1
|
Fröhlich H, Balling R, Beerenwinkel N, Kohlbacher O, Kumar S, Lengauer T, Maathuis MH, Moreau Y, Murphy SA, Przytycka TM, Rebhan M, Röst H, Schuppert A, Schwab M, Spang R, Stekhoven D, Sun J, Weber A, Ziemek D, Zupan B. From hype to reality: data science enabling personalized medicine. BMC Med 2018; 16:150. [PMID: 30145981 PMCID: PMC6109989 DOI: 10.1186/s12916-018-1122-7] [Citation(s) in RCA: 187] [Impact Index Per Article: 31.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 03/28/2018] [Accepted: 07/09/2018] [Indexed: 02/08/2023] Open
Abstract
BACKGROUND Personalized, precision, P4, or stratified medicine is understood as a medical approach in which patients are stratified based on their disease subtype, risk, prognosis, or treatment response using specialized diagnostic tests. The key idea is to base medical decisions on individual patient characteristics, including molecular and behavioral biomarkers, rather than on population averages. Personalized medicine is deeply connected to and dependent on data science, specifically machine learning (often named Artificial Intelligence in the mainstream media). While during recent years there has been a lot of enthusiasm about the potential of 'big data' and machine learning-based solutions, there exist only few examples that impact current clinical practice. The lack of impact on clinical practice can largely be attributed to insufficient performance of predictive models, difficulties to interpret complex model predictions, and lack of validation via prospective clinical trials that demonstrate a clear benefit compared to the standard of care. In this paper, we review the potential of state-of-the-art data science approaches for personalized medicine, discuss open challenges, and highlight directions that may help to overcome them in the future. CONCLUSIONS There is a need for an interdisciplinary effort, including data scientists, physicians, patient advocates, regulatory agencies, and health insurance organizations. Partially unrealistic expectations and concerns about data science-based solutions need to be better managed. In parallel, computational methods must advance more to provide direct benefit to clinical practice.
Collapse
Affiliation(s)
- Holger Fröhlich
- UCB Biosciences GmbH, Alfred-Nobel-Str. Str. 10, 40789 Monheim, Germany
- University of Bonn, Bonn-Aachen International Center for IT, Endenicher Allee 19c, 53115 Bonn, Germany
| | - Rudi Balling
- University of Luxembourg, 6 avenue du Swing, 4367 Belvaux, Luxembourg
| | - Niko Beerenwinkel
- Department of Biosciences and Engineering, ETH Zurich, Mattenstr. 26, 4058 Basel, Switzerland
| | - Oliver Kohlbacher
- University of Tübingen, WSI/ZBIT, Sand 14, 72076 Tübingen, Germany
- Max Planck Institute for Developmental Biology, Max-Planck-Ring 5, 72076 Tübingen, Germany
- Quantitative Biology Center, University of Tübingen, Auf der Morgenstelle 8, 72076 Tübingen, Germany
- Institute for Translational Bioinformatics, University Medical Center Tübingen, Sand 14, 72076 Tübingen, Germany
| | - Santosh Kumar
- Department of Computer Science, University of Memphis, 2222 Dunn Hall, Memphis, TN 38152 USA
| | - Thomas Lengauer
- Max-Planck-Institute for Informatics, 66123 Saarbrücken, Germany
| | - Marloes H. Maathuis
- ETH Zurich, Seminar für Statistik, Rämistrasse 101, 8092 Zurich, Switzerland
| | - Yves Moreau
- University of Leuven, ESAT, Kasteelpark Arenberg 10, 3001 Leuven, Belgium
| | - Susan A. Murphy
- Harvard University, Science Center 400 Suite, Oxford Street, Cambridge, MA 02138-2901 USA
| | - Teresa M. Przytycka
- National Center of Biotechnology Information, National Institute of Health, 8600 Rockville Pike, Bethesda, MD 20894-6075 USA
| | - Michael Rebhan
- Novartis Institutes for Biomedical Research, 4056 Basel, Switzerland
| | - Hannes Röst
- Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, 160 College Street, Toronto, ON M5S 3E1 Canada
| | - Andreas Schuppert
- RWTH Aachen, Joint Research Center for Computational Biomedicine, Pauwelsstrasse 19, 52074 Aachen, Germany
| | - Matthias Schwab
- Dr. Margarete Fischer-Bosch Institute of Clinical Pharmacology, Aucherbachstrasse 112, 70376 Stuttgart, Germany
