Deep Learning in Drug Discovery and Diagnostics, 2017 - 2035

Deep Learning in Drug Discovery and Diagnostics, 2017 - 2035

  • February 2017 •
  • 235 pages •
  • Report ID: 4805990 •
  • Format: PDF
Deep learning is a novel machine learning technique that can be used to generate relevant insights from large volumes of data. The term Deep Learning was coined in 2006 by Geoffrey Hinton to refer to algorithms that enable computers to analyze objects and text in videos and images. Fundamentally, deep learning algorithms are designed to analyze and use large volumes of data to improve the capabilities of machines. Companies, such as Google, Amazon, Facebook, LinkedIn, IBM and Netflix, are already using deep learning algorithms to analyze users’ activities and make customized suggestions and recommendations based on individual preferences. Today, in many ways, deep learning algorithms have enabled computers to see, read and write. In light of recent advances, the error rate associated with machines being able to analyze and interpret medical images has come down to 6%, which, some research groups claim, is even better than humans.

The applications of the technology are being explored across a variety of areas. Specifically in healthcare, the American Recovery and Reinvestment Act of 2009 and the Precision Medicine Initiative of 2015 have widely endorsed the value of medical data in healthcare. Owing to several such initiatives, medical big data is expected to grow approximately 50-fold to reach 25,000 petabytes by 2020. Since 80% of this is unstructured, it is difficult to generate valuable / meaningful insights using conventional data mining techniques. In such cases, deep learning has emerged as a novel solution. Lead identification and optimization in drug discovery, support in patient recruitment for clinical trials, medical image analysis, biomarker identification, drug efficacy analysis, drug adherence evaluation, sequencing data analysis, virtual screening, molecule profiling, metabolomic data analysis, EMR analysis and medical device data evaluation are examples of applications where deep learning based solutions are being explored.

The likely benefits associated with the use of deep learning based solutions in the above mentioned areas is estimated to be worth multi billion dollars. There are well-known references where deep learning models have accelerated the drug discovery process and provided solutions to precision medicine. With potential applications in drug repurposing and preclinical research, deep learning in drug discovery is likely to have great opportunity. In diagnostics, an increase in the speed of diagnosis is likely to have a profound impact in regions with large patient to physician ratios. The implementation of such solutions is anticipated to increase the efficiency of physicians providing a certain amount of relief to the overly-burdened global healthcare system.

The “Deep Learning: Drug Discovery and Diagnostics Market, 2017-2035” report examines the current landscape and future outlook of the growing market of deep learning solutions within the healthcare domain. Primarily driven by the big data revolution, deep learning algorithms have emerged as a novel solution to generate relevant insights from medical data. This continuing shift towards digitalization of healthcare system has been backed by a number of initiatives taken by the government, and has also sparked the interest of several industry / non-industry players. The involvement of global technology companies and their increasing collaborations with research institutes and hospitals are indicative of the research intensity in this field. At the same time, the pharma giants have been highly active in adopting the digital models. Companies such as AstraZeneca, Pfizer and Novartis continue to evaluate the digital health initiatives across drug discovery, clinical trial management and medical diagnosis. Some notable examples of such digital health initiatives include GSK and Pfizer’s collaboration with Apple for the use of the latter’s research kit in clinical trials, Biogen’s partnership with Fitbit for using smart wearables in clinical trial management, and Teva Pharmaceuticals’ partnership with American Well to use Smart Inhalers for patients with asthma and COPD.

Backed by funding from several Venture Capital firms and strategic investors, deep learning has emerged as one of the most widely explored initiatives within digital healthcare. The current generation of deep learning models are flexible and have the ability to evolve and become more efficient over time. Despite being a relatively novel field of research, these models have already demonstrated significant potential in the healthcare industry.

