Abstract
Myelodysplastic syndromes (MDS) constitute a group of heterogeneous hematopoietic disorders characterized by cytopenias, dysplasia of hematopoietic cells and a propensity to transform into acute myeloid leukemia (AML). At the time of diagnosis, nearly all MDS patients have anemia, of whom approximately half require therapeutic interventions. In 20-40% of patients, low levels of neutrophils and thrombocytes are present as well. Due to the heterogeneity of MDS, the diagnosis and treatment of MDS poses considerable challenges. In this thesis, we aimed to improve diagnosis, classification, and treatment using computational and experimental methods. In the first part of this thesis, we focused on improving the diagnosis of MDS. We developed and validated a computational tool that included pre-processing of flow cytometry data, FlowSOM for cell population detection, and a machine-learning classifier. This tool demonstrated an improved performance in comparison to existing diagnostic flow cytometry approaches in terms of accuracy and efficiency. Further, we examined inter-analyst agreement for myeloid progenitor assessment by flow cytometry, and illustrated the efficacy and concordance of both manual and computational approaches for analysis. In the second part, we explored classification and genotype-phenotype associations in MDS, focussing on the association of SF3B1 mutations with bone marrow immunophenotype and natural killer (NK) cells. We identified specific erythroid, myelomonocytic, and progenitor features associated with SF3B1 mutations and in case of co-occurrence with a deletion of chromosomal arm 5q (del(5q)). In the study focussing on SF3B1 mutations and NK cells, we identified associations with certain immunophenotypic profiles and impaired NK cell functionality. Additionally, we extensively characterized stem- and progenitor cells of MDS patients of different risk categories, and identified specific cell surface proteins associated with these risk categories. In the third part, we focused on the treatment of MDS, investigating new treatment regimens, identifying potential therapeutic biomarkers and exploring reasons for the paucity of available agents. We conducted a randomized phase-II study that assessed the efficacy and safety of lenalidomide with or without erythropoietin stimulating agents in lower risk MDS patients, and identified several biomarkers for response. Additionally, the monitoring of therapy response through flow cytometric assessment of MDS-associated dysplasia provided insights into treatment effectiveness in the presence or absence of a del(5q). Lastly, we conducted an analysis of MDS trials over a 20-year period to identify reasons for the paucity of available effective agents for the treatment of MDS. In summary, this thesis provides strategies for improving the clinical management of MDS and offers insights into the pathogenesis. The obtained results may contribute to the development of more effective diagnostic tools, to the refinement of classification strategies, and to the expansion of treatment options for MDS patients.
Original language | English |
---|---|
Qualification | Doctor of Philosophy |
Awarding Institution | |
Supervisors/Advisors |
|
Award date | 7 Nov 2023 |
Print ISBNs | 9789493278608 |
DOIs | |
Publication status | Published - 7 Nov 2023 |
Keywords
- Classification
- Diagnosis
- Flow cytometry
- Hematology
- Machine learning
- Myelodysplastic syndromes
- Treatment