Please login to the form below

Not currently logged in
Email:
Password:

AI: the smart money is on the smart thinking

Looking at the three key areas where AI is being applied: drug discovery, clinical decision-making and clinical trials

AI

The life sciences industry is particularly primed to benefit from Artificial Intelligence (AI). This transformative technology has the potential to save time, resources and, ultimately, lives.

In 2020, the amount of health data is expected to double every 73 days – and this Big Data both necessitates and supports the use of AI and machine learning.

It’s not just data that is seeing a surge in numbers – the financials are too. The market for AI in healthcare is expected to grow exponentially, projected to be worth $36.1bn by 2025, up from $2.1bn in 2018.

This is coupled with venture capital (VC) being increasingly attracted to the sector. In 2017, VC firms pumped in $1.3bn across 103 deals, rising to $2.7bn across 264 deals in 2018 and $3.11bn for 261 deals in just the first three-quarters of 2019.

As COVID-19 brings the life sciences industry to the fore, investment in AI looks set to establish itself as a key component to spur productivity in the industry.

This influx of capital means there are already close to 200 start-ups using AI in drug discovery. Three key areas in which the technology is being applied are: drug discovery, clinical decision- making and clinical trials. It is worth taking a closer look at each in turn.

Drug discovery

AI has the potential to significantly reduce the costs and time associated with drug discovery, across medical divisions and also specifically in handling the immediate COVID threat. Take, for example, the case of baricitinib, originally an Eli Lilly rheumatoid arthritis drug, now repurposed as a promising treatment for COVID-19. Its potential to treat COVID-19 was discovered by UK firm BenevolentAI.

Using machine learning, a small team of four researchers were able to identify an initial 370 kinase inhibitors, and then quickly narrow this down to the six most likely to work – a task which would have been impossible for a small team without AI.

Another recent case, separate to COVID-19, would be that of Insilico Medicine. The company designed, synthesised and validated a novel drug candidate in the space of just 46 days, which is 15 times faster than something the most advanced pharmaceutical companies can normally achieve.

Insilico Medicine was able to do this through development of an AI drug discovery system: GENTRL (Generative Tensorial Reinforcement Learning). Using GENTRL, the company was able to design a novel DDR1 kinase inhibitor from scratch in just 21 days, and the company then synthesised and validated it in 25 days.

Such a combined process would typically take a traditional pharmaceutical player around two years to achieve. Within the first 21 days, GENTRL was responsible for developing 30,000 novel small molecules that could be utilised against fibrosis.

In the following 25 days, Insilico was able to locate and synthesise six of the most promising compounds and perform in vitro tests for selectivity and metabolic stability. The most promising candidate was then tested on live mouse models and displayed favourable activity.

Decision-making

AI could also have a transformative effect on clinical decision-making through the utilisation of the huge levels of genomic, biomarker, phenotype, behavioural, biographical and clinical data that is generated across the health system.

Bayer and Merck & Co provide a perfect example of this. They have developed an AI software system to support clinical decision-making of chronic thromboembolic pulmonary hypertension (CTEPH) – a rare form of pulmonary hypertension.

The software helps differentiate patients from those suffering with similar symptoms that are actually a result of asthma and chronic obstructive pulmonary disease (COPD), and therefore diagnose CTEPH more reliably and efficiently.

The CTEPH Pattern Recognition Artificial Intelligence obtained FDA Breakthrough Device Designation in December 2018. It works by using machine learning to assess image findings of pulmonary vessels, lung perfusion and cardiac check-ups, coupled with checking the clinical history of the patient.

Bayer and Merck & Co hope that radiologists will be able to use this software to identify CTEPH patients sooner, allowing them to be treated earlier.

Clinical trials

When it comes to clinical trials, remarkably, just 9.6% of drug development programmes successfully progressed from phase 1 trials to FDA approval between 2006 and 2015. Admittedly, most trials fail as a result of not being able to demonstrate enough efficacy or safety.

However, some were felled by bad study design, participant drop-out or non-compliance, shortage of money or failure to recruit enough participants in the first place.

AI has the potential to massively streamline this process, save considerable amounts of money and consequently help promising projects get to the finish line.

A case in point is Deep 6 AI. It develops solutions to find participants for clinical trials. For example, it helped researchers at the Cedars Sinai Medical Center identify 16 patients in 30 minutes leading to the recruitment of eight in three weeks – impressive considering researchers had previously recruited just two patients for a cardio study in the space of six months.

The Deep 6 AI system will now enable recruiters at Cedars-Sinai to find patients for 30 trials a year, whereas they only managed one in the previous year.

The compatibility of Deep 6 AI is further shown in the firm’s quick response to applying its capabilities towards COVID-19. The platform can look for symptoms related to a disease that has not yet been diagnosed.

This is particularly useful when trying to identify patients who could have a new disease, like COVID-19, and grouping patients into different categories, such as likely symptomatic patients, likely at-risk people and patients who have similar but non-COVID-19- related symptoms.

Talent, systems and processes

Medical AI specialists are as niche as they sound, and so for the life sciences to really embrace AI there is undoubtedly a question about access to talent. Indeed, despite AI being set to become more and more central to life sciences in the next few years, talent is still drawn to the wider scope of IT and tech companies.

Even in medical AI-focused companies, only 15% of the staff, on average, are AI experts. Recognising this, big pharma firms are already looking to consolidate with AI solutions, if not through acquisitions, then through partnerships.

For example, the aforementioned BenevolentAI has arrangements in place with Novartis and AstraZeneca (the latter also works with DeepMatter), while Boehringer Ingelheim has a partnership with Insilico Medicine.

Not only is this creating the potential for ‘have nots’ – those who fail to build up AI provisions and thus miss out on the rapid advances that medical AI can provide, it also poses a challenge to their future altogether as Big Tech increasingly encroaches into healthcare and provides further competition.

Google, Microsoft, Apple, Amazon and Facebook have all signalled their interest in the sector – and their ability to provide scale from the outset could contribute to the stifling of start- up innovation, similar to the way in which Big Tech has impacted other industries.

There is, of course, no easy fix to this. Adapting to AI requires significant companywide infrastructure changes – it is not merely a software upgrade. Rather it necessitates a rethink on how data is obtained and stored.

But these factors provide a compelling argument that companies and investors should ensure that they are prepared for AI, and that the relevant talent, systems and processes are in place to achieve company goals.

AI to continue its push into life sciences

The aforementioned practical applications of AI certainly indicate that there will be growing industry and investor interest in the technology.

Indeed, this has been evidenced recently by Thermo Fisher’s acquisition of Qiagen in May for $11.5bn. Qiagen employs AI to develop assay technologies to intensify and enrich biomolecules, making them more accessible for analysis.

In the recent past, Qiagen also collaborated with Intel to find a solution that analyses human genomes for as little as $22, and the company has lately made headlines for being one of many AI-enhanced firms to make progress towards a cure for COVID-19.

Therefore, it is highly likely that Big Pharma and venture capital will continue to target AI-focused companies in the coming months and years. In other words – and as is so often the case in the life sciences – the smart money will continue to follow the smart thinking.

Arshad Ahad is a Research Analyst, Life Sciences, at finnCap

10th August 2020

Arshad Ahad is a Research Analyst, Life Sciences, at finnCap

10th August 2020

From: Research

Share

Tags


COVID-19 Updates and Daily News

Featured jobs

PMHub

Add my company
Wilmington Healthcare

A heritage to envy Wilmington Healthcare brings together the data intelligence specialisms of Binley’s, NHiS, Wellards and HSJ to create...

Infographics