Katharina Martha Götz (1), Katharina Och (1), Amin Turki (2), Saskia Leserer (2) and Thorsten Lehr (1)
(1) Clinical Pharmacy, Saarland University, Saarbrücken, Germany, (2) Department of Bone Marrow Transplantation, West-German Cancer Center, University Hospital Essen, Germany
Introduction: Allogeneic Hematopoietic Cell Transplantation (HCT) can cure or improve the outcome in a variety of hematological diseases. Currently, predictive scores of severe complications such as viral infections, graft-versus-host disease (GVHD) and relapse after HCT are insufficient. Consequently, early lifesaving interventions cannot always be applied on time. The collaborative project XplOit [1] started in 2016 with the aim to improve the outcome of adult patients after HCT by the application of systems medicine modelling, in order to reveal underlying processes of HCT complications. Furthermore, patients’ individual course of disease should be predicted.
Objectives: The objective was to describe, characterize and predict myelosuppression after conditioning chemotherapy and recovery of blood cells after HCT via mathematical models of leukopoiesis and thrombopoiesis.
Methods: Retrospectively collected daily measured cell counts of patients with HCT due to different diagnoses at University Hospital Essen, Germany, were analyzed. Model development and validation were performed independently for each of the two submodels, both consisting of a structural and stochastic model. Each structural model was developed stepwise by examining turnover models and further literature models. Given the developed models of leukopoiesis and thrombopoiesis and their estimated parameters, we informed the models based on a smaller dataset until day three after HCT and predicted individual profiles from day three to six weeks after HCT (Bayesian approach). Parameter estimation and simulations were performed using non-linear-mixed-effects methods implemented in the software NONMEM (version 7.4.3) and the graphical interface Pirana (version 2.9.9). Statistical evaluation and graphics were created within the software R (version 3.4.3) and its graphical interface RStudio (version 1.1.423).
Results: 58,731 individual leukocyte and 58,776 thrombocyte measurements of 1245 HCT patients were used for leukopoiesis and thrombopoiesis model development. The final structural models were adapted from Friberg et al. [2], originally developed to describe suppression and recovery of neutrophils in patients after administration of myelosuppressive chemotherapeutics. Due to missing information on the exact conditioning regimen, we assumed the drug effects of the conditioning regimen prior to HCT. The following population parameters were estimated: a leukocyte baseline of 5.67 cells/nL and a thrombocyte baseline of 89.50 cells/nL. To cover the feedback mechanisms of circulating on proliferating cells, the estimated parameter γ is 1.9 higher in the thrombopoiesis model compared to the leukopoiesis model (0.11 and 0.06, respectively), whereas the drug effect on proliferating stem cells is comparable in both models (0.72 and 0.66, respectively). Thrombocytes showed an increased mean transit time of 6.00 days compared to leukocytes (4.42 days). In the validation cohort (457 patients) the observed individual, hematologic recovery data matched very well with the model’s predictions.
Conclusions: We developed models of hematopoietic recovery among adult patients with HCT that show a good descriptive performance. The estimated leukocyte baseline presents a normal value. Similarly, the estimated thrombocyte baseline presents with mild thrombocytopenia according to the common toxicity criteria for adverse events (CTCAE) [3]. In conclusion, with a priori information the models adequately predict nadir and recovery of leukocytes and thrombocytes in a smaller dataset from day three to six weeks after HCT.
References:
[1] XplOit – Semantic Support for Predictive Modelling in Systems Medicine (2016): http://www.xploit-idsem.de/englische-zusammenfassung [accessed 28 February 2019].
[2] Friberg LE et al. Model of chemotherapy-induced myelosuppression with parameter consistency across drugs. J Clin Oncol (2002) 20: 4713–4721.
[3] National Cancer Institute. Common Terminology Criteria for Adverse Events Version 5.0 (CTCAE). U.S. Department of Health and Human Services (2017) v5.0: 1-147.
Reference: PAGE 28 (2019) Abstr 9000 [www.page-meeting.org/?abstract=9000]
Poster: Drug/Disease Modelling - Oncology