Advancing Past Ambition in AI + Blockchain for Drug Development: Report from the 2022 Financial Times Global Pharma and Biotech Summit

Advancing Past Ambition in AI + Blockchain for Drug Development: Report from the 2022 Financial Times Global Pharma and Biotech Summit

No alt text provided for this image

The number of medical publications is doubling every 70-75 days across the globe. It is becoming impossible to analyze medical information to inform decisions not only for the public or healthcare professionals, but for the scientists themselves. As a result, medical, drug and health related information is distributed more widely than ever today as the pharma and biotech industry has begun to share it publicly across more channels of communication than ever before in the recent years. In fact, in the first-ever Medical Affairs Innovation Olympics in the pharma and device industry that ended in October hosted by Amedea Pharma , our slogan was "Dare to Share for Our Healthcare". And during the month-long global cross-company ideation event, we witnessed a new 2022 innovation spectrum in which global companies and startups alike competed across the industry, then collaborated, formed important business partnerships that were disclosed publicly, and ultimately accelerated innovation. The definition of innovation in life sciences is expanding exponentially not only due to advances in medicine or research methods but also due to integrating technology, advanced analytics, AI (artificial intelligence) and Machine Learning (ML) from different fields. Yet the speed and proportion of patients it ultimately reaches also depends on our regulatory and geopolitical ecosystems, priorities across health systems, investors, and types of stakeholders required to make decisions that determine access to care or clinical studies for patients.

No alt text provided for this image

To learn how the contemporary leaders are accelerating the implementation of innovation on a larger scale across the industry and global health care system, a few weeks after the completion of our Innovation Olympics, I flew to London to attend one of my favorite events in my field, the Financial Times Global Pharma and Biotech Summit November 7-9. The event explored the most critical factors that link innovation in life science research to investments and ultimately real world experience of new treatments for patients.

In this issue of "InWeekend" I address potential doubts many of the AI, Machine Learning, or even Blockchain skeptics may have across our industry as well as those resistant to face the current inefficiencies in the drug development process and existing inequalities in healthcare. The article summarizes some of the latest cutting edge case studies and most introspective expert panel discussions among chief global pharma and startup executives, scientists, investors, AI, blockchain experts and decision-makers at this event that captivated a global audience. In accordance to the spirit the event inspired, they shared a refreshingly transparent glimpse into the most pressing, universal issues, challenges as well as advances along this evolving spectrum of innovation that prove that AI, ML and blockchain technology are no longer parts of a nebulous ambition but rather a reality that also comes with greater social responsibility for all of us to monitor emerging developments closely and contribute to in some way in the future as well.

This detailed report covers 8 topics across some of the expert panels and interviews I have selected:

1. LOW PRODUCTIVITY OF TRADITIONAL R & D MODELS

2. EXTERNAL ALLIANCES AT EARLIER STAGE OF DRUG DEVELOPMENT

3. INCREASING INVESTMENT IN HEALTH ACROSS THE GLOBAL STAKEHOLDERS AND ECOSYSTEMS

4. INNOVATION THROUGH PHARMA COLLABORATION

5. AI AND MACHINE LEARNING INCREASE SUCCESS RATE OF DRUG DEVELOPMENT

6. IMPROVING BACK TRANSLATION FROM CLINICAL DEVELOPMENT INTO DISCOVERY TO REPURPOSE DRUGS OR FAILED EXPERIMENTS

7. EMERGING DEALS AND INVESTMENTS IN AI AND R & D IN LIFE SCIENCES

8. COMBINING AI WITH BLOCKCHAIN AND NEW INCENTIVES FOR PATIENTS TO SHARE THEIR DATA TRANSPARENTLY

Previously, I have referenced Bayer as one of the innovative big global pharma giants which is investing significant resources in educating the lay public across different media platforms including LinkedIn or podcasts, together with other companies such as Pfizer, Boehringer Ingelheim, Lundbeck, Eli Lilly, among others. At the Financial Times Pharma Summit (), the President of Bayer, Stefan Oelrich was interviewed by Jamie Smyth to discuss his views on maximizing the potential of collaboration to build stronger ecosystems.

1. LOW PRODUCTIVITY OF TRADITIONAL R & D MODELS

No alt text provided for this image

First, Oelrich openly recognized that traditional R & D models are no longer productive in pharma and biotech - an opinion shared among the leaders across various panel discussions, including David Loew , Chief Executive Officer at Ipsen , Frank Nestle , Global Head of Research and Chief Scientific Officer at Sanofi , among others. In fact, Oelrich acknowledged that the most dated knowledge now sits within pharma companies and they need to adopt a more continuous exchange of data and knowledge between academia and research. In a different expert panel, David Loew noted that most of the research at Ipsen is carried out externally. However, he cited a range of difficulties of working with CMC (chemistry, manufacturing and controls) development companies as biotechs are running into increasing trouble finding the right molecules, structure and production process.

Until 2020 there was a decline in R & D productivity. Generally, drug discovery is complex as there are 20000 human genes, and understanding where, when they are expressed, in which tissues, variants, which disease states, or how the products of those genes interact with each other through thousands of cellular pathways takes years of research and leads to a 90% failure rate of pre-clinical candidates and frustration. An expert panelist, Werngard Czechtizky Chair of 阿斯利康 Global Chemistry Leadership referred to the traditional approach to drug discovery as "reductionist that identified one pathway, and focused on identifying a target that is related." Thus, pharma is searching for ways to design clinical trials more effectively, discover targets to all disease states which can hopefully increase productivity and return on investment for R & D.

