVHD at a median onset of 70 days (range 15-452). The majority developed aGVHD prior to day 100 (49/77, 64%) whereas 28/ 77 (36%) had aGVHD after day 100. With a median follow-up of 22 months (range 12-85), the cumulative incidences (cum.inc.) of grade II-IV and III-IV aGVHD at day 180 were 18% (95% CI: 14-23) and 6% (95% CI: 3-9), respectively. Among the 63 patients with aGVHD on day 180, the skin was the most commonly affected organ (n¼ 41, 65%), followed by gastrointestinal tract (n¼ 33, 52%), and liver (n¼ 5, 8%). Treatment response (complete or partial) by day 28 in these patients was 81%. Univariate analysis was performed to determine risk factors for grade III-IV aGVHD (Table). It demonstrated that recipient CMV negativity was the only variable associated with higher grade III-IV aGVHD (HR 5, p¼ 0.01), whereas the use of TBI in the conditioning regimen (HR 2.8, p¼ 0.06) and mismatched donors (HR 2.5, p¼ 0.09) showed a trend. cGVHD developed in 13 patients at a median of 160 days (95-1645): 9 had classical cGVHD (6 interrupted/ progressive, 3 de novo) and 4 had overlap syndrome. The cum.inc. of any grade of cGVHD at 3 years was 5% (95% CI: 3-8). Of the 229 patients who were alive and in remission at day 100, 201(88%) had no GVHD after day 100. The 1-year probability of CRFS was 65% (95% CI: 59-70) (Figure). Conclusions: CD34+ selected TCD allo-HSCT is a calcineurin inhibitor (CNI)-free intervention associated with low rate of aGVHD and cGVHD. Recipient CMV positivity was associated with lower day 180 grade III-IV aGVHD. This unexpected finding could be explained by the low incidence of aGVHD and needs further investigation in a larger population. The 1year CRFS demonstrated to be high in this population. The development of a CNI-free allo-HSCT intervention is a promising and feasible approach in patients with acute leukemia and MDS in remission.