One of the key objectives of this study was to identify the various deep learning solutions that are currently available / being developed to cater to unmet medical needs, and also evaluate the future prospects of deep learning within the healthcare industry. These solutions are anticipated to open up significant opportunities in the field of drug discovery and diagnostics as the healthcare industry gradually shifts towards digital solutions. In addition to other elements, the study covers the following:

- The current status of the market with respect to key players, specific applications and the therapeutic areas in which these solutions can be applied.
- The various initiatives that are being undertaken by technology giants, such as IBM, Google, Facebook, Microsoft, NVIDIA and Samsung. The presence of these stakeholders signifies the opportunity and the impact that these solutions are likely to have in the near future. Specifically, we have presented a comparative analysis of the deep learning solutions developed by IBM and Google.
- Detailed profiles of some of the established, as well as emerging players in the industry, highlighting key technology features, primary applications and other relevant information.
- The impact of venture capital funding in this area. It is important to mention that since the industry has witnessed the emergence of several start-ups, funding is a key enabler that is likely to drive both innovation and product development in the coming years.
- An elaborate valuation analysis of companies that are involved in applying deep learning in drug discovery and diagnostics. We built a multi-variable dependent valuation model to estimate the current valuation of a number of companies focused in this domain.
- Future growth opportunities and likely impact of deep learning in the drug discovery and diagnostics domains. The forecast model, backed by robust secondary research and credible inputs from primary research, was primarily based on the likely time-saving and its associated cost-saving opportunity to the healthcare system.

For the purpose of the study, we invited over 100 stakeholders to participate in a survey to solicit their opinions on upcoming opportunities and challenges that must be considered for a more inclusive growth. Our opinions and insights presented in this study were influenced by discussions conducted with several key players in this domain. The report features detailed transcripts of interviews held with Mausumi Acharya (CEO, Advenio Technosys), Carla Leibowitz (Head of Strategy and Marketing, Arterys) and Deekshith Marla (CTO,

Most of the data presented in this report has been gathered via secondary and primary research. We have also conducted interviews with experts in the area (academia, industry, medical practice and other associations) to solicit their opinions on emerging trends in the market. This is primarily useful for us to draw out our own opinion on how the market will evolve across different regions and technology segments. Where possible, the available data has been checked for accuracy from multiple sources of information.

The secondary sources of information include
- Annual reports
- Investor presentations
- SEC filings
- Industry databases
- News releases from company websites
- Government policy documents
- Other analyst's opinion reports

While the focus has been on forecasting the market over the coming two decades, the report also provides our independent view on various non-commercial trends emerging in the industry. This opinion is solely based on our knowledge, research and understanding of the relevant market gathered from various secondary and primary sources of information.

Chapter 2 provides an executive summary of the report. It offers a high level view on where the deep learning market for drug discovery and diagnostics is headed in the long term.

Chapter 3 is an introductory chapter that presents details on the digital revolution in the medical industry. It elaborates on the growth of artificial intelligence and machine learning tools, such as deep learning algorithms, along with a discussion on their potential applications in solving some of the key challenges faced by the healthcare industry. The chapter also gives an overview on the rise of big data and its role in providing personalized and evidence based care to patients.

Chapter 4 includes information on close to 100 companies that are evaluating potential applications of their proprietary deep learning solutions in the healthcare industry. The classification system used for these solutions was based on their application areas. These include drug discovery, diagnostics, clinical trial management and drug adherence programs. In addition, we have highlighted specific geographical pockets that we identified as innovation hubs in this sector.

Chapter 5 provides detailed profiles of some of the key stakeholders in this space with detailed information on their technologies, funding, collaborations and partnerships, intellectual capital, awards and recognition and activity on social media.

Chapter 6 presents a case study on two technology giants in this field, namely IBM and Google. It provides a detailed description of the initiatives being undertaken by these companies to explore the applications of deep learning in the medical field. In addition, the chapter provides a comparison of the two companies based on their respective deep learning expertise, and partnerships and acquisitions.

Chapter 7 provides information on the various investments that have been made into this industry. Our analysis revealed interesting insights on the growing interest of venture capitalists and other stakeholders in this market. In addition, we identified some of the key investors in this market.

Chapter 8 presents detailed projections related to the growth of the deep learning industry in healthcare from 2017 to 2035. To quantify the opportunity for deep learning in the drug discovery space, we have provided optimistic and conservative forecast scenarios, along with our base forecast to account for the uncertainties associated with the adoption of these technologies. The insights presented in this chapter are backed by data from close to 50 countries and highlights the opportunity for deep learning companies in diagnostics within the same regions.

