Abdelrheem, Ahmed

ABSTRACT Predicting delayed graft function (DGF) relies on donor, recipient, and perioperative factors. Despite the growing recognition that DGF duration strongly influences patient outcomes, no models currently address or predict its impact—highlighting an important gap in current research and clinical practice. This study aimed to develop an ensemble‐based machine learning model using perioperative data to predict DGF occurrence and duration. The gradient‐boosted decision trees model was trained and validated on 2725 patients, with k‐fold cross‐validation in an external cohort of 284 patients. Model performance was evaluated based on accuracy, ROC‐AUC, and other metrics using R and Python libraries. The binary DGF prediction model achieved a ROC‐AUC of 0.77, while the DGF duration classification model had an accuracy of 79.2%. DGF duration AUC values by time interval were 0.76 for 0–1 weeks, 0.81 for >1–2 weeks, 0.80 for >2–3 weeks, and 0.87 for >3 weeks; the overall macro‐averaged AUC was 0.81. The mean Brier score for multi‐class predictions was 0.14. External validation showed a 78% accuracy for DGF duration prediction. Acute kidney injury (AKI) and donor donation after circulatory death (DCD) status were key DGF predictors. The use of gradient‐boosted decision trees (GBDT) improves the prediction of both the likelihood and duration of DGF, addressing a current gap in kidney transplant patient care. By facilitating personalized transplant care, this model supports more effective perioperative planning and timely interventions, which may contribute to better patient outcomes.

Abstract Background A previous multicenter study of patients with TP53-mutated myeloproliferative neoplasms showed a significantly shorter survival in chronic phase myelofibrosis (MF-CP) that included primary (PMF) and secondary MF and was mostly based on research-based mutation screening and was not systematically controlled for patients without TP53 mutations (AJH 2025;100:552). The objective of the current study was to examine the prognostic impact of TP53 mutations (VAF ≥2%) that were discovered during clinical NGS testing among consecutive patients with PMF-CP. Impact on overall (OS) and leukemia-free (LFS) survival was analyzed in the context of other genetic risk factors and calculated from the time of mutation detection. Methods The current retrospective study includes consecutive patients with PMF-CP who underwent clinical NGS testing at the Mayo Clinic between 2015 and 2025. Diagnoses of PMF-CP and leukemic transformation (LT) were retrospectively fitted to be in accordance with the International Consensus Classification criteria (Blood 2022;140:1200). Mutation and cytogenetic analyses were performed according to institutional protocols for clinical use and reported in conformity with international guidelines. OS and LFS analyses were censored for treatment with allogeneic stem cell transplantation (ASCT) and calculated from the date of NGS. Conventional statistical methods were utilized (JMP Pro 17.1.0, SAS Institute, Cary, NC, USA). Results A total of 633 patients with PMF-CP underwent clinical NGS testing prior to LT and prior to ASCT (median age 65 years; males 59%). Median interval from initial diagnosis/referral to NGS testing was 2 months; 357 (56%), 392 (62%), and 429 (68%) patients had NGS performed within 3, 6 and 12 months of diagnosis, respectively. Risk distribution according to the dynamic International Prognostic Scoring System (DIPSS; Blood 2010;115:1703) were low (19%), intermediate-1 (44%), intermediate-2 (31%), or high (6%). TP53 mutations (TP53MUT) with variant allele frequency (VAF) of ≥2% were detected in 28 (4.4%) patients with a median VAF of 25.3% (range 2.4-85; ICC-defined multihit 43%). Patients with TP53MUT were more likely to harbor IDH2MUT (14% vs. 3%; p<0.01), EZH2MUT (14% vs. 4%; p=0.03), and unfavorable karyotype (36% vs. 19%; p=0.047) and less likely