Another panelist, Reza Sadeghi, Chief Executive Officer of BIOVIA , a premier global software company for chemical, materials, and bioscience research companies, provided a general perspective by pointing out that aspirin is a small molecule that contains only 20 atoms; large molecule drugs can approach 2000 atoms; biologics - 20000 atoms, and in cell / gene therapies the numbers increase exponentially as does the cost.

As a result, Ger Brophy , Executive Vice President, Biopharma Production at AVANTOR stated that the true product in drug manufacturing and development is the process. When pharma and biotech companies outsource CMC companies, they are likely to understand the impurity levels and requirements well in monoclonal antibodies, but in the areas of gene and cell therapy, nucleic acid drugs, they are changing quickly. He says that "for us to continue at the pace and continue to be successful, the relative input from the different parties have to be aligned appropriately, reviewed internally, and combined in a manner that is synergistic."

In addition, processing and analysis of data from patients who participate in clinical trials continue to be redundant and inefficient in pharma. Loew shared an example - in clinical trials, a pharma / biotech company programs a database and then asks the hospitals to fill in a separate case report form which is far from being efficient. Not every institution requires electronic medical records (EMR) completely, and many EMR systems do not operate with each other across hospitals or health systems.

Thus, Loew recommends to ask the hospitals to stop capturing all different data fields, and start standardizing fields required to capture across different specialties to have a harmonized database. Such an initiative will likely require collaboration with physician associations in Cardiology, Oncology, and Hematology, for example. So in the panel discussion, he illustrated real world success stories in advancing healthcare continuity of Kaiser Permanente, which follows patients from the general practitioner to the specialist all the way to the hospital completely, and also in prospective randomized trials of vaccines which fully run on EMR with Finland and Israel that are leading the way in this category, in his opinion.

Another limitation of pharma's traditional communication model with healthcare systems may be successfully bringing a drug to approval for a high priority condition yet which is rarely if ever considered for treatment in the real world, possibly worsening healthcare quality. The example Loew provides is a stroke patient who develops spasticity with evidence showing that only 15% of spasticity patients receive the neurotoxin now indicated. Thus, the three approaches Loew recommended in order for the pharma and biotech industry to work together with governments, policymakers, and consumers externally were:

  • Partner externally across pharma/biotech and earlier to advance drug development
  • Harmonize EMR across specialties and health systems to accelerate clinical trial completion
  • Partner with health systems more systematically to directly improve healthcare quality and appropriate clinical point of care decisions

So, in addition to 1) increasing difficulties in identified drug target interactions based on currently available pharma databases in complex disease areas and therapies such as gene and cell therapy, 2) limited cost effectiveness and speed to address complexity, 3) insufficient partnerships at the right stage of development; 4) redundant or fragmented data communication in clinical trials and study designs; 5) inconsistent partnerships with health systems mentioned above, 6) slow, challenging, wrong patient recruitment, one additional critical source of inefficiency in R & D and hot topic of discussion in another panel highlighted later in this issue was: 7) the lack of Back Translation or a Reverse Feedback Loop after drug candidates fail.

In fact, learning about why certain drugs do not work and in which patients was a topic Guru Sonpavde, MD had discussed on my @Alloutcoach podcast. At the time, he was the Bladder Cancer Institute Director at Dana Farber / Harvard Medical School, and is currently Genitourinary (GU) Oncology Director and Christopher K. Glanz Chair for Bladder Cancer Research at the Adventhealth Cancer Institute in Orlando, Florida.

2. EXTERNAL ALLIANCES AT EARLIER STAGE OF DRUG DEVELOPMENT

No alt text provided for this image

One important strategy to improve R & D productivity that is emerging is to form alliances with other companies at an earlier stage of drug development. Oelrich presented Bayer's approach in his interview.

Bayer is known as a pioneer in cardiovascular medicine which continues to be its primary focus, joined by oncology, cell and gene platforms for rare and immunological diseases. The company has completed over 40 merger and acquisition deals in the last 4 years.

He highlighted significant investments his company has made into "Leaps by Bayer", which he refers to as "a venture capital strategic engine with a mission that goes after 10 most pressing leaps that threaten humanity."

In the recent years, Bayer built a strategy to form an alliance with BlueRock, a small company that investigated curative treatment options through regenerative stem cells.

This was a starting point of a strategy of a smaller biotech that would tap into regenerative stem cells, gene editing and augmentation as BlueRock has the industry-leading induced pluripotent stem cell (iPSC) platform which enables?programming mature, differentiated cells back to iPSCs to regenerate cardiac cells, restore heart function after a cardiac event, or injecting cells that produce dopamine neurons to restore brain function to possibly cure neurodegenerative conditions such as Parkinson's.

Bayer has recently attempted to build a portfolio of companies and alliances, acquired Asklepios BioPharmaceutical, Inc. (AskBio) to expand its gene therapy platform, formed partnerships with Mammoth Biosciences based on CRISPR/CAS9 to advance gene editing as a result of which 6-7 products are now already being investigated in the clinic.

The pharma giant now owns full rights to Vividion Therapeutics, Inc. ’s proprietary discovery platform, which comprises three integrated, synergistic components: a novel chemoproteomic screening technology, an integrated data portal and a proprietary chemistry library.

Oelrich explained the advantage of this collaboration with Vividion. The company has discovered a technology that allows us to target molecules previously thought not to be "druggable". Most of the drug targets function like a deep pocket, where the molecule goes into the pocket, up or downregulates the target, but 90% of the targets are shallow and difficult for attachment.

As a result of its partnership strategy, Bayer has two clinical-stage programs with gene therapy and cell therapy to cure Parkinson's disease at the moment as Oelrich enthusiastically leads his company to "let innovation run its course".