- University of Tübingen, Departments of Clinical Pharmacology and of Pharmacy and Biochemistry, Tübingen, Germany
| | - Rainer Spang
- University of Regensburg, Institute of Functional Genomics, Am BioPark 9, 93053 Regensburg, Germany
| | - Daniel Stekhoven
- ETH Zurich, NEXUS Personalized Health Technol., Otto-Stern-Weg 7, 8093 Zurich, Switzerland
| | - Jimeng Sun
- Georgia Tech University, 801 Atlantic Drive, Atlanta, GA 30332-0280 USA
| | - Andreas Weber
- Institute for Computer Science, University of Bonn, Endenicher Allee 19a, 53115 Bonn, Germany
| | - Daniel Ziemek
- Pfizer, Worldwide Research and Development, Linkstraße 10, 10785 Berlin, Germany
| | - Blaz Zupan
- Faculty of Computer and Information Science, University of Ljubljana, Večna pot 113, SI-1000 Ljubljana, Slovenia
| |
Collapse
|
3
|
Abstract
The development of new drug therapies requires substantial and ever increasing investments from the pharmaceutical company. Ten years ago, the average time from early target identification and optimization until initial market authorization of a new drug compound took more than 10 years and involved costs in the order of one billion US dollars. Recent studies indicate even a significant growth of costs in the meanwhile, mainly driven by the increasing complexity of diseases addressed by pharmaceutical research.Modeling and simulation are proven approaches to handle highly complex systems; hence, systems medicine is expected to control the spiral of complexity of diseases and increasing costs. Today, the main focus of systems medicine applications in industry is on mechanistic modeling. Biological mechanisms are represented by explicit equations enabling insight into the cooperation of all relevant mechanisms. Mechanistic modeling is widely accepted in pharmacokinetics, but prediction from cell behavior to patients is rarely possible due to lacks in our understanding of the controlling mechanisms. Data-driven modeling aims to compensate these lacks by the use of advanced statistical and machine learning methods. Future progress in pharmaceutical research and development will require integrated hybrid modeling technologies allowing realization of the benefits of both mechanistic and data-driven modeling. In this chapter, we sketch typical industrial application areas for both modeling techniques and derive the requirements for future technology development.
Collapse
Affiliation(s)
- Lars Kuepfer
- Computational Systems Biology, Bayer Technology Services GmbH, Leverkusen, Germany
- Institute of Applied Microbiology, RWTH Aachen University, Aachen, Germany
| | - Andreas Schuppert
- Lehrstuhl für datenbasierte Modellierung in CES, Joint Research Center for Computational Biomedicine, AICES, RWTH Aachen University, Augustinerbach 2, Aachen, 52062, Germany.
| |
Collapse
|
4
|
Balabanov S, Braig M, Brümmendorf TH. Current aspects in resistance against tyrosine kinase inhibitors in chronic myelogenous leukemia. DRUG DISCOVERY TODAY. TECHNOLOGIES 2015; 11:89-99. [PMID: 24847658 DOI: 10.1016/j.ddtec.2014.03.003] [Citation(s) in RCA: 57] [Impact Index Per Article: 6.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 01/28/2023]
Abstract
Resistance against tyrosine kinase inhibitors (TKIs) represents a relevant clinical problem in treatment of chronic myelogenous leukemia (CML). On the basis of their activity against the spectrum of BCR-ABL mutations that have shown to be the most prominent mechanism of resistance to imatinib, new TKIs have been classified as second generation (such as nilotinib, dasatinib and bosutinib) or third generation (also cover- ing T315I such as ponatinib) TKIs. However, mutations in BCR-ABL only account for about half of the cases of treatment failure under TKI and other mechanisms either rendering the leukemic cells still dependent of BCR-ABL activity or supporting oncogenic properties of the leukemic cells independent of BCR-ABL signaling have been identified. A detailed understanding of the different underlying resistance mechanisms will be the prerequisite to eventually overcome clinical resistance and for the successful use of tailored combinations of targeted inhibitors in the future.
