Citations

Abstract Background Mitochondrial dysfunction in Alzheimer's Disease (AD) is characterized by impaired energy production, oxidative stress, and disrupted calcium homeostasis, which collectively contribute to neuronal damage, synaptic dysfunction, and the progression of AD pathology. A significant knowledge gap in studying mitochondrial dysfunction in Alzheimer's Disease (AD) lies in understanding the precise molecular mechanisms linking mitochondrial metabolic alterations to neuronal damage, particularly how sex‐specific and cell‐type‐specific mitochondrial processes contribute to the progression of AD pathology. Method This study integrates multiple datasets, post‐mortem brain RNAseq from ROSMAP, Mayo Clinic brain bank and Mount Sinai Brain Bank cohort data available via AD Knowledge Portal hosted by Sage Bionetworks), Human Protein Atlas (HPA), MitoCarta3, genome‐scale human metabolic reconstruction, mitochondriome analysis, and brain & blood metabolomics data (data generated by the Alzheimer's Disease Metabolomics Consortium (ADMC)) to identify mitochondrial genes associated with disrupted metabolism in AD. Additionally, biodomain analysis was used to identify key processes associated with pathology. Result A total of 62 common genes were identified across these approaches, highlighting their potential significance in mitochondrial processes. We also developed the framework to perform metabolite‐gene correlation analyses revealing key mitochondrial‐metabolic interactions (eg. homocysteine and MTHFR gene; acetyl‐CoA and PDHA1 gene). Integrated in silico metabolic analysis identified reaction fluxes, uncovering sex‐specific differences in mitochondrial transport and tricarboxylic acid (TCA) cycle subsystems, with notable alterations in females. Biodomain enrichment analysis focused on mitochondrial processes, identifying disruptions in pathways critical for energy production and redox homeostasis. Additionally, we analyzed datasets related to metformin and complex I inhibitors, linking mitochondrial complex I modulation with metabolic alterations in AD. This analysis identified druggable targets and potential therapeutic candidates for mitochondrial metabolic pathways. Conclusion Our findings integrate multiple data sources to provide a systems‐level understanding of mitochondrial metabolism in AD. Sex‐specific differences associated with mitochondrial dysfunction in AD were identified. We identify important mitochondrial transport proteins and metabolites that could potentially be used as druggable targets in AD.

Noninvasive electrophysiological source imaging (ESI) is a valuable tool for localizing and imaging brain activity, with significant potential to aid presurgical planning in focal drug-resistant epilepsy (fDRE) patients. Scalp electroencephalography (EEG) biomarkers, including interictal spikes, high-frequency oscillations (HFOs), and seizures, each offer unique capabilities in estimating the epileptogenic zone (EZ). However, there is a limited quantitative understanding of how these biomarkers differ in source-imaging precision, requiring distinct processing pipelines. Here, we developed a spatial–temporal–spectral imaging (STSI) framework for precision source imaging, and quantitatively evaluated various epilepsy biomarkers for source imaging in 2,081 individual events (spikes, HFOs, and seizures) from a cohort of 42 fDRE patients, comparing results to clinical ground truth such as surgical resection outcomes and intracranial EEG-defined seizure onset zones. The STSI enabled quantitative comparisons across key EEG epilepsy-related biomarkers, with averaged localization errors of 6.67 mm for seizures, 8.73 mm for HFOs overlapping with spikes (pHFO), 10.28 mm for HFO-riding spikes (pSpike), 19.59 mm for general spikes (aSpike), and 36.53 mm for general HFOs (aHFO), respectively, for seizure-free patients. These findings indicate that HFOs overlapping with spikes is the most spatially accurate interictal biomarker for mapping the EZ. The proposed STSI framework not only establishes a unified analysis approach for epileptic biomarkers to enhance presurgical planning in focal drug-resistant epilepsy, but could also generalize as a versatile tool for mapping event-related potentials, neural oscillations, and dynamic brain states, within a single framework to advance cognitive neuroscience research and clinical management of neurological and psychiatric disorders.