Mergers and acquisitions involve a necessary organizational transformation which Oelrich also characterized by outlining Bayer's approach to managing alliances. Bayer manages the biotech companies it partners with at an "arms-length" with their executives serving on Bayer's Board, and they are independent. In return, the smaller companies have the benefit of utilizing Bayer's libraries, knowledge, pharmacovigilance and IT support which ultimately ensures productivity.

3. INCREASING INVESTMENT IN HEALTH ACROSS THE GLOBAL STAKEHOLDERS AND ECOSYSTEMS

No alt text provided for this image

Like Stefan Oelrich, David Loew at Ipsen, also shared his company's strategy to partner with companies as early as three years before drugs are tested in humans. Earlier stage alliances can improve drug development productivity, decrease risk, and also be more cost effective alternatives for pharma companies.

Rising complexity of drug development, inflation, and demands from the more educated health consumers call for a critical review of the global ecosystem and geopolitical climate to understand why investing more in healthcare despite uncertainty translates into a healthier economy as well.

Oelrich provided a global overview of recent trends to illustrate all the key factors and players in the geopolitical environment to advancing drug development and medicine. The event took place in London with a slight majority of speakers, executives, and attendees from Europe compared to other continents, so European industry perspectives and immediate needs were featured most often throughout the discussions. Yet Oelrich shared his opinions and latest statistics to provide a balanced, global outlook on drug development.

  • He described the venture and private capital investors in pharma and biotech in the US as more vibrant than in Europe at the moment, yet expressed Bayer's equal interest in Europe and China as well
  • Evaluations of companies have decreased over the last 1-2 years
  • Yet "the attractive [drug] assets are still well financed, especially pre-IPO. The assets and partnerships with Bayer are primarily with pre-IPO."
  • Europe still depends on century-old technology across healthcare and the most prominent digital companies are based in the U.S.
  • Governments are elected for 4 years, but the life science cycles across our industry are usually 10-20 years and this difference presents a problem

Asked directly from the Senior Clinical Director of the NHS from the U.K. about the idea collaboration model between the industry and the government, Oelrich cautioned other pharma and biotech companies to ensure that health ministers are not the ones who drive the Industrial development policy which he believes poses a clear risk to the pharma industry because "it is relegated to talk to the health authority when there needs to be a strong innovation and industrial agenda" that needs to be considered as well.

OECD is a global Organization for Economic Co-operation and Development which gathers many countries together to find solutions to common problems to improve their local and regional economic policies.

  • OECD spends about 8% of GDP on health, in Western Europe 8% is also a benchmark, but the UK is below the benchmark which is a gap in a critical investment into the health of our next generations.

A memorable call for action from Oelrich I found inspiring is to move the pharma / biotech sector away from being considered as a cost.

We can translate into numbers the dividends of a healthy population. If a country spends more of its GDP on health it also translates into a healthier economy.
-Stefan Oelrich, Bayer CEO

During the interview, Oelrich boldly tackled the controversial topic of drug pricing and how the industry is pivoting amidst the global inflation. The recent Inflation Reduction Act in the U.S. signed by Biden has granted the government more power to negotiate lower prices on certain medications and limit the growth of prices. The Bayer CEO acknowledged that he is beginning to scrutinize the company's pipeline more critically especially among small molecules to better understand the return on development, and emphasized innovative cell and gene editing therapies throughout his remarks. He also reminded everyone that the IRA is not likely to be implemented until the next U.S. President's administration.

The discussion then evolved into drug price differences in the U.S. compared to Europe. Americans pay a lot for their drugs but Europeans seem to benefit. Thus, Oelrich was asked directly whether he thought that "Europeans are free riding on American drug prices".

  • He referenced latest data demonstrating that the number of publications from Europe, China, and United States have now equalized, so European knowledge generation must convert into European capitalization.
  • In addition, a lot of the knowledge that is translated into capital value comes from European academia which is tax funded.
  • The proportion of the drug budget relative to the overall health expenses is the same in Europe and the U.S. Out of the overall healthcare spend, Europe spends 5-7% on the patented, new, innovative drug development. In the U.S. 7% of the overall healthcare spent is on drugs.

So, he cleverly turned the question back to the audience to reflect and ask if they are "free riding on our [European] universities".

Ultimately, Oelrich argues that the pricing problem is the amount of overall healthcare spending in the U.S. and the question is whether the health system in the U.S. is sufficiently efficient.

4. INNOVATION THROUGH PHARMA COLLABORATION

No alt text provided for this image

The covid19 pandemic was certainly the unfortunate necessity that accelerated collaboration across the competitive pharma and biotech sector, as well as research of antimicrobial resistance. And while the desire to collaborate more across the industry was evident across all leaders at the event, they seemed to agree across the different sessions that it is becoming more difficult to collaborate. Thus, addressing the observations that lead to the intriguing question of whether or not the industry will "naturally succumb to the competitive pressures and change its course from the recent trend to collaborate across different companies and stakeholders", Oelrich made a passionate plea "for the planet to focus on health" and for all the stakeholders in politics, the regulators, the private industry, and academia to interact more continuously. He reminded leaders from the NHS in U.K. and others in the audience that "we need medical progress to advance the health agenda" and increase the life expectancy of patients by investing in research and not invest a disproportionate amount of money into hospitals, for example.

Therefore, to ensure that the positive collaboration trends throughout the pandemic will continue to move forward beyond formal mergers and acquisitions, we have to understand the areas and types of collaborations in drug development to prioritize, key players and decision makers and how to influence them.

