Citations

Introduction The pathogenic role of somatic non-coding (nc) single nucleotide variants (SNVs) is underexplored in multiple myeloma (MM) and additional characterization of these regions has the potential to provide significant new insights into disease progression. To investigate the role of ncSNVs and their relationship with MM pathogenesis, we have characterized a novel series of WGS from 294 cases of MGUS, SMM, and MM using computational tools optimized to identify functional driver SNVs in these regions. We hypothesize the ncSNVs that co-localize to sites of super-enhancers (SEs) active in plasma cells (PC) are sites of high functional impact SNVs and these mutations lead to the dysregulated expression of genes that should be silent at the terminally differentiated stage of PCs, providing a selective advantage that shapes the fitness of the MM genomic landscape. Methods We analyzed 294 high-coverage matched tumor/normal pairs (80x tumor and 40x normal) derived from 21 MGUS, 86 SMM, 167 NDMM and 40 RRMM. In addition to previously sequenced samples, newly sequenced patients were obtained from NYU Langone Health, Memorial Sloan Kettering Cancer Center, University of Alabama at Birmingham and University of Miami following IRB-approved informed consent. DNA was extracted from CD138+ purified bone marrow plasma cells using standard approaches. We used the MGP1000 (https://github.com/pblaney/mgp1000) to process sequencing data which employs a union-consensus mechanism for determining the final set of variants. We collected enhancer elements (E/SEs) active in GC B-cells, ABC-DLBCL and GBC-DLBCL (Bal, et al. Nature, 2022) and PC specific E/SEs (Loven, et al. Cell 2013). FunSeq2 was used to score ncSNVs. GRAM was used to predict how variants alter TF binding and gene expression in a subset of 6 cases with paired WGS in our series and will expand this analysis on a validation set of 263 NDMM and 93 RRMM cases from University-Hospital Heidelberg enrolled in the GMMG-HD6 clinical trial. Results We identified 1,952,232 high-quality SNVs with 18,199 (0.93%) accumulating in 280 focal sites we termed “mutationally enriched regions” or MERs. To discern SNVs in MERs with the highest potential for imparting a selective advantage, we clustered mutations based on the density of FunSeq2 impact scores which uses reference data contexts on sequence conservation, modification of transcription factor (TF) binding, and gene linkage to regulatory elements and network hubs from pan-cancer studies. We show that MERs are enriched for ncSNVs with high impact scores (mean=2.57) compared to random similarly-sized regions of the genome (Fisher's exact test P<9.7e-12, OR=10.10 [3.88-37.40]). The high impact ncSNVs (hi-ncSNVs) in MERs (n=220) account for 8.09% of all hi-ncSNVs and are present in 46.3% of cases (136/294) with some patients having up to 7. The proportion of cases carrying a MER hi-ncSNVs was similar across disease stages (MGUS=42.9%, SMM=46.5%, MM=40.7%, and RR=47.5%). Analysis of the VAF of MER hi-ncSNVs shows a minor shift towards more clonal versus subclonal mutations (53.6% vs 46.4% VAF>=0.25) suggesting some high impact mutations are acquired in the GC but additional hi-ncSNV are acquired later in PC development, providing evidence that hi-ncSNVs accumulate in MERs throughout disease progression. We mapped these hi-ncSNVs to associated genes then intersected the loci with E/SEs active in GCBs and PCs to determine their functional context. This approach identified 12 genes affected by hi-ncSNVs including the previously identified nc-drivers NEAT1, MIR142, ZCCHC7, and IGLL5 as well as the PC differentiation genes IRF8, BACH2, proliferation genes RHOH, SPRED2, known coding MM drivers CCND1, ZFP36L1, and chr1q genes BTG2 and CROCC. We predicted the effect of ncSNVs on expression through modulation of TF binding at this set of putative nc drivers in MM and identified a trend towards increase in expression of SPRED2, CCND1, NEAT1, and IGLL5 compared to non-mutated cases (P=0.089). We will present data on the validation of this set of genes on a larger set of data derived from the Heidelberg group. Conclusions We identified a set of 12 putative nc driver genes in MM that modify expression and have the potential to significantly alter the pathogenesis of MM by providing a selective survival advantage. We will present the impact of these by clinical stage and on outcome.

