Describing the evolution of myeloid-derived leucocytes in treated B-lineage paediatric acute lymphoblastic leukaemia with a data-driven granulocyte-monocyte-blast model.

Acute lymphoblastic leukaemia (ALL) is associated with a compromised myeloid system. Understanding the state of granulopoiesis in a patient during treatment, places the clinician in an advantageous position. Mathematical models are aids able to present the clinician with insight into the behaviour of myeloid-derived leucocytes. The main objective of this investigation was to determine whether a proposed model of ALL during induction therapy would be a usable descriptor of the system. The model assumes the co-occurrence of the independent leukaemic and normal marrow populations. It is comprised of four delay-differential equations, capturing the fundamental characteristics of the blood and bone marrow myeloid leucocytes and B-lineage lymphoblasts. The effect of treatment was presumed to amplify cell loss within both populations. Clinical data was used to inform the construction, calibration and examination of the model. The model is promising-presenting a good foundation for the development of a clinical supportive tool. The predicted parameters and forecasts aligned with clinical expectations. The starting assumptions were also found to be sound. A comparative investigation highlighted the differing responses of similarly diagnosed patients during treatment and further testing on patient data emphasized patient specificity. Model examination recommended the explicit consideration of the suppressive effects of treatment on the normal population production. Additionally, patient-related factors that could have resulted in such different responses between patients need to be considered. The parameter estimates require refinement to incorporate the action of treatment. Furthermore, the myeloid populations require separate consideration. Despite the model providing explanation, incorporating these recommendations would enhance both model usability and predictive capacity.