m o c . e b o d a . k c o t s – i m e c n a C i n n a v o G © i : n o i t a r t s u l l I Märkte & Technologien It’s all about dose Precision Medicine and Personalized Chemotherapy People are different. But no matter how different, medicine considers patients in the same way: the same aspirin, the same ibuprofen, the same Pepto Bismol – one drug to heal them all. Precision medicine and personalized chemotherapy takes an individual look at the human being. By Anton Dolgikh A precise Approach Therapeutic dose monitoring (TDM)2, an example of the first type of approaches, is a method to adjust dosing according to patient-specific pharmacokinetic parame- ters. Measuring drug concentration in plasma makes it possible to obtain an indi- vidual dose-exposure relationship and to create a medication schedule so as to keep the optimal therapeutic drug concentra- tion, which results in increased therapy efficacy. From the population point of view, TDM opens opportunities for better thera- pies and improves the benefit-risk profile3. According to ClinicalTrials.gov, 56 clinical trials are recruiting or searching for appro- priate candidates or are active, dedicated to the optimization of chemotherapy by means of TDM. Another prominent approach involving empirical data is metronomics.4 Existing for almost twenty years, it has not become a routine procedure in clinical practice. ABOUT THE AUTHOR Anton Dolgikh leads KI and ML projects within the Healthcare and Life Sciences department of the global technology consultancy DataArt and gives seminars on the integration of machine learning into business solutions. Before joining DataArt, Dolgikh conducted research in the Department of Complex Systems at the Free University of Brussels. T he situation is a little bit oversimpli- fied and exacerbated but is right on average and oncologists know this better than others. Oncologists work with highly toxic medications, and while in most cases, erroneously prescribing aspirin will not lead to any serious conse- quences, non-appropriate anticancer treat- ment may in extreme cases cause severe harm to a patient’s health. In Search of Patterns A common and widely adopted approach in chemotherapy is to administer the maxi- mally tolerated dose – following the prin- ciple “the more, the better”. The problem here is that anticancer medications target not only the cancer cells but the healthy 1) Reddy, M., Yang, R. S., Andersen, M. E., & Clewell III, H. J. (2005). Physiologically based pharmacokinetic modeling: science and applications. John Wiley & Sons. 2) Barbolosi, D., Ciccolini, J., Lacarelle, B., Barlési, F., & André, N. (2016). Computational oncology — mathema- tical modelling of drug regimens for precision medicine. Nature Reviews Clinical Oncology, 13(4), 242–254; McKenna, M. T., Weis, J. A., Brock, A., Quaranta, V., & Yankeelov, T. E. (2018). Precision Medicine with Imprecise Therapy: Computational Modeling for Chemotherapy in Breast Cancer. Translational Oncology, 11(3), 732–742. 3) Joerger, M., von Pawel, J., Kraff, S., Fischer, J. R., Eberhardt, W., Gauler, T. C., Jaehde, U. (2016). Open-label, randomized study of individualized, pharmacokinetically (PK)- guided dosing of paclitaxel combined with carboplatin or cisplatin in patients with advanced non-small-cell lung cancer (NSCLC). Annals of Oncology, 27(10), 1895–1902. 4) André, N., Carré, M., & Pasquier, E. (2014). Metronomics: Towards personalized chemotherapy? Nature Reviews Clinical Oncology, 11(7), 413–431. the population cells as well, damaging them. Ideally, anti- cancer drugs are highly specific and hit only the cancer cells. However, response to treatment may differ between patients even for thoroughly tested drugs due to the personal variabilities in pharmaco- kinetic and pharmacodynamical characte- ristics. Precision medicine catches all these individual characteristics and finds patterns within that enables prescrib ing the optimal treatment for an individual patient. When it comes to finding such patterns, mathematics are key. Mathematical modelling is what moves personalized chemotherapy forward. Chemotherapy is complicated, and no unique mathematical model can resolve all its issues at once. There are two high-level approaches to mathematical modelling in personalized chemotherapy: 1. The empirical data approach – machine learning (ML) or mathematical statis- tics in general; 2. Building a model from the first prin- ciples of the patient’s specific biology leading to specific pharmacokinetics and pharma codynamics. An example of is Physiologically Based Pharmaco kinetics/Pharmacodynamics modelling (PBPM).1 this 36 ls 01-2019 „Smarte Medizin“