Abstract Dexamethasone is a key component of induction for newly diagnosed multiple myeloma (NDMM), despite common toxicities, including hyperglycemia and insomnia. In the randomized ECOG E4A03 trial, dexamethasone 40 mg once weekly was associated with lower mortality than higher doses. However, the performance of dexamethasone dose reductions below this threshold with regard to progression-free survival (PFS) and overall survival (OS) in NDMM has not been fully characterized. We conducted a secondary pooled analysis of the SWOG 0777 and SWOG 1211 studies of NDMM, which used lenalidomide and dexamethasone (Rd) alone, with or without bortezomib, and with or without elotuzumab. The planned dexamethasone intensity was 40 to 60 mg weekly in all arms. Patients were categorized into FD-DEX (full-dose dexamethasone maintained throughout induction) or LD-DEX (lowered-dose dexamethasone or discontinuation; only permitted for grade 3+ toxicities per both study protocols). Of the 541 evaluated patients, the LD-DEX group comprised 373 patients (69%). There were no differences in PFS or OS between the FD-DEX and LD-DEX groups, which were balanced in terms of age, stage, and performance status. Predictors of PFS and OS in the multivariate models were treatment arm, age ≥70 years, and thrombocytopenia. FD-DEX did not significantly improve either outcome. Our study suggests that dexamethasone dose reductions are common in multiple myeloma, even within clinical trials. Given the many toxicities and unclear benefits of dexamethasone in the era of modern treatment regimens, dexamethasone dose reduction during NDMM induction warrants further prospective studies. These trials were registered at www.clinicaltrials.gov as #NCT00644228 and NCT01668719.

Abstract Artificial intelligence (AI)–enabled interpretation of electrocardiogram (ECG) images (AI-ECGs) can identify patterns predictive of future adverse cardiac events. We hypothesized that such an approach would provide prognostic information for the risk of cardiac complications and mortality in patients undergoing hematopoietic cell transplantation (HCT). We retrospectively subjected ECGs obtained before HCT to an externally trained, deep-learning model designed to predict the risk of atrial fibrillation (AF). Included were 1377 patients (849 autologous [auto] HCT and 528 allogeneic [allo] HCT recipients). The median follow-up was 2.9 years. The 3-year cumulative incidence of AF was 9% (95% confidence interval [CI], 7-12) in patients who underwent auto HCT and 13% (10%-16%) in patients who underwent allo HCT. In the entire cohort, pre-HCT AI-ECG estimate of AF risk correlated highly with the development of clinical AF (hazard ratio [HR], 7.37; 95% CI, 3.53-15.4; P < .001), inferior survival (HR, 2.4; 95% CI, 1.3-4.5; P = .004), and greater risk of nonrelapse mortality (NRM; HR, 95% CI, 3.36; 1.39-8.13; P = .007), without increased risk of relapse. Association with mortality was only noted in allo HCT recipients, where the risk of NRM was greater. The use of cyclophosphamide after transplantation resulted in greater 90-day incidence of AF (13% vs 5%; P = .01) compared to calcineurin inhibitor–based graft-versus-host disease prophylaxis, corresponding to temporal changes in AI-ECG AF prediction after HCT. In summary, AI-ECG can inform risk of posttransplant cardiac outcomes and survival in HCT patients and represents a novel strategy for personalized risk assessment.