John Haughey , 德勤 ’s Global Life Sciences & Health Care Consulting Leader pointed out several trends and offered a personal guide to success for the biopharma sector during the Panel: Partnerships - Driving innovation through collaboration led by Clive Cookson, Science Editor at Financial Times . Haughey recommends three areas in which to drive innovation through collaboration:

1. R & D

  • While prior to 2020 there was a decline in R & D productivity, the proportion of new drugs that were co-developed increased from 32 to 46% in 2021
  • In fact, co-development is more likely to lead to drug approval and success as nearly half of the forecasted revenues from late-stage pipelines were a result of collaborations or scientific partnerships

John predicted that the positive correlation between drug development collaboration and success will continue as there is significant interest in investments from venture and financing

2. Data and AI

  • There is an explosion of data tech investments across the life sciences value chain. The needs for bio-informatics and computational biologists will only continue to increase.

3. Environment, Social, and corporate Governance (ESG) and Sustainability

  • Healthcare industry is responsible for 4.4% of global greenhouse emissions. If it were a country, it'd be the 5th largest emitter.

As a result, Haughey predicted that an organization's ability to engender trusting relationships among peers, suppliers, governments, regulators, and healthcare professionals will play the decisive role in success and differentiation from competition.

Luckily, tremendous global partnerships are underway independent of any current drug development program, which offer unprecedented promise to patients across the globe. Nerida Scott , Head of Johnson and Johnson Innovation, EMEA highlighted one recent winning data collaboration between a big pharma company and the government - the partnership between J & J and the UK Biobank, which holds the largest human genome sequence project in the world to sequence the genomic data from 500,000 people. J & J contributed whole genome sequencing to the UK Biobank, which is a long-term state investment into the industry.

Growing TRUST in the Pharma and Biotech Industry

Latest research from Deloitte suggests that trust in the pharma industry has still not improved significantly despite the evolution of development throughout the pandemic according to John Haughey. Although, David Loew from Ipsen views such results with caution reminding the panel that published opinion is often inconsistent with public opinion.

In order to unequivocally improve trust in the pharmaceutical industry, Haughey of Deloitte recommended the following four priority areas in which all of the key biopharma stakeholders are making demands today.

  1. Humanity - is the intention of the sector to really benefit patients?
  2. Transparency - reporting study results or adverse events in lay terms patients understand
  3. Capability - demonstrating an ability to deliver products seamlessly across healthcare
  4. Reliability - ability to bring quality products based on robust methods in clinical trials

Loew believes that a certain amount of time spent together, preferably in person can establish trust to dismantle tense situations, which often call for root cause analysis by a third party. Thus, more human interaction with deliberation is critical to form collaborations we may not have even yet imagined possible. Loew emphasized the different drivers of human dialogue to accelerate new relationships:

We need some deliberate redundancy and inefficiencies to rely on others either transactionally, contractually or in goodwill. The infrastructure is open to engender such relationships.
-David Loew, CEO Ipsen

A number of health inequalities emerged during the Covid19 pandemic, with underrepresented populations in trials or with low access to care, understanding of health literacy. Kate Bingham ( Kate Bingham ), Managing Partner of SV Health Investors, spoke about NIHR (National Institute for Health and Care Research) 's experience in partnering with various pharma companies to recruit patients for vaccine trials. NIHR translated educational materials related to the vaccine in 10 languages, set up a registry of 500 thousand people, a third of which was over the age of 60 but only 8% represented minority populations. In retrospect, Bingham recalls how collaborating with patient advocacy groups significantly increased patient recruitment in certain institutions.

Some of the gaps in trust can simply be traced back to what Loew refers to many governments' "acute lack of deep insights of the mechanisms of the industry and the possibilities of partnerships". He believes Singapore, Switzerland, Ireland are some of the governments that have figured out the needs of the industry from a pharmaceutical production aspect while many of other countries in Europe have a more superficial understanding of how the industry works, and as soon as there is a government change any learning progress until that moment is eliminated and restarts from zero.

5. AI AND MACHINE LEARNING INCREASE SUCCESS RATE OF DRUG DEVELOPMENT

No alt text provided for this image

Even more unclear and variable is not only the government but the internal understanding of how AI and ML increases the probability of success of new drug candidates across pharma and biotech companies. The Panel: Drug discovery - Applying data science in early-stage R&D moderated by Naveed Panjwani provided a practical, state of the art perspective on extent of success to expect realistically when integrating AI, ML, and Blockchain into drug development. Reza Sadeghi of BIOVIA 达索系统 of the BIOVIA group recognized that the industry is gathering more insights using advanced analytics and applying AI methods.

  • Research shows that the probability of success for orphan drugs ranges from 3 - 6% in Oncology to ~33% in Vaccines
  • "We have a true chance of doubling these rates of success through AI because we are benefitting from billions of records that will allow us to assess the function, not only the structure of the new drug candidate"
  • There is a huge benefit of having human data indicated by the doubling of the success rate of phase 2 clinical trials in which there is some human genetic validation of that disease.

How does AI help us identify not just novel targets but disease relevant targets?

Frank Nestle , Sanofi noted that even after identifying a relevant disease target, scientists must manage thousands of different attributes of the drug such as potency, specificity, safety, bio-distribution in the body. Each of these steps can now be modeled using predictive AI models.

In addition to chemical and pharmacological variables, the effect of a drug candidate on the biology requires analysis of functional genomics in a knowledge graph, all of which have to be brought together by integrating different ways patients and their diseases express themselves, or their phenotypes, according to Werngard Czechtizky , Head of Medicinal Chemistry for Respiratory and Immunology at 阿斯利康 . Thus, according to Czechtizky, as a consequence of poly-pharmacology, in which a drug is designed to act on multiple pathways, it can then be repositioned and modified to improve its benefit/risk ratio profiles.