Abstract Background: Post-transplant lymphoproliferative disorder (PTLD) occurs in up to 30% of high-risk solid organ transplant recipients and is strongly associated with the intensity and duration of immunosuppression. While Epstein-Barr virus (EBV)-positive monomorphic PTLD and diffuse large B-cell lymphoma (DLBCL) share overlapping morphologic features and expression of canonical B-cell markers, they exhibit distinct EBV transcriptional programs. Immunosuppressive regimens, essential for preventing graft rejection, impair EBV-specific cytotoxic T lymphocyte (CTL) responses, thereby promoting EBV-driven lymphomagenesis. We hypothesized that these differences in viral transcription and immune evasion reflect fundamentally distinct immune microenvironments and host responses in EBV-positive PTLD compared to DLBCL. Methods This prospective, non-interventional study has been conducted at Moffitt Cancer Center since 2019. To characterize tumor-specific T-cell responses in EBV-positive PTLD versus DLBCL, peripheral blood mononuclear cells (PBMCs) from enrolled patients were infected with EBV (ATCC VR-1492) to generate autologous, immortalized lymphoblastoid B-cell lines (LCLs). EBV antigen-specific immune responses were quantified using an IFN-γ ELISpot assay targeting EBNA1, LAMP1, LAMP2A, BARF1, and a CEF peptide pool as a control. To optimize cell viability and functional readout post-thaw, we implemented an overnight resting protocol for PBMCs, which demonstrated superior IFN-γ release compared to the conventional 1-hour rest. Results We present pilot data from 11 consecutive patients (55% male; median age 63.4 years), including 3 with PTLD (2 EBV-positive; 2 monomorphic) and 8 with EBV-negative DLBCL. A total of 12 evaluable PBMC samples were analyzed: 7 pre-treatment, 1 mid-treatment, and 4 post-treatment. Successful generation of autologous LCLs was predominantly achieved using pre-treatment samples, whereas samples collected after B-cell–targeted therapy (Rituximab ± CHOP) showed reduced transformation efficiency. Notably, early pre-treatment samples exhibited a relative enrichment of CD27⁺ memory B cells. Baseline flow cytometric profiling of B cells undergoing LCL transformation revealed low CD19 expression (range <1%–15%). Repeat immunophenotyping at 3 weeks identified residual T cells within cultures, suggesting ongoing B-cell clearance. In matched flow cytometry and ELISpot assays from 7 evaluable patients with pre-treatment samples, EBV-positive PTLD cases demonstrated pronounced overexpression of PD-1 and TIM-3, but not LAG-3, alongside markedly diminished T-cell responses to EBV antigens. This distinct immunophenotypic and functional signature suggests T-cell exhaustion or dysfunction, likely driven by prior immunosuppression, which may impair immune surveillance and facilitate EBV-driven lymphomagenesis. Conclusions Our findings reveal distinct immunologic profiles in EBV-positive monomorphic PTLD compared to EBV-negative DLBCL, characterized by immune exhaustion and impaired T-cell functionality. Specifically, overexpression of PD-1 and TIM-3—but not LAG-3—in EBV-positive PTLD correlated with diminished antigen-specific T-cell responses, as measured by ELISpot, underscoring the role of prior immunosuppression in shaping ineffective tumor-specific immunity. Successful generation of autologous LCLs was predominantly achieved from pre-treatment PBMCs, highlighting the impact of B-cell–depleting therapies (Rituximab ± CHOP) on downstream immune profiling feasibility. The persistence of residual T cells in long-term LCL cultures suggests ongoing B-cell clearance and may reflect previously unrecognized immune dynamics in PTLD. Collectively, these data provide novel insights into the immune landscape of EBV-driven PTLD and support the rationale for therapeutic strategies aimed at reversing T-cell exhaustion—such as immune checkpoint blockade or adoptive T-cell therapies—to restore anti-EBV immunity in transplant recipients.