PURPOSE Outcomes for patients with newly diagnosed multiple myeloma (NDMM) are heterogenous, with overall survival (OS) ranging from months to over 10 years. METHODS To decipher and predict the molecular and clinical heterogeneity of NDMM, we assembled a series of 1,933 patients with available clinical, genomic, and therapeutic data. RESULTS Leveraging a comprehensive catalog of genomic drivers, we identified 12 groups, expanding on previous gene expression–based molecular classifications. To build a model predicting individualized risk in NDMM (IRMMa), we integrated clinical, genomic, and treatment variables. To correct for time-dependent variables, including high-dose melphalan followed by autologous stem-cell transplantation (HDM-ASCT), and maintenance therapy, a multi-state model was designed. The IRMMa model accuracy was significantly higher than all comparator prognostic models, with a c-index for OS of 0.726, compared with International Staging System (ISS; 0.61), revised-ISS (0.572), and R2-ISS (0.625). Integral to model accuracy was 20 genomic features, including 1q21 gain/amp, del 1p, TP53 loss, NSD2 translocations, APOBEC mutational signatures, and copy-number signatures (reflecting the complex structural variant chromothripsis). IRMMa accuracy and superiority compared with other prognostic models were validated on 256 patients enrolled in the GMMG-HD6 (ClinicalTrials.gov identifier: NCT02495922 ) clinical trial. Individualized patient risks were significantly affected across the 12 genomic groups by different treatment strategies (ie, treatment variance), which was used to identify patients for whom HDM-ASCT is particularly effective versus patients for whom the impact is limited. CONCLUSION Integrating clinical, demographic, genomic, and therapeutic data, to our knowledge, we have developed the first individualized risk-prediction model enabling personally tailored therapeutic decisions for patients with NDMM.

Background: The management of multiple myeloma (MM) continues to evolve with the emergence of new classes of therapy, most recently including BCMA-targeted therapies. Considering the expanding MM treatment armamentarium, we sought to identify contemporary practice patterns among healthcare professionals (HCPs) for MM and compare them with those of experts using an online decision support tool. Methods: An online decision support tool was developed with input from 5 experts providing therapy recommendations for 192 unique patient case scenarios based on considerations including disease setting, treatment history, access to therapy, and specific comorbidities. HCP tool users entered specific patient criteria to define a case along with their intended management for that case. The tool then showed the 5 expert recommendations for that same case scenario, and users were asked if the recommendations impacted their intended approach. An analysis of expert recommendations and user-selected therapy was performed. Results: Between April 2023 and July 2023, 207 patient cases were entered by 144 participating HCPs. HCP and expert treatment recommendations were frequently discordant, regardless of treatment setting (Table). The highest concordance between HCP and expert recommendations (47%; P <.0001) was in the induction setting (n = 111). In this setting, experts were more likely to recommend quadruplet regimens compared with HCPs (62% vs 27%, respectively; P <.0001). Treatment selection concordance was lower in the maintenance (n = 17; 24%; P <.0001) and relapsed/refractory settings (n = 79; 20%; P <.0001). The lowest concordance (10%; P <.0001) between HCP and expert treatment recommendations was noted in the relapsed/refractory setting specifically among patients who were refractory to earlier lines of therapy (n = 41; P <.0001). For patients who are refractory to earlier lines of therapy, experts were more likely to recommend an anti-CD38-based combination regimen compared with HCPs (81% vs 32; P <.0001). When recommending treatment for patients with triple- or penta-refractory disease, experts were significantly more likely to recommend BCMA-targeted agents (83% vs 39%; P <.0001). For HCPs whose treatment plan did not match expert recommendations in these scenarios, 53% of respondents indicated that expert recommendations confirmed or changed their intended therapy, but 31% indicated that there were barriers to implementing those recommendations. Conclusions: These data suggest ongoing challenges with incorporating the newest therapies and combinations across the spectrum of MM into patient care, particularly multidrug regimens in the induction and maintenance settings and BCMA-targeted therapies in those with triple- or penta-refractory disease. Continued education and development of resources for HCPs, including online decision support tools, may be increasingly important as the treatment of MM continues to grow in complexity.