to harbor CALR type 1/like (4% vs. 18%; p=0.02) and SF3B1MUT (0% vs. 6%; p=0.07). The association with unfavorable karyotype was mainly driven by multihit configuration. No significant associations with clinical features were apparent. At a median follow-up of 40 months for surviving patients, 157 (25%) deaths, 48 (8%) LTs and 135 (21%) ASCTs were documented. A near-two-fold lower, albeit not statistically significant OS was observed in the presence of TP53MUT (median 123 vs. 228 months; HR 1.9, 95% CI 0.94-3.7; p=0.074); the difference reached significance for multihit TP53MUT (median 81 months; HR 3.6, 1.7-7.7; p<0.01) but not for non-multihit TP53MUT (median not reached; HR 0.7, 0.2-2.8; p=0.6), both compared to TP53WT (median 228 months). The significant impact of multihit TP53MUT on OS was confirmed in multivariable analysis (MVA) that included other high molecular risk (HMR) mutations and unfavorable karyotype: multihit TP53MUT (HR 2.5, 1.1-5.5; p=0.02), SRSF2MUT (2.4, 1.5-3.8; p<0.01), ASXL1MUT (1.8, 1.2-2.6; p<0.01), U2AF1MUT-Q157 (4.8, 2.6-8.6; p<0.01), and absence of CALR type 1/like (4.0, 2.0-7.8; p<0.01); additional prognostic contribution was not evident for unfavorable karyotype (p=0.17), EZH2MUT (p=0.3), IDH1MUT (p=0.6), IDH2MUT (p=0.7), or any other HMR mutation. Multihit TP53MUT was also associated with inferior LFS (10-year risk 62% vs. 32% for non-multihit TP53MUT vs. 15% for TP53WT; p=0.03) but significance was lost during MVA that included SRSF2MUT, which was the only HMR mutation showing a significant effect on LFS (HR 5.0, 2.5-9.8; p<0.01). Post-ASCT survival among 135 transplanted patients was significantly shorter for patients with multihit TP53MUT (N=4; median 5 months) vs. those with non-multihit TP53MUT (N=7; median not reached) vs. patients with TP53WT (N=124; median not reached) [p=0.02]. Conclusion The current study implicates multihit TP53MUT as an HMR mutation-independent prognostic factor for OS in PMF-CP. By contrast, prognosis in PMF-CP patients with non-multihit TP53MUT was more likely to be influenced by other HMR mutations, as was the case for those with TP53WT.

Abstract Background Myeloproliferative neoplasms (MPNs) can progress from chronic phase to accelerated or blast phase (MPN-AP/BP), a transition with poor prognosis and limited treatment options. While some abnormalities have been described, the genomic and clonal drivers of transformation remain poorly defined. We characterized mutational and structural alterations in MPN-AP/BP, leveraging paired samples to trace clonal evolution. Methods We profiled 186 samples from 160 individuals, including 159 MPN-AP/BP and 27 chronic-phase MPN samples (26 MF, 1 PV), using targeted sequencing of a 588 gene panel (mean depth: 709x). Paired pre- and post-transformation samples were available for 26 patients (n= 52). A consensus pipeline was used to detect somatic mutations, copy number variations (CNVs), and copy-neutral loss of heterozygosity (cnLOH). Clinical cytogenetic data were integrated when available. Single-cell validation is ongoing. Results Genomic complexity was widespread: 99% of patients harbored ≥1 oncogenic mutation (median 4; range 1–10). In total, 859 mutations were identified, most commonly in JAK2 (68%), ASXL1 (28%), TET2 (26%), SRSF2 (24%), TP53 (22%), and RUNX1 (22%). 684 structural alterations were identified in 86% of patients (median 3; range 1-20), including arm-level (22.5%), segmental (>5 Mb; 60%), and focal (<2 Mb; 9%) alterations. Recurrent events included 9p cnLOH (16%), del(5q) (13%), del(17p) (10%), and focal deletion of 21q22.12 involving RUNX1 (10%). Complex karyotypes (≥3 abnormalities) were observed in 55% of patients. Biallelic mutations accounted for 21% (183/859) of mutations, affecting 58% of patients across 21 genes, most frequently JAK2 V617F (28%), TP53 (17%) and TET2 (14%). Patients with biallelic mutations in these genes had significantly higher blast percentages