Abstract Introduction Persistent symptom burden in classic myeloproliferative neoplasms (MPNs) i.e., polycythemia vera (PV), essential thrombocythemia (ET) and myelofibrosis (MF) is linked to shorter survival, more thromboembolic events and impaired quality of life. Identifying demographic, clinical, laboratory, and treatment-related predictors of symptom improvement can inform risk-adapted care and guide supportive interventions. Methods Patients with PV, ET, and MF were identified through a retrospective chart review. Total symptom scores (TSS) and scores of individual symptoms were recorded at the time of first TSS assessment. Demographic (age, sex, race, ethnicity), seasonal (summer: April–September; winter: October–March), clinical (time since diagnosis, depression, obesity, spleen size), treatment-related (phlebotomy, hydroxyurea, interferon-alpha, anagrelide, JAK inhibitors, antidepressants) and laboratory (complete blood counts, serum chemistry) factors were captured at the baseline. Symptom improvement was defined as a ≥50% TSS reduction from baseline. Cumulative incidence analyses, accounting for death as a competing risk, were performed to estimate the probability and timing of symptom improvement. Multicollinearity was assessed using variance inflation factors (VIFs). Highly correlated (VIF ≥5) features and features with ≥15% missing observations were excluded. Multivariable Fine-Gray hazards regression was performed to identify predictors of symptom improvement. Sub-distribution hazard ratios (sHR) with 95% confidence intervals (CI) were computed with p-values <0.05, indicating a statistically significant association with symptom improvement. Results A total of 532 patients (PV: 195, ET: 180, MF: 157) were identified between 2019 and 2024. The median age was 65 (IQR: 57-74) for PV, 65 (52-75) for ET, and 69 (63-75) for MF. Most patients were White (PV: 178, ET: 164, MF: 145) and female (PV: 99, ET: 134, MF: 84). Fatigue was the most frequently reported symptom (PV: 161, ET: 155, MF: 136) whereas weight loss was the least reported (PV: 9, ET: 12, MF: 26). Baseline TSS was highest in MF (median: 15; IQR: 7-27) followed by PV (13; 6-25) and ET (13; 5-23). Symptom improvement was observed in 89 (46%) PV, 71 (39%) ET, and 44 (28%) MF patients. Among those who improved, the cumulative incidence of symptom improvement reached 100% by 55 months in PV, 53 months in ET, and 58 months in MF. At 6, 12, and 24 months, cumulative symptom improvement was 22%, 34%, and 57% in PV; 15%, 25%, and 43% in ET; and 19%, 28%, and 39% in MF. Median time to improvement was not reached in any group. Death before symptom improvement was observed in 5 (3%) PV, 6 (3%) ET, and 37 (24%) MF patients. Multivariable regression analyses in PV identified higher baseline TSS (sHR: 1.02; 95%CI: 1.01-1.03) as associated with symptom improvement (p = 0.02). In ET, lower white blood cell count (sHR: 0.89; 95%CI: 0.81-0.98) and less time since diagnosis (0.99; 0.98-0.99) were associated with symptom improvement (p = 0.02). In MF, more months since diagnosis (sHR: 1.01 95%CI: 1.01-1.01) was associated with symptom improvement (p = 0.01). Conclusions Symptom trajectories differed across MPN subtypes. In PV, a higher baseline symptom burden; in ET, a lower white blood cell count and more recent diagnosis and in MF, a longer disease duration were associated with symptom improvement. These readily available parameters may help inform supportive strategies, tailor response expectations, and identify patients for symptom-directed interventions or clinical trials.