  • Goals of integrating AI into R & D are to increase the speed and quality of research as well as improve the cost and sustainability

Sadeghi noted that computing power has improved dramatically and agreed with Czechtizky about the complexity of system biology. He noted that certain events take milliseconds, while others can take months to manifest, such as atherosclerotic plaque. How molecules aggregate, fold, the structure of the mRNA, their immunogenicity are other factors that influence the probability of success.

Frank Nestle shared his optimism with Reza and other panelists. "We are living in the age of structural enablement", said Nestle, referring to protein databases and predictive models with the CryoEM (cryoelectron microscopy) model that enables us to visualize and understand large macromolecule structures better, at near-atomic 3D resolution. Small molecule pipelines have undergone a paradigm shift and the next frontier is to discover biologics which are more complex by orders of magnitude.

Mads Dalsgaard , CEO of Cureteq AG - a Partex Company , an AI supported Biotechnology and Asset Management company shared his recent search for PD1/PDL1 inhibitors for cancer which yielded 40,000 relevant results that are impossible to review with human capacity yet which AI helps digest and build upon the knowledge that keeps growing.

The ultimate questions that remain, however are what method or algorithm is used by AI and its accuracy, according to Sadeghi. In drug design some of the popular key performance indicators (KPIs) or metrics may be target success rates, target cycle times, early development self-originated success rates, and discovery productivity. New drugs fully discovered, or rather imagined with AI training sets using Deep Learning generative adversarial networks (GANs) are already entering human trials, as featured in previous issues of "InWeekend". Only a small proportion of databases used by pharmaceutical manufacturers describe the structure-activity relationship between molecules and their biological functions. Deep learning approaches that employ machine learning such as GANs augment this gap to accelerate drug design, repurpose old drugs, and improve drug target prediction accuracy compared to all other tested methods. Public databases including PubChem, ChEMBL, Binding DB, ExCAPE-DB, ChemDiv, ZINC, QM9, L1000, GDB-17 and Reaxys Database (commercial database) standardize and integrate multiple-sourced datasets. Still, some models of drug target interactions are based on small benchmarking datasets, do not sufficiently validate their performance and have low predictive value.? In September, I presented a systematic review of various metrics currently used to evaluate GANs/Deep Learning for drug design at the 6th Annual International Conference on Drug Delivery and Formulations conference in Scotland (request a copy of the abstract which includes a table of the metrics and benchmarks used). Despite the progress demonstrated, more quality benchmarks and consensus about metrics of these advanced approaches for drug discovery are needed across pharma and biotech.

In this Drug Discovery panel, Reza Sadeghi, recognized the need to improve the AI algorithms in drug design in order not to require as much power of computation because the number of products synthesized is decreasing, time to approvals are decreasing, but new drug targets are also becoming more complex.

Frank Nestle from Sanofi looked ahead favorably at the ambition of AI decreasing the time to IND submission of a small drug candidate from an average of 5-6 years by ~20%, or about one year.

How are pharma companies sharing R & D data with each other to accelerate drug development in the entire ecosystem?

In one year, Top 20 pharma disclosed 50 partnerships, mergers, or acquisitions. From a collaboration perspective during drug discovery, big pharma companies currently submit their R & D data to KMR, a global leader in benchmarking, analytics and performance management with an exclusive focus on biopharmaceutical R&D. About 3000-5000 compounds are tested to get to a clinical candidate and if this can be reduced to 300-500 compounds it reflects tremendous savings in time.

  • However, Nestle pointed out that most of the current success stories are based on molecules where the chemical space and databases are rich with information used to train models yet it is much more difficult to generate new compounds de novo with AI regardless of your databases.

Ultimately, R & D costs and productivity are best expressed by time savings and reduction attrition of compounds.

The majority of the cost (~2/3) of drug development is burned in R & D, which can be expedited from the conceptualization and ideation stage all the way through validation. The critical time and costs of waiting for in-vitro, pre-clinical or clinical validation can be reduced significantly if a digital twin or a synthetic control can be used in a clinical trial and therefore predict how a monoclonal antibody can bind in a clinical trial, according to Mads Dalsgaard.

Nestle shared one of the most exciting current trends to confirm the real world value of AI in drug discovery that is already taking shape.

At KMR, we [Sanofi] are #1 in biologic success rates among the 11 pharma companies. Applying predictive models is accelerating our productivity already.

While computational power has improved exponentially, Dalsgaard noted that the best engines crawl public databases however "there are lots of proprietary data which are still too isolated."

For true breakthroughs, I believe big pharma is sitting on tremendous treasures of data which is a competitive edge but finding a solution for sharing them will take us to the next level!
-Mads Dalsgaard (CEO, Cureteq AG - a Partex Company )

One of the best, most popular examples of a cross pharma collaboration in drug development discussed in detail at the panel was the MELODDY project.

Last June, 10 major pharmaceutical companies — Amgen, Astellas, AstraZeneca, Bayer, Boehringer Ingelheim, GSK, Institut De Recherches Servier, Janssen, Merck, and Novartis — inked an agreement to build a shared platform called MELLODDY (Machine Learning Ledger Orchestration for Drug Discovery). In partnership with Nvidia, Owkin, and others, the group sought to leverage techniques like?federated learning?to collectively train AI on datasets without having to share any proprietary data.

