We investigated the long-term safety and disease control data obtained with i.v. busulfan (Bu) combined with clofarabine (Clo) in patients with acute lymphoblastic leukemia (ALL) undergoing allogeneic hematopoietic stem cell transplantation (HSCT). A total of 107 patients, median age 38 years (range, 19 to 64 years) received a matched sibling donor (n = 52) or matched unrelated donor (n = 55) transplant for ALL in first complete remission (n = 62), second complete remission (n = 28), or more advanced disease (n = 17). Nearly one-half of the patients had a high-risk cytogenetic profile as defined by the presence of t(9;22) (n = 34), t(4;11) (n = 4), or complex cytogenetics (n = 7). Clo 40 mg/m2 was given once daily, with each dose followed by pharmacokinetically dosed Bu infused over 3 hours daily for 4 days, followed by hematopoietic cell infusion after 2 days of rest. The Bu dose was based on the drug clearance as determined by a test Bu dose of 32 mg/m2. The target daily area under the curve was 5500 µmol/min for patients aged <60 years and 4000 µmol/min for patients aged >59 years. With a median follow-up of 3.3 years among surviving patients (range, 1 to 5.8 years), the 2-year progression-free survival (PFS) for patients undergoing HSCT in first complete remission (CR1), second complete remission (CR2), or more advanced disease was 62%, 34%, and 35%, respectively. The regimen was well tolerated, with nonrelapse mortality (NRM) of 10% at 100 days and 31% at 2 years post-HSCT. The incidence of grade II-IV and III-IV acute graft-versus-host disease (GVHD) was 35% and 10%, respectively; 18% patients developed extensive chronic GVHD. The 2-year overall survival (OS) for patients undergoing HSCT in CR1, CR2, or more advanced disease was 70%, 57%, and 35%, respectively. Among 11 patients aged >59 years treated with reduced-dose Bu in CR1 (n = 7) or CR2 (n = 4), 4 remain alive and disease-free, with a median follow-up of 2.6 years (range, 2 to 4.7 years). Only the presence of minimal residual disease at the time of transplantation was associated with significantly worse PFS and OS in multivariate analysis. Our data indicate that the Clo-Bu combination provides effective disease control while maintaining a favorable safety profile. OS and NRM rates compare favorably with those for traditional myeloablative total body irradiation-based conditioning regimens.

Polygenic risk scores (PRS) are commonly used to quantify the inherited susceptibility for a trait, yet they fail to account for non-linear and interaction effects between single nucleotide polymorphisms (SNPs). We address this via a machine learning approach, validated in nine complex phenotypes in a multi-ancestry population. We use an ensemble method of SNP selection followed by gradient boosted trees (XGBoost) to allow for non-linearities and interaction effects. We compare our results to the standard, linear PRS model developed using PRSice, LDpred2, and lassosum2. Combining a PRS as a feature in an XGBoost model results in a relative increase in the percentage variance explained compared to the standard linear PRS model by 22% for height, 27% for HDL cholesterol, 43% for body mass index, 50% for sleep duration, 58% for systolic blood pressure, 64% for total cholesterol, 66% for triglycerides, 77% for LDL cholesterol, and 100% for diastolic blood pressure. Multi-ancestry trained models perform similarly to specific racial/ethnic group trained models and are consistently superior to the standard linear PRS models. This work demonstrates an effective method to account for non-linearities and interaction effects in genetics-based prediction models.

We construct non-linear machine learning (ML) prediction models for systolic and diastolic blood pressure (SBP, DBP) using demographic and clinical variables and polygenic risk scores (PRSs). We developed a two-model ensemble, consisting of a baseline model, where prediction is based on demographic and clinical variables only, and a genetic model, where we also include PRSs. We evaluate the use of a linear versus a non-linear model at both the baseline and the genetic model levels and assess the improvement in performance when incorporating multiple PRSs. We report the ensemble models performance as percentage variance explained (PVE) on a held-out test dataset. A non-linear baseline model improved the PVEs from 28.1 to 30.1% (SBP) and 14.3% to 17.4% (DBP) compared with a linear baseline model. Including seven PRSs in the genetic model computed based on the largest available GWAS of SBP/DBP improved the genetic model PVE from 4.8 to 5.1% (SBP) and 4.7 to 5% (DBP) compared to using a single PRS. Adding additional 14 PRSs computed based on two independent GWASs further increased the genetic model PVE to 6.3% (SBP) and 5.7% (DBP). PVE differed across self-reported race/ethnicity groups, with primarily all non-White groups benefitting from the inclusion of additional PRSs. In summary, non-linear ML models improves BP prediction in models incorporating diverse populations.