Abstract Pseudouridine (Ψ) is the most prevalent RNA modification in human cells. Although Ψ was among the first post-transcriptional modifications discovered (1951), its biological functions remain poorly understood. More recently, Ψ was incorporated into COVID-19 mRNA vaccine, where it enhances RNA stability and reduces innate immune sensing. However, the precise roles of RNA Ψ in cancer, including leukemia, are largely unknown. Here, we identify RNA Pseudouridine Synthase D4 (RPUSD4) as a novel vulnerability in acute myeloid leukemia (AML) via directly catalyzing mitochondrial RNA (mt-RNA) Ψ formation. We further provide proof-of-concept evidence that targeting RPUSD4 suppresses leukemogenesis and stimulates anti-leukemia immunity. To investigate whether RNA Ψ dysregulation is involved in leukemia, we first compared RNA Ψ levels between primitive CD34+ cells (LSC-enriched) and CD34- bulk cells from multiple primary AML specimens. CD34+ AML cells exhibited significantly elevated RNA Ψ levels. We also examined Ψ levels across various RNA types in AML cells, identifying that mt-RNA contained much higher Ψ levels than other RNA types. These findings imply that RNA Ψ, particularly mt-RNA Ψ, may play a role in LSC biology and leukemogenesis. Further unbiased CRISPR screening revealed RPUSD4 as the most critical Ψ synthase for LSC maintenance. Large-scale transcriptomic and patient data analyses showed that RPUSD4 is highly expressed in AML specimens (P = 3.5 × 10-9), particularly within LSCs (P = 1.4 × 10-18). RPUSD4 knockout (KO) significantly suppressed AML proliferation, reduced mitochondrial OxPhos, and decreased LSC frequency in vitro and substantially delayed AML progression in vivo (MLL-r and FLT3ITD models). Of note, RPUSD4 KO-mediated anti-leukemia effects could be fully rescued by wild-type but not catalytically inactive RPUSD4 both in vitro and in vivo, demonstrating that Ψ synthase activity is essential for its function in leukemia. To profile RPUSD4-dependent Ψ sites, we applied two orthogonal technologies, direct nanopore sequencing and PRAISE sequencing (PMID: 36997645), both capable of detecting Ψ at single-base resolution. These approaches converged on a single RPUSD4-dependent Ψ at position 3,067 on mt-16S rRNA. This Ψ site is essential for maintaining rRNA stability. RPUSD4 KO induced mt-16S rRNA decay, thereby suppressing mitochondrial translation. This led to OxPhos inhibition, growth suppression, and LSC eradication in AML. Mechanistically, RPUSD4 is exclusively localized in mitochondria, where it directly interacts with the mt-RNA exonuclease PNPase. Under physiological conditions, PNPase resides within mitochondria and degrades mt-RNA to prevent accumulation of mitochondrial double-stranded RNA (mt-dsRNA). Upon RPUSD4 depletion, PNPase translocates from mitochondria to the cytosol, resulting in two major consequences: (i) mRNA decay:Cytosolic PNPase degrades mRNAs, particularly the nuclear-encoded mitochondrial mRNAs, such as DHODH(Dihydroorotate dehydrogenase). By retrogradely upregulating DHODH, RPUSD4 promotes de novo pyrimidine synthesis in AML. (ii) Aberrant mt-dsRNA accumulation: The buildup and cytoplasmic release of mt-dsRNA activates “viral mimicry” response that sensitizes AML cells to T cell cytotoxicity. Finally, to therapeutically target RPUSD4, we developed an RNA-based inhibitor, CpG-siRNARPUSD4, by conjugating a specific RPUSD4 siRNA to a Toll-like receptor (TLR9) ligand. CpG-siRNARPUSD4 is specifically internalized by TLR9+ cells, including AML cells, thereby maximizing selectivity and minimizing toxicity. Preclinical studies demonstrate that pharmacologically targeting RPUSD4 significantly prolonged survival across multiple AML models, including patient-derived xenograft (PDX) models, and rendered AML cells more susceptible to T cell–mediated cytotoxicity. Furthermore, the RPUSD4 inhibitor exhibited a promising synergistic effect with DHODH inhibitor in treating relapsed AML in vivo. For instance, in a representative relapsed PDX model, median survival times (in days) were: Ctrl – 28, RPUSD4 inhibitor – 36, DHODH inhibitor – 39, and combination – over 60. In conclusion, our findings uncover a previously unrecognized role of RPUSD4 and mt-RNA Ψ in leukemic metabolism, anti-leukemia immunity, and LSC maintenance. This study also bridges a critical knowledge gap regarding the role of RNA Ψ in leukemogenesis.