Frank Nestle openly expressed his enthusiasm about MELLODDY stating it represents the future framework of drug development by allowing the industry to leave the data at the original site while still sharing it to improve the models and databases continuously. Werngard Czechtizky, AstraZeneca, agreed there is a clear commitment from the industry to contribute more data. She added that molecules may be characterized in different ways without necessarily always sharing the entire structure, and also reminded the panel about the importance of the quality of data shared, not just the quantity. Reza Sadeghi provided a more cautious, data science perspective on this milestone project achievement, which in his words has been a partial success as he believes there was some sharing of data, though not all of it was useful as data science is a critical factor to ensure that confidence and security of the data is provided where it is needed. In Sadeghi's opinion, the evolution of blockchain and graph data is the critical turning point to ensuring such data integrity on a wider scale without revealing proprietary secrets.

6. IMPROVING BACK TRANSLATION FROM CLINICAL DEVELOPMENT INTO DISCOVERY TO REPURPOSE DRUGS OR FAILED EXPERIMENTS

No alt text provided for this image

Like some of the most popular, best recipes that are based on an unintended initial mistake in the kitchen, many medications are discovered based on errors in the research lab.

Unfortunately, the feedback loop from clinical development back to drug discovery that informs the scientists of the mechanisms of failed experiments to analyze the reasons for the results in more detail, has been broken. Werngard from AstraZeneca says this is not a closed loop but it is improving in better characterizing patient cohorts, for example.

According to Frank Nestle, a change in culture of communication in R & D is necessary. He says "We require a change of culture - the KPIs we focus on is key response objective, safety, PK (pharmacokinetics), drug levels, certain PD (pharmacodynamic) endpoints, but back translation to learn what causes non-response is one of the most important questions to learn in healthcare. It allows us to understand the molecular mechanism of action and reposition it to the right indication and patients, maybe even to the right individual patient so Back Translation is a must-have in our industry"

Dalsgaard also points out that failed drugs pose no concerns in data sharing and the FDA has a critical path to facilitate and prioritize Back Translation.

How are the advancements in AI and ML influencing the types of clinical endpoints pharma are now testing?

Dalsgaard and Nestle agreed that clinical endpoints may be digital endpoints, and that they do not have to necessarily take 52 weeks to be evaluated in a trial. If both public and private databases can be categorized using ontology that links the target or biological pathway to the patient we can create new endpoints versus traditional ones found in the guidelines. Thus, changing clinical endpoints is a critical challenge which must be addressed to pave the way for new indications, and new endpoints. In neurogenerative diseases in particular, new clinical endpoints may be absolutely transformational in R & D. Then the next step is to influence regulatory authorities to accept and validate them.

7. EMERGING DEALS AND INVESTMENTS IN AI AND R & D IN LIFE SCIENCES

No alt text provided for this image

The prospect of accelerating productivity in R & D and ultimately quality of treatments to be developed in the future is attracting a diverse community of investors. Though the evaluation of many pharma companies has decreased and money is becoming more difficult to raise there is no shortage of investors nor capital available in life sciences by any means. A lively panel titled Dealmakers' dialogue: Where are investors in life sciences putting their money? examined the critical steps and latest trends of investments into AI for drug development and R & D in life sciences in general.

First, there is a substantial difference between the perspectives of the private and public investor even throughout the recent uncertainty. Many biotechs have begun to enter the market at a premature stage with an evaluation that is 30-90% lower than the IPO, however these are prime candidates due to their intrinsically higher value for some of the private investors who double down when the public does not, according to Jeanne Cunicelli EVP, UPMC Enterprises, one of the experts on this panel. Nevertheless, some biotech companies submit incomplete pre-clinical data to the FDA too early, said Melanie Lee , CEP at LifeArc, which as a result contributes to their lack of maturity and subsequent undervaluation. According to James Peyer, PhD the best returning period for private equity venture capitalists was 2009-2020, when they made many Series B investments in companies that did not have a diversified drug portfolio, unlike many other "tourist investors" who did not want to take on the high level of risk at an early stage. Thus, Peyer' company Cambrian Bio is supporting new solutions for life sciences to "efficiently de-risk the most innovative breakthroughs" in pharmaceutical research in academic institutions to invest less money and earlier, while attracting not only specialist but a full range of types of investors.

Cunicelli reassured the audience that there are innovating ways to find financing and models that increasingly include long-term royalties. For example, some investors are demanding synthetic royalty payments even before a drug is discovered. Peyer noted that AI companies commonly ask for royalties in exchange for access to their target discovery platforms which can identify viable targets quickly and more efficiently (less attrition of molecules).

As a PhD scientist and successful investor and pharma executive, James discussed the role of his background in influencing the focus of his company Cambrian Bio . He had observed that a single intervention in the lab can extend a mouse's life span by 30% by improving various preventable chronic conditions at once, which he had found most exciting throughout his work. Yet while it is not feasible to run a clinical trial for aging, life science companies can create breakthrough treatments by studying the underlying causes of aging individually. Thus, Cambrian is built around supporting longevity research that addresses upstream causes of chronic disease that are preventative, he says. As a result, Peyer believes that we have many promising preventative drugs that can prevent cancer, cardiovascular disease, immune dysfunction, and neurodegenerative disease which the growing data, AI and ML technology, and knowledge can demonstrate faster in the coming years.

Looking ahead into the future, James made two predictions during the panel discussion.

  • Hypothesis #1 - Tech companies will have increasing access to AI that will be exclusive to them, and not available to the public, and pharma will join the tech sector to own this technology.
  • Hypothesis #2 - There will be more "sticky companies that will prefer to own the lifelong patient" with a more long-term approach to healthcare quality, similar to Amazon's "One Medical" to gain a larger influence in primary, preventive care.

