The Artificially Intelligent Pharma & Healthcare Sector M. Morris Hosseini, MSc, PhD Senior Partner CC Pharma & Healthcare Roland Berger Grand Hyatt Athens, September 24 th 2018
What are the therapies of the future in the digital health era for Pharma and Healthcare and why is Artificial Intelligence so crucial? How does Artificial Intelligence work and where can it help in leveraging and expanding our existing knowledge pool in Pharma and Healthcare? How will Artificial Intelligence affect the stakeholder landscape in Pharma and Healthcare and which opportunities and threats arise? 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 2
Modern medicine can reach an ever larger share of the population, however ever smaller populations Population shift along advancing medicine in Pharma and Healthcare "One size fits all" "Individualized" Personalization focus precision "P4" medicine blockbuster medicine Untreatable Digital Health as accelerator Precision "P4" Data - Predictive Co-diagnostics Test - Preventive as accelerator - Participatory Test Pill - Personalized Pill Stratified Pill Blockbuster Past Present Future Source: L. Hood, Roland Berger 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 3
An enormous amount of health-related data becomes available but needs to be interpreted for modern medicine Digital Health data sources and according application opportunities Data generation technologies Digital Health Enabled/ (Digital Health Data Sources) Enhanced applications > Allogeneic Stem Cells Cyto- Prediction of diseases > iPS 1) mics > CRISPR 2)- Cas9 Histo- Identification of disease > Advanced imaging related agents and patterns mics > In-situ hybridization > Intracellular transport Microbio- Monitoring of health state / visualization Maintenance of wellbeing mics > Microbiome-genomics Novel therapies and Metabolo- > IVD 3) /wearables transport mechanisms mics > Micro-array sensors Data Proteo- Identification of cell > uHTS 4) mics differentiation pathways > Mass spectroscopy interpretation > Genome sequencing Geno- Regenerative therapies / > Epigenomic profiling mics Gene therapies > Transcription mapping Source: Roland Berger 1) induced Pluripotent Stem Cells 2) Clustered Regularly Interspaced Short Palindromic Repeats 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 3) In Vitro Diagnostics 4) ultra High Throughput Screening 4
AI is the core technology to help manage complexity of systems biology to create actionable solutions Complexity of Digital Health in systems biology Complexity and variability of humans Multi-Omics Major challenges due to digital health complexity ✓ > To find relevant signals within this enormous amount of Organisms individual data and enhance Populations the signal-to-noise ratio Cytome … Organs > To analyze and interpret Microbiome 50 organs the data signals and enable actionable health related Metabolome Cells 100 trillion cells decisions Proteome Artificial Intelligence Transciptome Molecules 20,000 proteins as core technology to Epigenome alleviate complexity 35,000 orfs Genome Genes challenge 6 bn nucleotides Source: Roland Berger 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 5
There is a big 'buzz' around AI in healthcare , which attracts approx. 18% of global AI investment Share of global investment in AI by major industry Financial Services Artificial Intelligence represents one of technology's most 20% important priorities and healthcare is perhaps AI's most urgent application. — Peter Lee, Others 45% Director of Research Health 18% I believe we will reach a point Care around 2029 when medical technologies will add one additional year every year to your life expectancy 17% — Ray Kurzweil, Chief Futurist Retail Source: IDC, Roland Berger 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 6
A myriad of AI startups have emerged in the Pharma and Healthcare space along a great variety of use cases Landscape of AI startups in Pharma and Healthcare Source: IDC, CBInsights, Roland Berger 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 7
What are the therapies of the future in the digital health era for Pharma and Healthcare and why is Artificial Intelligence so crucial? How does Artificial Intelligence work and where can it help in leveraging and expanding our existing knowledge pool in Pharma and Healthcare? How will Artificial Intelligence affect the stakeholder landscape in Pharma and Healthcare and which opportunities and threats arise? 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 8
AI does not need a defined algorithm – It "creates" one based on enormous amounts of data Comparison between classical programming and AI > Fixed "if this, than that" algorithms are developed during program design Classic pro- Input data > Algorithm is designed for Output gramming a specific pattern in input data/ Pre-defined "Smart data answers algorithm > All heuristics need to be heuristics" specifically considered during design > Machine learning "generates" the algorithm Input data Output data based on large input data Machine sets – the more data, the learning better the algorithm Self – Feedback > The algorithm adapts with learning "AI" output data feedback from output data algorithm ("the network is trained") Source: Roland Berger 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 9
Evidence-based medicine ensures that clinical decisions are made based on what is known Evidence-based decision making in clinical practice (1/2) Quadrants of medical knowledge known known Evidence-Based Medicine knowns unknowns > Evidence based medicine (EBM) is the conscientious, explicit, judicious and reasonable use of modern, best evidence in making decisions about the care of individual patients > EBM integrates clinical experience and patient values with the best available research information > EBM aims to increase the use of high unknown unknown quality clinical research in clinical decision making knowns unknowns Source: Acta Inform Med, NIH, D. Rumsfeld , Roland Berger 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 10
AI can help us both for the unknown knowns as well as for the known unknowns with its adaptive algorithms Artificial Intelligence leverage points along the quadrants of medical knowledge Informed Trial & Error/ Treatments Research Doctor Knowledge and mechanistic understanding Targeted expansion of knowledge known knowns known unknowns Routinely access- > Routine anamnesis > Hypothesis-driven non-clinical ible and leveraged > Readily accessible knowlegde and clinical research and GP knowlegde > Experience pool of GP doctor development Advanced > Treatment guidelines > AI-powered high-throughput Specialist medical > New publications screening and systems biology expertise > Rare specific/orphan cases > Full leverage of current "Smart Heuristics" knowledge base Artificial > Leverage of already existing > Serendipity-driven unexpected Intelligence Unintended surprise but hitherto untapped experiences experience base discoveries Doctor Intuition unknown knowns unknown knowns unknown unknowns Artificial Intelligence enabled leverage points along quadrants of medical knowledge Source: Roland Berger 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 11
Stanford's researchers developed an AI that can detect skin cancer after machine learning with 130,000 images Improved skin cancer detection employing AI AI-Example: KNOWN UNKNOWNS ✓ Problem Functionality: Visual Processing > Every year about 5.4 million new > AI powered pattern recognition employing deep learning and pre- skin cancer cases in the US; rate of diagnosed image database survival decreases from 97% to 14% if detected in a later stage Approach > The technology is fueled by deep learning programs and a 130,000 image database of high-quality and pre-diagnosed medical imagery > The AI is build up on Google's already present AI that was trained to identify 1.28 million images from 1,000 object categories Advantage > Technology achieved the accuracy of board-certified dermatologists > Future goal is an app that can be used as a scanner on human skin lesions to detect skin cancer Source: Stanford News: Roland Berger Source: Stanford News, Roland Berger 18_09_24 Economist Presentation AI in Pharma Hosseini fv.pptx 12
Recommend
More recommend
