RESUMO
ETV6::RUNX1 is the most common fusion gene in childhood acute lymphoblastic leukaemia (ALL) and is associated with favorable outcomes, especially in low-risk children. However, as many as 10% of children relapse within 3 years, and such early relapses have poor survival. Identifying children at risk for early relapse is an important challenge. We interrogated data from 87 children with low-risk ETV6::RUNX1-positive B-cell ALL and with available preserved bone marrow samples (discovery cohort). We profiled somatic point mutations in a panel of 559 genes and genome-wide transcriptome and single-nucleotide variants. We found high TIMD4 expression (> 85th-percentile value) at diagnosis was the most important independent prognostic factor of early relapse (hazard ratio [HR] = 5.07 [1.76, 14.62]; p = 0.03). In an independent validation cohort of low-risk ETV6::RUNX1-positive B-cell ALL (N = 68) high TIMD4 expression at diagnosis had an HR = 4.78 [1.07, 21.36] (p = 0.04) for early relapse. In another validation cohort including 78 children with low-risk ETV6::RUNX1-negative B-cell ALL, high TIMD4 expression at diagnosis had an HR = 3.93 [1.31, 11.79] (p = 0.01). Our results suggest high TIMD4 expression at diagnosis in low-risk B-cell ALL in children might be associated with high risk for early relapse.
RESUMO
Forecasting of severe acute graft-versus-host disease (aGVHD) after transplantation is a challenging 'large p, small n' problem that suffers from nonuniform data sampling. We propose a dynamic probabilistic algorithm, daGOAT, that accommodates sampling heterogeneity, integrates multidimensional clinical data and continuously updates the daily risk score for severe aGVHD onset within a two-week moving window. In the studied cohorts, the cross-validated area under the receiver operator characteristic curve (AUROC) of daGOAT rose steadily after transplantation and peaked at ≥0.78 in both the adult and pediatric cohorts, outperforming the two-biomarker MAGIC score, three-biomarker Ann Arbor score, peri-transplantation features-based models and XGBoost. Simulation experiments indicated that the daGOAT algorithm is well suited for short time-series scenarios where the underlying process for event generation is smooth, multidimensional and where there are frequent and irregular data missing. daGOAT's broader utility was demonstrated by performance testing on a remotely different task, that is, prediction of imminent human postural change based on smartphone inertial sensor time-series data.