Background: Despite the development and approval of effective therapies for multiple myeloma (MM), MM remains an incurable disease. Moreover, patients with high-risk cytogenetics, like translocation t(4;14), continue to have a poor prognosis compared to standard-risk patients. No novel agents have been approved specifically for this population. There remains an unmet need for precision therapies for MM that can target specific drivers of high-risk MM like t(4;14). Multiple myeloma SET domain-containing protein (MMSET) overexpression observed in 100% of patients with t(4;14) acts through the MMSET-H3K36me2 axis, which promotes malignant transformation through regulation of multiple oncogenic programs. KTX-1001, a methyl transferase inhibitor, is a novel, first-in-class, potent, oral, small molecule that inhibits MMSET and the methylation of lysine K36. KTX-1001 had favorable drug-like properties, demonstrated in vivo activity, and was well tolerated in non-clinical studies. Based on these data, KTX-1001 is now being studied in a Phase 1 dose escalation and expansion study in relapsed and refractory multiple myeloma (RRMM). Study Design and Methods: This multi-country, Phase 1, open-label study (NCT05651932) is designed to assess the safety, tolerability, and initial clinical activity of KTX-1001. The study will also examine the recommended phase 2 dose (RP2D), pharmacokinetics (PK) and pharmacodynamics (PD). Part A of this trial includes an accelerated titration, followed by 3+3 dose escalation with backfill of cohorts for dose optimization prior to moving to Part B, which is an expansion phase at the RP2D (please see Figure 1 for study design schematic). The expansion phase will only include patients with t(4;14). KTX-1001 will be administered orally daily and given until disease progression, unacceptable toxicity, or withdrawal from the study. Estimated enrollment for Parts A and B will be approximately 60 patients. The study will enroll patients with RRMM who have received ≥3 prior lines of therapy, including a proteasome inhibitor (PI), immunomodulatory drug (IMiD), and anti-CD38 antibody. Patients must have measurable disease, be ≥18 years of age, have an Eastern Cooperative Oncology Group score of ≤2, and adequate bone marrow, liver, and kidney function. Key exclusion criteria include history of or current plasma cell leukemia, active central nervous system (CNS) disease, or active unstable cardiovascular function. The primary endpoints of safety and tolerability include incidence of dose-limiting toxicity (DLTs), treatment-emergent adverse events (AEs), and treatment-related AEs. Secondary endpoints include plasma concentrations of KTX-1001 and efficacy outcomes such as objective response rate (ORR), duration of response (DOR), progression-free survival (PFS), and overall survival (OS). Exploratory biomarker and pharmacodynamic assessments will also be performed. The sample size was not determined by formal statistical methods but will be sufficient to provide preliminary information on safety, PK, PD, and efficacy. Statistical analyses will be primarily descriptive. Conclusion: This Phase 1 study will evaluate a potential treatment for RRMM utilizing KTX-1001, which is a novel, first-in-class, potent, oral small-molecule MMSET inhibitor. The study is open and actively enrolling patients at centers in the United States, Spain, France, and Canada.

