Springe direkt zu Inhalt

Pharmacist Marian Klose

Bild: Marian Klose

Institute of Pharmacy

Clinical Pharmacy & Biochemistry

Freie Universität Berlin

Kelchstr. 31
Room 138 A
12169 Berlin

Curriculum Vitae

Since 03/2023

PhD student (PharMetrX: Pharmacometrics & Computational Disease Modelling) at the Dept. of Clinical Pharmacy and Biochemistry, supervised by Prof. Dr. C. Kloft, Freie Universität Berlin

10/2022 – 12/2022

Internship, Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim am Rhein


License to practise as a pharmacist

11/2021 – 05/2022

Pre-registration Pharmacist, Center for Pharmacometrics and Systems Pharmacology, University of Florida, Florida, USA

05/2021 – 10/2021

Pre-registration Pharmacist, MediosApotheke an der Charité, Berlin

09/2018 – 10/2018

Internship, Sanofi-Aventis Deutschland GmbH, Frankfurt am Main

04/2017 – 03/2021

Pharmacy studies, Freie Universität Berlin, Berlin

06/2015 – 03/2017

Chemistry and Biochemistry studies, Freie Universität Berlin and Humboldt-Universität zu Berlin

07/2014 – 05/2015

Work and Travel, Australia and New Zealand


A-levels, St. Ursula Gymnasium, Attendorn

Optimising Pharmacotherapy Through Novel Model-Informed Precision Dosing Strategies With a Special Focus on Anticancer Drugs.

The global impact of cancer-related deaths requires the efficient development of effective and safe anticancer drugs as well as an optimised use of these drugs. However, achieving appropriate plasma concentrations of these drugs can be challenging. While underexposure due to a suboptimal dosing regimen leads to lack of efficacy, overexposure of anticancer drugs increases the risk of potentially life-threatening severe adverse drug reactions. These two opposing risks must be constantly balanced in drug therapy to achieve an optimal benefit-risk ratio for the individual patient. Complicating matters further, many of these drugs show significant pharmacokinetic (PK) and pharmacodynamic (PD) variability between patients. As a result, a one-dose fits all approach cannot be meaningfully applied and a patient-specific therapy needs to be utilised. Which method is best for therapy-individualisation is part of an ongoing discussion. In oncology, dosing for many drugs is based on body surface area, although there are strong indications that other approaches may be more appropriate. One of these alternative approaches is the use of quantitative pharmacometric modelling and simulation frameworks. The use of such models can characterise patient-drug-disease relationships and reveal sources of interindividual variability. In clinical settings, the Model-informed precision dosing (MIPD) strategy has the potential to further improve the therapy outcome while simultaneously reducing the risk of adverse drug events. In this context, my PhD research will focus on optimising pharmacotherapy using novel Model-Informed Precision Dosing Strategies with an emphasis on, but not limited to, anticancer drugs. Various research questions shall be addressed within specific projects:

The focus of project 1 is on methotrexate (MTX), a folic acid analogue widely used for the treatment of various cancers. When patients are treated with MTX, they receive supportive therapy to mitigate side effects. This includes urinary alkalinisation, hydration, and delayed administration of calcium folinate. Because some patients exhibit decreased elimination of MTX, this supportive therapy must be continued for an extended period of time until plasma MTX concentration fall below a certain threshold. In clinical practice, these patients are of particular concern, and it would be beneficial to identify patients at risk for decreased MTX elimination at the early phases of pharmacotherapy. I plan to achieve this goal by developing a pharmacokinetic nonlinear mixed-effects model on real world data. This model will be used to identify patients at risk for decreased MTX clearance for whom closer Therapeutic Drug Monitoring (TDM) and supportive care can be established. Such modelling frameworks can also be applied to provide insights into optimal design of the blood sampling schedule with the goal to increase efficiency of the TDM program. For the model building step, I have obtained clinical data from 256 patients undergoing MTX therapy retrospectively collected at Klinikum Stuttgart. The data includes MTX dosing, MTX plasma concentrations, patient characteristics, and biomarker data. To make the usage of our predictions more user-friendly and more accessible, the NLME pharmacokinetic model will be implemented in an RShiny application that can be used in clinical practice. In addition, I will evaluate the predictive performance of existing NLME PK models based on our dataset. In doing so, I will apply novel approaches such as model averaging to determine whether these new approaches can improve the performance of existing NLME PK models.

Project 2 will focus on paclitaxel, a chemotherapeutic agent widely used against various types of cancer, including lung cancer. Previously, a novel bone-marrow-exhaustion model to predict the time course of neutrophil count over multiple cycles was developed upon this dataset. During my PhD, I want to examine if the predictive performance of this novel model is also higher for predictions of platelet count compared to a commonly used reference model from literature. Furthermore, I would like to explore different approaches on how to derive reliable individual parameter estimates in a multiple imputation framework, since a NLME PK model was previously built based on this approach and this dataset. For a subsequent link to e.g., the platelet count model, reliable individual parameter estimates are needed and the correct derivation of these estimates is being discussed in the scientific community. For this project, a dataset from the CEPAC study (CESAR Study of PAClitaxel Therapeutic Drug Monitoring) will be utilised.

In Project 3 I plan to work on new methodological concepts to optimise predictions within a MIPD framework. Previously published approaches of how ‘classical’ NLME models can be combined with machine learning (ML) showed promising results while preserving the well-established nature of NLME modelling approaches. Such hybrid concepts include e.g., ML assisted model-selection and ML assisted flattening of model priors for estimation of the maximum a-posteriori (MAP) parameters of an individual. I plan to extend and refine these existing concepts and investigate novel approaches on how machine learning can be integrated into existing modelling structures to improve the predictive performance of models such as in projects 1 and 2.

Within these different projects, the topic of my work 'Optimising pharmacotherapy through Novel Model-Informed Precision Dosing Strategies with a special focus on anticancer drugs' is explored from different perspectives, ranging from applied to more methodological aspects.