Abstract Background Acute graft-versus-host disease (GVHD) is a major cause of non-relapse mortality (NRM) after allogeneic hematopoietic cell transplantation (HCT). High-dose steroids are standard first-line therapy for grade II–IV acute GVHD, whereas systemic treatment (tx) is generally not recommended for grade I. However, in practice many patients (pts) with grade I GVHD receive systemic steroids. A prior GITMO randomized trial (n=173) of grade I GVHD to either systemic steroid tx or “watch and wait” found that preemptive steroid tx reduced progression to grade II but not grade III/IV GVHD, increased serious infections, and resulted in similar chronic GVHD, NRM and survival (Bacigalupo, Haematologica 2017). In this study we analyzed real world evidence for early systemic steroid use and the potential impact of serum biomarkers in grade I GVHD. Methods We analyzed clinical data and serum samples from pts with grade I acute GVHD at 24 Mount Sinai Acute GVHD International Consortium (MAGIC) centers between 2014 and 2022 who were treated according to physician discretion. Pts were divided into a preemptive steroid tx group or a watch-and-wait group (systemic steroids administered only if progression to grade II–IV GVHD). Preemptive steroid use was treated as a time-dependent covariate and a frailty proportional hazards model with random effects was used to adjust for center effects to account for possible heterogeneity in clinical practices across transplant centers. Serum samples collected at the diagnosis of grade I GVHD were analyzed for ST2 and REG3α to calculate validated Ann Arbor (AA) scores (1, 2, 3). Results Grade I GVHD was diagnosed in 1143 pts; 591 (52%) received preemptive tx and 552 (48%) were treated by a watch-and-wait strategy. In the preemptive tx group, steroids were initiated at a median of 0 days and within 7 days for >90%. The median initial dose of steroids was 1 mg/kg of prednisone or its equivalent. In the watch and wait group, only 184 pts (33%) required systemic steroids for progression to grade II-IV GVHD at a median of 14 days from grade I diagnosis. Preemptive steroid tx reduced the risk of grade II–IV acute GVHD (HR, 0.71; 95% CI, 0.56–0.89; P=0.003), but not the risk of grade III/IV acute GVHD (HR, 1.03; 95% CI, 0.69–1.52; P=0.899) or chronic GVHD (HR, 1.03; 95% CI, 0.69–1.52; P=0.899), consistent with the GITMO trial. The 12-month incidence of NRM was higher for preemptive tx vs. watch and wait (13% vs 7%, P=0.002). In multivariate analysis preemptive steroid use independently predicted an increased risk of NRM (HR, 1.63; 95% CI, 1.06–2.50; P=0.026) as did several other known risk factors (older recipient age, umbilical cord blood donor, HCT-CI ≥3, and myeloablative conditioning). Greater NRM in the preemptive tx group was driven primarily by 3-fold more infectious deaths; deaths due to uncontrolled GVHD were similar for both tx groups. The risk for relapse was unexpectedly lower in the preemptive tx group (HR 0.64; 95% CI 0.47-0.88; P=0.006) and overall survival was therefore not different between groups, but risk factors for relapse such as minimal residual disease status were not available precluding further analysis of this finding. Serum was collected in 871 (76%) pts at onset of grade I GVHD for retrospective biomarker analysis. AA scores were AA1 (80%), AA2 (15%), and AA3 (5%). AA scores were highly predictive of progression to grade III/IV GVHD (AA1 vs AA2 vs AA3: 9% vs 14% vs 44%, all P<0.001) and the risk of 12-month NRM (8% vs 16% vs 39%, all P<0.001); this pattern was observed in both tx groups. The negative effect of preemptive steroid tx was most pronounced in pts with grade I/AA1 GVHD who experienced 5-fold more infectious related deaths compared to pts managed with a watch and wait approach. Conclusion Treatment of grade I GVHD with systemic steroids did not reduce the incidence of grade III/IV GVHD or deaths due to uncontrolled GVHD and instead increased the risk of NRM from infections compared to pts who were treated only after GVHD progressed. Importantly, 2/3 of the watch and wait pts never needed systemic steroid tx. These findings remained robust after adjusting for multiple confounders, including center effects, and were consistent with the GITMO trial. MAGIC biomarkers at the onset of grade I GVHD predicted outcomes and a watch and wait strategy may be best for the 80% of pts with AA1 GVHD. Better treatment options are needed for the 20% of pts with grade I and AA2/3 GVHD.