Thus, James described his enthusiasm for a preventive and more proactive, rather than reactive approach commonly seen across life science research to address tough, intractable conditions. This "dealmakers" expert panel shared quite a promising outlook on the growing diversity of drug development deals to expect in 2023 and beyond. Ultimately, the experts illustrated both a patient and business case for the entire pharma and biotech ecosystem to take initiative, intervene and invest earlier in both the disease and the drug development process.

8. COMBINING AI WITH BLOCKCHAIN AND NEW INCENTIVES FOR PATIENTS TO SHARE THEIR DATA TRANSPARENTLY

No alt text provided for this image

So it is clear that pharma collaboration, healthcare data, advanced tech and AI software continue to accelerate drug development as well as improve its quality. However, the relationship between diseases, patients, clinical trials, access and affordability of new or investigational treatments, key decision-makers and stakeholders across the geo-political and economic spectrum promises to be complex. As disease and drug development demands grow, so do those of the patients. One interview at the Financial Times Pharma event CEO interview: Can the healthcare industry achieve a true digital pharma revolution? featured a courageous, heartfelt discussion with Gunjan Bhardwaj , CEO of Partex, that explored the critical factors that determine how likely the ultimate healthcare consumer, the patient, is to access the best specialist physician, medication, or investigational treatment in a clinical trial and potential models that distribute incentives, ownership, and accountability for growing healthcare data more evenly and transparently across the entire life science value chain.

Bhardwaj is the author of the book "Inside the cockpit: navigating the complexity of drug development with AI and Blockchain." and throughout the interview, he challenged the pharma and biotech industry to truly redefine patient centricity by addressing its core, underlying problems directly in the context of health equity.

First, Gunjan questions the true extent of significant progress made in centering on patients' true needs throughout the drug development cycle. He is inspired by people dear to him in his life with cancer who were not fortunate enough to receive quality care or opportunities to enroll in clinical trials, and illustrates the grim reality that patients diagnosed with cancer really do not know much about their particular type of cancer, the most appropriate key opinion leader (KOL), researcher, or the appropriate current, ongoing clinical studies investigating cutting edge treatments in which to enroll. Notably, Johan Lauritsen is a chronic patient, CEO of a startup called Probe which links patients to appropriate clinical trials seamlessly, who recently won the Evidence Generation event at the first-ever Innovation Olympics in pharma organized by Amedea Pharma who I think shares a tremendous passion with Bhardwaj to close this gap in patient access to clinical trials.

The most essential questions pharma needs to address at a root cause or systems level, in Gunjan's opinion, are

  1. Ownership of proprietary data
  2. Decision-making in clinical trial validation, indication prioritization, and drug candidate identification, and finally
  3. Business development and licensing cycle times of drug assets.


  1. DATA OWNERSHIP. He believes the inherent problem lies in two options to manage proprietary healthcare data currently - to either anonymize the owner of the data or to require patients to be philanthropic and donate their data. So Bhardwaj asks us to consider why patients who themselves create their data should also own "the lion's share of the value associated with this data". As a result, his company has released the world's largest real world data exchange where patients can license their own data to drug discovery programs, facilitated by AI and blockchain. Patients upload their data in a data vault. AI summarizes the case history, the summary is "hashed" (tagged appropriately) and is placed securely on the blockchain database for full integrity. By the end of 2022, Partex will have indexed between 350,000 and 500,000 patients.
  2. CLINICAL DECISION-MAKING. While Bhardwaj acknowledges tremendous progress in drug development with AI, and the most advanced IP to predict clinical trials, indication prioritization, and target identification, he references certain "big ticket" failures as well and therefore believes that in house validation capabilities, or Human Intelligence (HI) is still needed for target identification, a) to validate a hypothesis, and b) to provide feedback to the AI algorithms to improve them.
  3. DRUG ASSET BUSINESS DEVELOPMENT AND LICENSING PROCESS. The speed and frequency of a drug portfolio evaluation in a typical big pharma is too low, usually once every year or two years, which represents a huge disadvantage for patients. Bhardwaj tackles the real world root cause of the problem by suggesting that when that decision is made to focus on an asset or an indication, it is not color-blind and is often likely to be based on passion, politics and economics more than the science to unlock the full potential of the drug. He recognizes why such decisions are relevant for commercial success and viability of an organization with growing complexity of relationships with key stakeholders, payers, and governments. However, what I found inspiring in Bhardwaj's narrative is his evident patient perspective and call for more urgency. Because he explains that if Business Development secures an asset in an average of 2 to 5 years, then looks to partner with companies for another 3 - 5 years, the exclusive patent window is soon gone in the process as a result of which he says "patients who could rightfully have access to this drug are denied this life changing opportunity". So he imagines an ecosystem in which companies scan their pipelines not every 12 or 24 months but every month or quarter, while in addition reviewing external assets outside of their company with which to possibly also combine their drugs, which ultimately accelerates patient access to cutting edge therapies.

Thus, the digital healthcare model supported by Gunjan is revolutionary but quite intuitive and more transparent than ever attempted before. It distributes ownership, incentive, and accountability more evenly than we have seen as well. He mentioned in his interview that decentralized autonomous organizations can only be built using AI and blockchain. So he has created a blockchain marketplace where the patient provides approval to share his/her healthcare data and AI hashes and clips the summary of the data with a unique identifier record. Asked about the quality of the data, he conceded that sometimes data may be incomplete, ICD diagnosis codes recorded, and scans can be illegible, yet he describes the marketplace as a social service for the patient community and incentivizes them to share quality data such that "if they are not able to benefit or survive from it, maybe others could."