Introduction Patients (pts) with multiple myeloma (MM) have a higher risk of progression/death following an inadequate response to frontline ASCT (van de Velde Eur J Haematol 2017). Therapies (Tx) with novel mechanisms of action may further improve response outcomes. Ide-cel previously demonstrated frequent, deep, and durable responses in KarMMa (Munshi N Engl J Med 2021) and significantly improved median progression-free survival (PFS) and overall response rate (ORR) versus standard regimens in KarMMa-3 (Rodríguez-Otero N Engl J Med 2023) in pts with triple-class-exposed relapsed and refractory MM. KarMMa-2 (NCT03601078) is a multicohort, phase 2, multicenter trial evaluating the efficacy and safety of ide-cel in pts with MM; cohort 2c included pts with NDMM who had an inadequate response to frontline Tx with ASCT. In cohort 2c, with a median follow-up of 27.9 months, ide-cel demonstrated deep and durable responses; complete response (CR) rate was 74.2% (ORR, 87.1%), with a 24-month PFS rate of 83.1% (Dhodapkar Blood 2022). Here, we report results with an additional 11.5 months of median follow-up for updated efficacy and safety data, and health-related quality of life (HRQoL) outcomes in cohort 2c. Methods Eligible pts had NDMM, received ≥ 3 cycles of induction Tx (including a proteasome inhibitor, an immunomodulatory agent, and dexamethasone), had < very good partial response (VGPR) 70-110 days after ASCT (single/tandem), and had Eastern Cooperative Oncology Group performance status ≤ 1. Pts received a single infusion of ide-cel (dose range: 150-450 × 10 6 CAR+ T cells) after lymphodepletion and could receive maintenance Tx post-infusion at investigator's discretion. The primary endpoint CR rate and secondary endpoints ORR, duration of response (DOR), PFS, overall survival (OS), safety, and patient-reported outcomes on HRQoL, assessed via European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Core 30 Questionnaire (QLQ-C30) and EORTC Quality of Life Multiple Myeloma Module 20 Questionnaire (QLQ-MY20), are reported. The exploratory endpoint of minimal residual disease (MRD) negativity (sensitivity, 10 -5) is also reported. Results At data cutoff (May 3, 2023), with a median follow-up of 39.4 (range, 31.6-44.7) months, all 31 pts who were infused with ide-cel were alive; 28 (90.3%) remained in follow-up, with 3 (9.7%) having discontinued due to disease progression. The ORR and CR rates were 87.1% and 77.4%, respectively ( Table); best overall response in 1 pt improved from VGPR to stringent CR. The median DOR was not reached (NR; 95% CI, NR-NR), neither was median PFS (95% CI, 38.0-NR) or median OS (95% CI, NR-NR); the 36-month event-free rates were 80.9%, 76.8%, and 100.0%, respectively. In pts who did not receive lenalidomide (LEN) maintenance, 9 (29.0%) had disease progression; of the 8 pts who received LEN maintenance post ide-cel infusion, none experienced disease progression. At 36 months, MRD negativity was confirmed in 9/14 (64.3%) evaluable pts; all 9 achieved ≥ CR. In this extended follow-up, safety results were generally consistent with the previous data cutoff. All pts experienced ≥ 1 any-grade adverse event (AE), most commonly hematologic, including neutropenia (80.6%), anemia (38.7%), leukopenia (32.3%), and thrombocytopenia (29.0%); infection and infestation AEs occurred in 74.2% of pts. No grade 5 AEs were reported. Incidence of any-grade cytokine release syndrome (58.1%; no grade ≥ 3) was unchanged and there were no additional reports of parkinsonism since the previous report. Overall, HRQoL improved ( Figure) and MM disease symptoms were stable over time based on the EORTC QLQ-C30 and QLQ-MY20, respectively, following ide-cel treatment; by month 3, > 75% of pts had improved or stable global health status/QoL. Conclusions Ide-cel continued to demonstrate deep, durable responses in pts with an inadequate response to frontline ASCT, no new safety signals were observed with extended follow-up, and no deaths were reported. No pts who received LEN maintenance post ide-cel experienced disease progression. Ide-cel treatment resulted in stable or improved HRQoL. Overall, ide-cel continued to demonstrate a favorable clinical benefit-risk profile in NDMM after ASCT. Study support 2seventy bio and Celgene, a Bristol-Myers Squibb Company.