Abstract Background Accurate diagnosis and risk stratification in myeloproliferative neoplasms (MPNs) rely on expert interpretation of complex clinical records. Outside major academic centers, specialist expertise is scarce, and time-consuming chart reviews extend clinic visits and delay treatment decisions, underscoring the need for automated, point-of-care phenotyping for MPN patients. Methods Patients with the ICD codes of PV, ET and MF were identified from the electronic health record. Cases were defined as patients meeting the 2017 WHO diagnostic criteria for PV, ET or MF or those with a definitive MPN diagnosis determined by a hematologist. Controls were the patients who met neither the diagnostic criteria nor had a hematologist-assigned diagnosis. Two clinicians extracted demographic, clinical (symptoms, thrombotic history), laboratory (complete blood count, serum chemistry), genetic (mutations, karyotype), bone marrow morphology, diagnostic mentions and prognostic score information from unstructured clinical notes and pathology reports. Prognostic scores included the conventional risk model for PV, the International Prognostic Score of Thrombosis (IPSET) for ET and the Dynamic International Prognostic Scoring System (DIPSS), DIPSS-plus or Mutation-enhanced International Prognostic Scoring System (MIPSS-70) for MF. Discrepancies in data extraction were resolved by a consensus. A random sample of 510 patients (PV: 85, ET: 85, MF: 85, controls: 255) was split into prompt development (n=60) and test (n=450) sets for iterative workflow engineering and evaluation. The workflow included: (i) a rule-based algorithm for automated retrieval of demographic details, the first bone marrow biopsy report, the first clinical note after biopsy and the latest laboratory results within 90 days preceding biopsy (ii) a HIPAA-compliant LLM (GPT-4 turbo) for data extraction from the clinical notes and pathology reports (iii) a regular expressions-based algorithm for post-processing extracted data into structured tables, (iv) a clinician-informed source prioritization algorithm to select data from preferred sources, (v) a modified WHO-informed diagnostic characterization tool incorporating hematologist-assigned diagnoses to distinguish cases from controls and; (vi) an NCCN guidelines-based algorithm for prognostic tool selection and risk-stratification based on available data. Structured prompts were used to extract demographic, clinical, laboratory, genetic and pathological data in a zero-shot deterministic setting. The performance of automated data extraction, diagnostic characterization and prognostication was compared against the manually curated data. Accuracy and F1 scores were computed for data extraction and prognostication tasks while sensitivity and specificity were calculated to assess diagnostic performance. Results A total of 450 (PV: 150, ET: 150, MF: 150) pathology reports and 50 (PV: 52, ET: 55; MF: 65) clinical notes were identified for 450 test set patients. The test set data extraction accuracy was 95% (reports: 97%, notes: 92%) for PV, 93% (reports: 92%, notes: 94%) for ET and 98% (reports: 99%, notes 97%) for MF. The test set F1 score was 0.95 (reports: 0.96, notes: 0.93) for PV, 0.95 (reports: 0.95, notes: 0.96) for ET and 0.98 (reports: 0.99, notes: 0.97) for MF. PV patients were diagnosed with 100% sensitivity and 98% specificity whereas ET and MF patients were diagnosed with 100% sensitivity and specificity. The accuracy of risk stratification was 100% for PV and ET and 94% for MF. The F1 score for risk stratification was 1.00 for PV and ET and 0.94 for MF. Conclusions The first-in-class automated framework successfully identified pertinent clinical information, performed accurate data extraction from unstructured records and enabled clinician-guided source prioritization for accurate diagnosis and risk stratification of MPN patients. Prospective validation of the framework is required for widespread clinical implementation.