Bhardwaj took Jamie Smyth and the audience along a tumultuous journey of his company Partex. For example, he has found many of the European governments supportive of transparently sharing investigator sponsored trials (ISTs) which are typically locked, hidden, and safeguarded as data of failed assays or experiments offers huge value for others, yet while the governments are ready to fund such a cross-company IST database, the industry says no.

His company had filed 30-40 patents, and first approached European Hospitals regarding sharing patient data with their consent while paying them in a marketplace yet they refused for 2-3 years, after which they began to approach the patients directly and promised to completely respect their data. In return, within 12 months, the company had seen over 2 million downloads from patients who had made their choice. As a result, Partex suddenly had 10-12 hospital partners across Europe who had no other choice but to follow the patients' lead.

In addition to sharing data, Gunjan's company is building a panel of hospitals and tumor boards who have expertise with respect to specific indications. AI manages the workflow and summarizes the case history of patients. And all of the patients in the database now quickly gain access to the top-tier, global experts for that indication or disease state to provide a "second opinion" which most of them would not be otherwise able to access.

Looking ahead at patient powered research networks, Bhardwaj feels such organizations should be orchestrated and decentralized, and using a governance protocol, patients should know how their health information is used and for what purpose. Data protection laws are different in Europe compared to the U.S. as in Europe patients' explicit consent are required and data pools are needed, according to Bhardwaj. He recalls one health minister in Europe who argued that because the state subsidizes healthcare, data should belong to the state. He then reminds us that in many Eastern Europe countries or in India, patients cannot even afford molecular diagnostics. So Bhardwaj believes that since patients create their health information, governments should not take over control or in some form steal data but instead incentivize them. After all, transforming health data into a liquid currency "is what putting patients first truly means", he says.

Gunjan Bhardwaj and his company have embarked on a mission to shift the scales and transform the hierarchy, value, and transparency of healthcare that begins with patients' practical interests and ends with acceleration of the drug development process. Partex has acquired a high throughput screening facility in Germany and is in the process of acquiring AI validation capabilities, while also looking for experimental data sets and partnerships with universities and grants.

The CEO interviews and expert panels I highlighted provide a glimpse into the recurrent themes and many other presentations at this truly memorable global summit organized by Financial Times in London this November. It was filled with a passionate, courageous, candid, and a truly distinguished group of leaders, panelists, journalists, and attendees who dared to ask and address the kinds of questions that gave voice to millions of others at a critical yet promising time ahead in life sciences!

Make sure to subscribe to the "InWeekend" Newsletter and share it with your friends! I will soon be sharing the audio version in a new "InWeekend" series on my @Alloutcoach podcast to keep building on ideas and keep the conversations alive so stay tuned. Those of you in life sciences, visit @Amedea Pharma to stay up to date with our latest news, interviews, new data we are generating and sharing continuously on our website and blog!

Leave me your comments, questions, ideas, or topics for future issues!

-Tim Mikhelashvili

CEO and Co-Founder of @Amedea Pharma

Host of the @Alloutcoach Podcast and YouTube Channel where sportsmanship meets the scientific method. Subscribe to both to stay tuned with global, multicultural and real world approaches to leadership, innovation across pharma and beyond, mentorship, company culture and organizational change and interviews with Fortune 100 and 500 Executives, Independent Thinkers, Innovators, Scientists, Olympians, Researchers and more!

Podcast links and availability to the Alloutcoach podcast across the most popular platforms are listed below.

SPOTIFY

https://open.spotify.com/show/6zJgLr6eSlSUliMlKKzKwb?

ANCHOR

https://anchor.fm/alloutcoach

APPLE

https://itunes.apple.com/us/podcast/alloutcoach-tim/id1455249798?mt=2&uo=4

STITCHER

https://www.stitcher.com/show/alloutcoach-tim

GOOGLE PODCASTS

https://www.google.com/podcasts?feed=aHR0cHM6Ly9hbmNob3IuZm0vcy85ODhmNzgwL3BvZGNhc3QvcnNz

RADIO PUBLIC

https://radiopublic.com/alloutcoach-tim-6nBRvZ

POCKET CASTS

https://pca.st/9eu7

OVERCAST

https://overcast.fm/itunes1455249798/alloutcoach-tim

AMAZON MUSIC

https://music.amazon.com/podcasts/e84974a8-0393-439d-8954-92a544afc711/alloutcoach-tim

Great body of work Tim, a lot for us collectively to digest. The core of asset development and the acceleration of the implementation of innovation on a larger scale across these assets is generating unprecedented data. And like you said more than we can keep pace with. Technology is growing at an exponential rate. We have to think differently about approachs and outcomes. No easy answer … this is truly Darwin at work. Keep the great work coming … love this article!!

Chris Cook

2x Olympian, Keynote Speaker & Non-Executive Director

1 年

This is really insightful Tim The AI piece is interesting and you’ve really found a way of simplifying the complex here

Steve Royle

Company Founder | Invited Speaker/Chair/Moderator | Expertise in stakeholder engagement and insight generation

1 年

Great read Tim, very relevant too. You have a knack for making the complex topics relatable!

Rick Johansen

Introduces innovative approaches for KOL identification and the strategic utilization of regional and community leaders. Talks about #KOLs, #comunity leaders, #market access, #launch, and #speaker programs

1 年

Love the take on AI on how it can help in R&D. Worth the read

Johan Lauritsen

Making Treatments Reach Patients Faster @ TrialSync | Chronic Patient @ Spinal Muscular Atrophy | Content Creator @johan_lauritsen

1 年

Exciting and purposeful summary! ???? Everyone wanting to get a jumpstart into the frontier of innovation should give it a read! ????

要查看或添加评论,请登录

社区洞察

其他会员也浏览了