Personalized Risk Schemes and Machine Learning to Empower Genomic Prognostication Models in Myelodysplastic Syndromes

Abstract

Myelodysplastic syndromes (MDS) are characterized by variable clinical manifestations and outcomes. Several prognostic systems relying on clinical factors and cytogenetic abnormalities have been developed to help stratify MDS patients into different risk categories of distinct prognoses and therapeutic implications. The current abundance of molecular information poses the challenges of precisely defining patients’ molecular profiles and their incorporation in clinically established diagnostic and prognostic schemes. Perhaps the prognostic power of the current systems can be boosted by incorporating molecular features. Machine learning (ML) algorithms can be helpful in developing more precise prognostication models that integrate complex genomic interactions at a higher dimensional level. These techniques can potentially generate automated diagnostic and prognostic models and assist in advancing personalized therapies. This review highlights the current prognostication models used in MDS while shedding light on the latest achievements in ML-based research.

Machine learning prediction of the failure of high-flow nasal oxygen therapy in patients with acute respiratory failure

“…This traditional approach requires significant manual intervention and expert knowledge to exclude unnecessary features. ML algorithms can be helpful in developing more precise prognostication models that integrate complex interactions at a higher dimensional level 25 . Physicians now have access to a variety of resources to learn about ML fundamentals and techniques 26 , 27 .…”

Molecular Patterns Identify Distinct Subclasses of Myeloid Neoplasia

“…While MDS classi cation schemes evolved as useful clinical diagnostic or prognostic tools, diagnostic criteria according to genomic signatures re ective of molecular pathogenesis have not been established 1,11,27 . Furthermore, previous attempts to incorporate mutations into prognostic schemes to increase their predictive precision resulted in considering only a handful of consequential mutations 17 . One of the reasons for the notorious inability to establish reproducible genotype/phenotype associations might be the application of primarily supervised strategies using traditional statistics and clinical classi cations (reliant on subjective nosology and time-dependent parameters) as a gold standard.…”

“…Therefore, new strategies may be needed to deconvolute this molecular diversity and generate subdivisions of patients with MDS whose disease ts within molecular pattern similarities, better re ecting prognosis and which could then be targeted with specialized therapeutic approaches. Machine learning (ML) analytic methods, as demonstrated in acute myeloid leukemia (AML) 17 , provide new opportunities to integrate the molecular pathogenesis by identifying relevant patterns, which could serve as molecular sub-entities 11,16,18,19 .…”

How I Manage Transplant Ineligible Patients with Myelodysplastic Neoplasms

“…Artificial intelligence (AI) and machine learning algorithms have also demonstrated converging pathobiological routes by unveiling latent commonalities, but have also highlighted patterns unique to specific MDS clusters with prognostic implications [17,29,30]. Besides obvious potentialities, current pitfalls of such approaches lie in the statistical power of the sample size needed in case of rare mutational events, the consideration of specific mutational characteristics (e.g., variant allelic frequency-VAF, type of mutations), and the inherent 'black-box' nature of the AI methods [31].…”