Abstract INTRODUCTION Allogeneic hematopoietic cell transplantation (alloHCT) is indicated in patients (pts) with FLT3-ITD mutated AML. The randomized BMT CTN 1506 trial (Levis et al. JCO) demonstrated that maintenance with FLT3 inhibitor gilteritinib (gilt) significantly improved relapse-free survival (RFS) among pts with detectable peri-alloHCT minimal residual disease (MRD) but was also associated with a greater degree of myelosuppression and infections. In older pts, aged >60 years, increasing post-HCT toxicity may adversely impact outcomes. On this basis, we performed a post-hoc analysis of pts aged > 60 years enrolled onto BMT CTN 1506 to examine the impact of gilteritinib maintenance in older pts. METHODS Pts ≥60 years randomized on BMT CTN 1506 (gilt=47, placebo=57; N=104) were included. The primary endpoint of this post-hoc analysis was RFS. We additionally assessed overall survival (OS), incidence of relapse, non-relapse mortality (NRM) in all pts and in pts with detectable MRD. RESULTS Baseline characteristics were balanced. Age was similar (65.2 vs 64.7 years; p = 0.12), as was prior TKI exposure, and NPM1 mutation status. Myeloablative conditioning (MAC) was used in 8 (17%) pts randomized to gilt and 18 (32%) pts randomized to placebo (p=0.08). Median f/u of survivors was 43.8 months (range:1-61 mo.). Estimated RFS and OS at 48 months were not different between groups (RFS: gilt: 57.0%, 95%CI 44.1-73.7%, placebo: 57.1%, 95%CI 45.4-71.8%; OS: gilt: 56.9%, 95%CI 42.9-46.7%, placebo: 58.4%, 95%CI 46.7-73.1%). While gilt assignment did not affect survival, several predictors of poorer survival were identified including use of reduced intensity conditioning (RIC) (HR = 3.46; p = 0.014), use of calcineurin inhibitor/MTX for graft-vs-host disease (GVHD) prophylaxis (HR = 2.29; p = 0.041), high-risk cytogenetics (HR = 2.64; p = 0.021), and MRD (+) pre-HCT (HR = 2.29; p = 0.018). Finally, relative to gilt assignment, opposing rates of relapse and NRM were observed. The incidence of relapse was lower with gilt (16%, 95%CI 6.8-28% vs 32%, 95%CI 20-45%, p=0.046) but NRM was higher (27%, 95%CI 15-41% vs 11%, 95% CI 4.3-21%, p=0.033). We next analyzed the 49 pts (gilt=22, placebo=27) with detectable peri-HCT MRD. RFS and OS again were not different (RFS: gilt: 44.4%, 95%CI 27.1-72.7%, placebo: 43.8%, 95%CI 28.4-67.5%, OS: gilt: 49.2%,95%CI, 31.3-77.4%, placebo: 55.6%, 95%CI 39.4-77.8%). The incidence of relapse trended lower in gilt pts (25%, 95%CI, 5.3-44.7% vs 45% (95%CI, 25-63%), p=0.2) while NRM trended higher (31%, 95%CI 12-52% vs 11%, 95%CI, 2.7-26%, p=0.12). Seventeen (36%) patients completed planned 2 year maintenance with gilt compared to 29 (51%) completing therapy with placebo. Median time on gilt was 13.2 months. The most common reason for discontinuation was adverse events (n=19, 40%), of which cytopenias (n=8) and infections (n=4) were most frequent. Non-relapse causes of death among gilt treated pts included infections (n=8), intracranial hemorrhage related to thrombocytopenia (n=1), second cancer (n=1), and Guillain-Barre syndrome (n=1). Most infections (n=5) were attributed to GVHD as an underlying cause. CONCLUSION Among pts aged ≥60 years, gilt maintenance did not improve RFS or OS. Gilt use was associated with reduced relapse regardless of peri-HCT MRD status demonstrating disease effect; however, higher NRM limited overall benefit. In older FLT3 mutated pts, employing strategies to mitigate HCT related toxicity, e.g. alternative GVHD prophylaxis strategies, may better accommodate use of maintenance therapy including gilt. Lower doses of gilt (particularly with azole co-use) may be worth exploring to mitigate drug AEs, improve compliance, thus further improving benefit.