Chapter 9 features a comprehensive valuation analysis of the companies that are developing deep learning solutions for applications in drug discovery and diagnostics. The chapter provides insights based on a multi-variable dependent valuation model. The model is based on the future potential of the companies’ technologies, their current popularity, funding received, year of establishment and the employed workforce in these companies.

Chapter 10 presents the opinions expressed by selected key opinion leaders on the applications and challenges associated with deep learning in the healthcare sector. The chapter provides key takeaways from presentations and videos of these experts, highlighting the future opportunity for these models within the healthcare industry.

Chapter 11 summarizes the overall report. In this chapter, we provide a recap of the key takeaways and our independent opinion based on the research and analysis described in the previous chapters.

Chapter 12 is a collection of interview transcripts of the discussions held with key stakeholders in this market. We have presented the details of our discussions with Mausumi Acharya (CEO, Advenio Technosys), Carla Leibowitz (Head of Strategy and Marketing, Arterys), and Deekshith Marla (CTO,

Chapter 13 is an appendix, which provides tabulated data and numbers for all the figures in the report. In addition, the chapter includes a detailed analysis of the survey conducted with several companies to estimate the opportunity for deep learning in drug discovery and diagnostics.

Chapter 14 is an appendix, which provides the list of companies and organizations mentioned in the report.

1. During our research, we identified close to 100 industry / non-industry players that are exploring their proprietary deep learning based technologies in drug discovery and diagnostics. A majority of these companies (61%) were founded post 2010. In fact, between 2013 and 2016 alone, the industry saw the emergence of 50 startups in this field.
2. More than 55% of the companies working in this space are applying their deep learning models for diagnostic purposes. Of these, 78% of the companies offer solutions for medical imaging analysis. Notable examples include (in alphabetical order) Arterys, AvalonAI, Bay Labs,, Butterfly Network, CAMELOT biomedical systems, Cyrcadia Health, Enlitic, iCarbonX, Lunit and Zebra Medical Vision.
3. On the other hand, close to 35% of the companies engaged in this domain are focused on applying deep learning models in drug discovery. 57% of these companies provide deep learning powered drug discovery platforms. Examples of players in this segment include (in alphabetical order) Atomwise,, BERG Health, Cloud Pharmaceuticals, Cyclica, Hummingbird Bioscience, InSilico Medicine, Mind the Byte, Molplex Pharmaceutical, nference, Numedii, Numerate, Standigm, twoXAR, Verge Genomics, Vium and SparkBeyond.
4. In addition, there are companies that are focused in applying deep learning in both drug discovery as well as diagnostics. Examples of such companies include (in alphabetical order) 23andMe, Appistry, Deep Genomics, Desktop Genetics, Globavir Biosciences, Google, IBM, SolveBio and Wuxi NextCODE.
5. During the last three years, heavy investments have been made in this domain. Of the overall amount invested in last 10 years (USD 1.8 billion), USD 1.6 billion was invested into deep learning initiatives in and after 2014. There are several recent examples. iCarbonX raised USD 214 million in three funding rounds (January 2016, April 2016 and July 2016), Flatiron Health received USD 175 million in Series C funding (January 2016), LAM Therapeutics witnessed funding of USD 40 million (February 2016), and Human Longevity closed a Series B funding round amounting to USD 220 million (April 2016)..
6. In the drug discovery segment, the deep learning solutions have shown to significantly reduce the cost and time spent in bringing a drug to the market. Taking a drug from discovery stage to the market is known to cost up to USD 2.5 billion and takes, on an average, close to 12 years. Deep learning models are likely to save as much as 50% of this cost and save a significant amount of time. By 2035, we have predicted annual cost savings of over USD 100 billion for the global healthcare system.
7. The adoption of deep learning models in diagnostics is also likely to provide several cost and time saving opportunities. According to our estimates, by 2035, deep learning solutions can result in annual savings of over USD 35 billion in the diagnostics segment alone. The activity is likely to be relatively more prominent in the high income countries in the near term. However, in the long term, the low radiologist to patient ratio in middle income countries is likely to provide ample growth opportunities in these countries.

We are very sorry, but an error occurred.
Please contact [email protected] if the problem remains.