Background: Dex is a key component of induction regimens for NDMM despite common toxicities including insomnia and anxiety. The E4A03 randomized trial (Rajkumar, Lancet Oncol 2010) demonstrated that dex dosed at 40 mg weekly outperformed higher-intensity dex and established this regimen as the standard of care for NDMM. In standard practice as well as clinical trials, however, dex doses are often reduced to 20 mg weekly or lower (if not stopped entirely) among older patients or among patients experiencing toxicities. As a first step toward establishing the equipoise of planned dex dose reductions in NDMM, we investigated outcomes with patients from two previous SWOG trials to understand the impact of dex dose reductions on post-induction outcomes. Methods: This was a secondary analysis of data from two completed NDMM studies, S0777 (comparing VRd vs Rd) and S1211 (comparing Elo-VRd vs VRd). For this secondary analysis, we analyzed the subset of patients who completed all eight 3-week cycles of induction (or six 4-week cycles in the S0777 Rd arm). As shown in Figure 1A, we defined DEX fullDose as patients for whom the starting dex dose (either 40 mg once per week or 20 mg on Days 1, 2, 4, 5, 8, 9, 11, and 12 of 21-day cycles) did not change during induction. In contrast, DEX lowered was defined as patients for whom the end-of-induction dex dose and/or frequency were lower than their starting values. As an exploratory analysis, we also compared PFS and OS between the DEX fullDose (no changes to dex strength during induction) and DEX veryLow (dex dosing reduced ≥50% during induction) subgroups. All analyses were landmarked from the end of induction, defined as 24 weeks from registration, to account for the time-dependent nature of dex treatment received. Survival curves were based on the Kaplan-Meier method and comparisons were made using the log-rank test. Results: Of 541 evaluable patients ( n = 441 from S0777, n = 100 from S1211), dex was reduced during induction in 76% ( n = 410). Compared to DEX fullDose patients, DEX lowered patients were more likely to be anemic (37% vs 26%, p = 0.02) but had similar distributions of age ≥70 (24% of DEX lowered vs 26% of DEX fullDose), Performance Status > 1 (15% vs 12%), ISS stage 3 disease (33% vs 27%), ≥ 60% bone marrow plasma cell burden (37% vs 31%), M-spike > 3 g/dL (48% vs 48%) platelets < 150,000 cells/mm 3 (18% vs 14%), baseline LDH ≥ 190 U/L (36% vs 41%), and baseline CRP ≥ 8 mg/L (27% vs 30%). PFS (Figure 1B) was comparable between DEX lowered (median 36 months, 95% CI 29 to 39 months) and DEX fullDose (median 33 months, 95% CI 24 to 48 months) with p = 0.90. Similarly, OS (Figure 1C) was comparable between DEX lowered [median 77 months, 95% CI 68 months to not reached (NR)] and DEX fullDose (median 92 months, 95% CI 65 months to NR) with p = 0.93. In the DEX veryLow subgroup with ≥50% reduction in dex dosing during induction, median PFS was 39 months (95% CI 34 to 50 months) and median OS was NR (95% CI NR to NR). As shown in Figure 1D and Figure 1E, PFS and OS were comparable between these patients and DEX fullDose patients. Discussion: To our knowledge, this is the first analysis of dex dosing strength in NDMM since E4A03. In two separate randomized trials, over half of patients required dex reductions below a starting dose of ~40 mg weekly. Dex dose reductions during induction had no impact on post-induction PFS and OS, even in the subset of patients who required significant dose reductions. Limitations of our post hoc analysis include heterogeneity between dex dosing schedules and lack of cytogenetic data for all patients. We could not discern specifically why dex was lowered, and it is possible that dex might have been lowered disproportionately commonly in patients who had achieved early responses and thus independently did better. Regardless, these results strongly suggest that maintaining dex 40 mg weekly for the entirety of NDMM induction may be unnecessary in the modern era. Cooperative efforts to study a planned dex de-escalation protocol (i.e., once no longer needed for symptom management or as a pre-medication) alongside patient-reported outcome assessments of symptom burden are being planned. Funding: NIH/NCI/NCTN grants U10CA180888, U10CA180819, U10CA180821, U10CA180820; and in part by Millennium Pharmaceuticals, Inc. (now part of Takeda Pharmaceutical Company), and by Celgene.