compared to those with monoallelic mutations (p<0.05), highlighting an association with aggressive disease. In 26 patients with longitudinal samples (chronic phase MPN and MPN-AP/BP), 87% (125/144) of mutations were already detectable during the chronic phase, on average more than 7 years prior to transformation. Only 11% (16/144) of mutations were newly acquired, including additional RUNX1 and TP53 hits, or structural changes that converted pre-existing JAK2 V617F and TP53 R248Q mutations from monoallelic to biallelic via 9p cnLOH or 17p loss, respectively. These findings suggest transformation is often driven by outgrowth of pre-existing high-risk clones, indicating that these clones likely have a competitive advantage and that the mutations within these clones have a cooperative biological interaction which underlies this competitive advantage. To investigate this hypothesis, we examined whether patterns of recurrent co-mutations were identifiable in this cohort. We identified recurrent co-occurring mutations in SRSF2-IDH2, ASXL1-SRSF2, and ASXL1-EZH2, which were enriched at transformation (p < 0.05; Fisher exact test). Importantly, presence of these co-mutations was consistently associated with a higher transformation risk in two independent cohorts of MF patients from Mayo Clinic (n= 405) and University of Florence (n= 518). Clonal architecture analysis revealed that these co-mutations frequently arose within the same malignant clone: 100% for SRSF2-IDH2, 80% for ASXL1-SRSF2, and 85% for ASXL1-EZH2, and are often within the dominant clone. Prior studies have demonstrated mechanistic interplay between SRSF2 and IDH2 (Yoshimi et al.), as well as ASXL1 and SRSF2 (Sui et al.) mutations in myeloid malignancies. We assessed functional synergy of ASXL1-EZH2 using a murine model and identified that dual deletion caused a rapidly fatal myeloid neoplasm with shorter survival than single-gene loss. These results reinforce that functional synergy emerges within a shared cellular context, supporting clonal co-dependence as a driver of progression. Single-cell RNA sequencing of ASXL1-EZH2 mutant cases is ongoing and will be presented. Conclusions Transformation to MPN-AP/BP is driven by gradual clonal remodeling and expansion of pre-existing high-risk abnormalities. Biallelic and multi-hit alterations in TP53, JAK2, TET2, and RUNX1 promote clonal dominance and blast progression. Recurrent co-mutations such as IDH2-SRSF2 and ASXL1-EZH2 arise within the same clone, exhibit functional synergy, and are linked to poor outcomes. These findings support the clinical utility of early clonal and co-mutational profiling to guide risk-adapted intervention.

ABSTRACTIntroductionThe incidence of mortality late in the pandemic, particularly after widespread vaccine availability, is not well understood. Herein, we elucidate the effect of this impact of the COVID pandemic as well as risk factors for mortality during it.MethodsThe primary end point was death with a functioning graft with secondary endpoints of mortality rates in subgroups and at different time intervals during the pandemic.ResultsDespite widespread vaccination, mortality rates for kidney transplant (KTx) recipients almost doubled during the COVID‐19 era (6.40 deaths per 100 person years vs. 3.54 pre‐COVID). Mortality increased in all racial/ethnic groups but increased more in Native Americans, Hispanics, and African Americans compared to non‐Hispanic Caucasians. The highest mortality rate occurred during the Delta and Omicron time frames. In contrast to the general population, mortality was more evenly spread across age groups in KTx recipients.ConclusionsMortality rates during the COVID‐19 era were extremely high, more than doubling in some racial/ethnic groups. We conclude that the KTx population was a particularly vulnerable group during the COVID‐19 era and suggests the need for further research into the management of variants in the future.