Collapse
|
5
|
Arvaniti K, Papadioti A, Kinigopoulou M, Theodorou V, Skobridis K, Tsiotis G. Proteome Changes Induced by Imatinib and Novel Imatinib Derivatives in K562 Human Chronic Myeloid Leukemia Cells. Proteomes 2014; 2:363-381. [PMID: 28250386 PMCID: PMC5302748 DOI: 10.3390/proteomes2030363] [Citation(s) in RCA: 4] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/24/2014] [Revised: 07/03/2014] [Accepted: 07/08/2014] [Indexed: 12/13/2022] Open
Abstract
Imatinib mesylate is the leading compound to treat chronic myeloid leukemia (CML) and other cancers, through its inhibition of Bcr-Abl tyrosine kinases. However, resistance to imatinib develops frequently, particularly in late-stage disease and has necessitated the development of new Bcr-Abl inhibitors. The synthesis of a new series of phenylaminopyrimidines, structurally related to imatinib, showed large interest since the introduction of nilotinib. Here, we compare the protein levels in K562 cells treated with either imatinib or with novel imatinib derivates. Our results revealed that among the 986 quantified proteins, 35 had significantly altered levels of expression by imatinib or its derivates. In a second series of experiments, we directly compared the proteomes of imatinib treated K562 cells with those K562 cells treated with any of the four imatinib derivates. More than 1029 protein were quantified, 80 of which had altered levels of expression. Both experiments pointed to changes in the expression of the ATP-dependent RNA helicase DDX3X and of two mitochondrial coiled-coil-helix-coiled-coil-helix domain-containing proteins.
Collapse
Affiliation(s)
- Katerina Arvaniti
- Division of Biochemistry, Department of Chemistry, University of Crete, P.O. Box 2208, GR-71003 Voutes, Greece.
| | - Anastasia Papadioti
- Division of Biochemistry, Department of Chemistry, University of Crete, P.O. Box 2208, GR-71003 Voutes, Greece.
| | - Maria Kinigopoulou
- Department of Chemistry, Section of Organic Chemistry and Biochemistry, University of Ioannina, 45110 Ioannina, Greece.
| | - Vassiliki Theodorou
- Department of Chemistry, Section of Organic Chemistry and Biochemistry, University of Ioannina, 45110 Ioannina, Greece.
| | - Konstantinos Skobridis
- Department of Chemistry, Section of Organic Chemistry and Biochemistry, University of Ioannina, 45110 Ioannina, Greece.
| | - Georgios Tsiotis
- Division of Biochemistry, Department of Chemistry, University of Crete, P.O. Box 2208, GR-71003 Voutes, Greece.
| |
Collapse
|
6
|
Stehle F, Schulz K, Seliger B. Towards defining biomarkers indicating resistances to targeted therapies. BIOCHIMICA ET BIOPHYSICA ACTA-PROTEINS AND PROTEOMICS 2013; 1844:909-16. [PMID: 24269379 DOI: 10.1016/j.bbapap.2013.11.006] [Citation(s) in RCA: 9] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 04/23/2013] [Revised: 10/17/2013] [Accepted: 11/13/2013] [Indexed: 12/20/2022]
Abstract
An impressive, but often short objective response was obtained in many tumor patients treated with different targeted therapies, but most of the patients develop resistances against these drugs. So far, a number of distinct mechanisms leading to intrinsic as well as acquired resistances have been identified in tumors of distinct origin. These can arise from genetic alterations, like mutations, truncations, and amplifications or due to deregulated expression of various proteins and signal transduction pathways, but also from cellular heterogeneity within tumors after an initial response. Therefore, biomarkers are urgently needed for cancer prognosis and personalized cancer medicine. The application of "ome"-based technologies including cancer (epi)genomics, next generation sequencing, cDNA microarrays and proteomics might led to the predictive or prognostic stratification of patients to categorize resistance mechanisms and to postulate combinations of treatment strategies. This review discusses the implementation of proteome-based analysis to identify markers of pathway (in)activation in tumors and the resistance mechanisms, which represent major clinical problems as a tool to optimize individually tailored therapies based on targeted drugs. This article is part of a Special Issue entitled: Biomarkers: A Proteomic Challenge.
Collapse
Affiliation(s)
- Franziska Stehle
- Martin Luther University Halle-Wittenberg, Institute of Medical Immunology, Magdeburger Str. 2, D-06112 Halle, Saale, Germany
| | - Kristin Schulz
- Martin Luther University Halle-Wittenberg, Institute of Medical Immunology, Magdeburger Str. 2, D-06112 Halle, Saale, Germany
| | - Barbara Seliger
- Martin Luther University Halle-Wittenberg, Institute of Medical Immunology, Magdeburger Str. 2, D-06112 Halle, Saale, Germany.
| |
Collapse
|