High dose chemotherapy followed by autologous hematopoietic cell transplantation (auto HCT) is a standard therapy for patients with plasma cell disorders (PCD). A barrier to improving candidate selection for auto HCT is appraisal of risk for transplant related complications. Electrocardiography (ECG) is a standard pre-HCT test to evaluate the potential for arrhythmia and structural heart disease. We hypothesized that an artificial intelligence (AI) based interpretation of ECG waveform data (AI-ECG) could enhance risk prediction for cardiac-specific and other outcomes following auto HCT for patients with PCD. To test this hypothesis, we used several AI-ECG tools developed using a deep-learning based approach that were trained to predict comorbidity (Attia et al, Lancet 2019 & Grogan et al, Mayo Clin Proc 2021). Model outputs include a scaled prediction value indicating increasing risk for atrial fibrillation/flutter (AF), reduced ejection fraction (EF), or cardiac amyloidosis. We evaluated this tool in 9,225 ECGs obtained from 849 patients (median 6 ECGs/patient, inter-quartile range: 3-12) who underwent first or subsequent auto HCT at Memorial Sloan Kettering Cancer Center between 2016-2022 and who had a diagnosis of multiple myeloma (N = 810), amyloidosis (N = 37), or other PCD (N = 2). The endpoints were to determine whether baseline AI-ECG prediction of AF risk correlated with peri-transplant AF events, and to explore whether model outputs correlated with survival and other cardiac clinical outcomes. The mean age of patients was 61.0 years (range 23.7 - 81.2) and median follow up was 2.7 years post auto HCT. Contributing cardiovascular risk factors included a prior history of AF (N = 101, 11.9%), coronary or cerebrovascular disease (N = 95, 11.2%), and diabetes mellitus (N = 98, 11.5%). Reduced dose melphalan (< 200 mg/m 2) was administered in 26%. The HCT-comorbidity index (HCT-CI) was high risk (≥3) in 52.7%. The one-year overall survival and progression-free survival in the whole cohort was 94% (95% confidence interval (95% CI): 92 - 95%) and 86% (84 - 89%), respectively. Among patients not in AF at pre HCT evaluation, the 90-day cumulative incidence of AF was 5.6% (4.2 - 7.3%). In univariable competing risk regression, baseline pre HCT AI-ECG appraisal of AF risk was strongly correlated with incidence AF in patients who were in sinus rhythm before HCT (hazard ratio (HR): 25.1, 95% CI: 11.7 - 53.7, p < 0.001). In a multivariable model adjusting for patient age and HCT-CI, AI-ECG AF predicted risk remained associated with incidence of post HCT AF (HR: 13.4, 95% CI: 5.9 - 30.4, p < 0.001). Considering a decision threshold identified during original model training, AI-ECG predicted AF events with an accuracy of 75% (Mcnemar's p < 2 x 10 -16). The precision, recall, and F1-score were 0.94, 0.77, and 0.85, respectively. AI-ECG predicted age correlated highly with actual patient age (Spearman rho = 0.59, p < 2 x 10 -16). Amyloid risk scores were highly correlated with a known clinical history of amyloidosis (median 0.45 versus 0.05, Wilcoxon rank-sum p < 0.001). Patients with EF < 50% had greater prediction scores for cardiomyopathy compared to patients with normal EF (median 0.23 versus 0.01, Wilcoxon rank-sum p <0.001). In univariable analysis, increasing AI-ECG predicted age (HR 1.02, CI: 1.003 - 1.04, p = 0.02) and predicted AF risk score (HR: 3.2, CI: 1.4 - 7.2, p = 0.006) were associated with risk for death, whereas predicted amyloid score and predicted low EF were not associated with decreased survival. In multivariable analysis of all patients that considered patient age, hematopoietic comorbidity index, and remission status at HCT, increasing predicted AF score trended towards association with increased risk for death (HR: 2.3, 95%CI: 1.0 - 5.5, p = 0.059). Predicted AF score correlated highly with the HCT-CI and patient age. Finally, we examined whether risk prediction for low EF correlated with peri-HCT fluid overload (FO) defined by absolute weight gain. The incidence of ≥ 10% weight gain (grade 2 FO) and ≥ 20% weight gain (grade 3 FO) at 180 days post HCT was 7.4% (95% CI: 5.8%, 9.3%) and 1.0% (95% CI: 0.5%, 1.8%), respectively. Predicted low EF was not associated with development of either grade 2 or grade 3 FO in this population. In summary, AI-ECG predicts episodes of AF in patients undergoing auto HCT for PCD and may augment transplant risk appraisal when considered in conjunction with other measures of comorbidity.