Abstract Background: Graft-versus-host disease (GVHD) remains the leading cause of non-relapse morbidity and mortality in allogeneic hematopoietic-stem-cell transplantation (allo-HSCT) recipients. Clinical scoring systems (e.g., EBMT, HCT-CI, DRI, etc.) can stratify risk but show modest predictive performance (C-statistics 0.49–0.67) when it comes to GVHD. To address this gap, we developed GVHD-Intel 1.0, a real-time, scalable, and interpretable machine learning (ML) framework from an international study to predict both acute (aGVHD) and chronic GVHD (cGVHD), providing individual and bulk predictions using common input variables and HLA alleles. Methods: This retrospective global study, approved by respective institutional review boards, included electronic health records from three tertiary care centers in Abu Dhabi, UAE, and King Hussein Cancer Center (KHCC), Jordan, from 2009 to 2023. The model was developed in phases. The UAE399 cohort (399 HSCT recipients) served as the baseline dataset. To enhance model generalizability, 55 adult allo-HSCT recipients from KHCC were added to the training set. Two distinct test cohorts were used: UAE46 (46 allogeneic HSCT recipients from UAE) and KHCC150 (150 adult allogeneic HSCT recipients from KHCC), collected later and independently of training data. Variables included donor type, sex mismatch, disease status, time from diagnosis to transplant, and agents used in conditioning and GVHD prophylaxis (e.g., TBI, ATG, alemtuzumab use etc.). HLA alleles (A, B, C, DQ, DR) were symbolically encoded and transformed via one-hot encoding. GVHD events were classified as aGVHD (<100 days) or cGVHD (≥100 days). Missing values were excluded, and class imbalance was handled via class weighting. GVHD-Intel 1.0, co-developed with MBZ University of Artificial Intelligence, features modular ensemble architecture, drug-level feature engineering, novel HLA encoding for low-resolution compatibility scoring, and built-in interpretability using SHAP and LIME. The framework is cloud-deployable for both real-time and batch predictions. Results: The GVHD-Intel 1.0 was externally tested on two independent, temporally and institutionally distinct cohorts (UAE46 and KHCC150), demonstrating strong and consistent predictive performance for both acute and chronic GVHD. For cGVHD prediction, the model achieved an AUC (area under the curve or C-statistic for binary outcomes) of 0.832, F1 score of 0.896, accuracy of 82.7%, PPV of 0.930, sensitivity of 0.864, Brier score of 0.178, and log loss of 0.543. For aGVHD, performance metrics included an AUC of 0.802, F1 score of 0.860, accuracy of 77.6%, PPV of 0.930, sensitivity of 0.799, Brier score of 0.196, and log loss of 0.581. Model calibration remained strong across both tasks, with Brier scores <0.20. Conclusions: GVHD-Intel 1.0 is a robust, interpretable machine learning framework developed from real-world, multicenter data. By integrating widely available clinical variables and HLA allele data, it outperforms conventional clinical risk scores in predicting both acute and chronic GVHD in allo-HSCT recipients. Slightly reduced performance in aGVHD prediction is likely due to the lack of relevant variables. Next steps include prospective validation across multiple transplant centers globally and continued refinement through expanded data integration. GVHD-Intel 1.0 holds strong potential as a clinically valuable, scalable decision-support tool that can be enacted widely and easily by practicing transplant physicians worldwide in all types of settings.