Biomarkers of Longevity 2.0

Biomarkers of Longevity 2.0: The Shift Toward Actionable, AI-Empowered Biomarkers of Aging, Health and Longevity

- By Franco Cortese, Partner at Deep Knowledge Group and Longevity.Capital

Key Points

The use of biomarkers is an indispensable component of Longevity industry analytics and assessment. It is the foundation upon which measurement of Healthy Longevity and the effectiveness of Longevity therapeutics is built.
Biomarkers are also the primary metric in P4 (precision, preventive, personalized and participatory) medicine, which involves continuous monitoring of the progress of a disease state and recommending a series of corrective interventions in response, to keep patients’ state of health in an optimal mode for as long as possible.
Aggregating biomarkers of aging (rather than biomarkers of disease) is particularly difficult however, as they must be sought in populations of healthy people rather than from among the health data of the hospital populations.
Furthermore, as the scope of P4 medicine broadens, the number of biomarkers and technologies will increase rapidly in the coming years. This makes the implementation of P4 medicine impractical by current, manual means.
Some possible solutions to these problems include the use of AI for the development of an optimal panel of biomarkers of aging, for the analysis of individual patients’ biomarkers of aging, and for orchestrating therapeutic interventions in response to fluctuations in those biomarkers.
As the number of data points increases, it becomes not only optimal, but strictly necessary, to use AI and big data analysis for these purposes. This is one of the foremost goals of the recently-established Longevity AI Consortium at King's College London.

I have worked with the Deep Knowledge Group and its various subsidiary and partner organizations in a number of capacities over the past three years. Before serving as Deputy Director of Aging Analytics Agency, I served as Deputy Director of the Biogerontology Research Foundation, and a member of its Board of Trustees, from 2017 - 2018. One of the projects I am most proud of during my tenure there was spearheading the Biogerontology Research Foundation’s part of a multi-institution project to classify aging as a disease during the World Health Organization’s most recent round of revisions to their International Classification of Disease (ICD-11) - a health care classification system that providing a framework of diagnostic codes for classifying diseases, which serves as the example followed by the majority of developed nations' drug regulatory agencies for determining which conditions are officially classified as diseases and, therefore, which conditions can be targeted by therapeutic interventions, tested, and ultimately approved.

The result of our proposal was the introduction by the World Health Organization of a new extension code for aging-related diseases - XT9T - that can be applied to both new and existing diseases. While the addition of an extension code for ‘aging-related’ via ICD-11 does not amount to classifying aging as a disease in full, it is a step in the right direction, and brings us one step closer to getting the World Health Organization to recognize aging as a pathological process with identifiable and quantifiable clinical indications, which can be intervened upon so as to enable human healthspan extension, compression of morbidity and prevention of age-related disease, during subsequent ICD revision processes.

Report (16_9) - Biomarkers of Longevity.

These efforts are about much more than just theory or semantics. Aubrey de Grey remarked when asked about the impact of the new extension code, XT9T: “The ICD is not just a taxonomy. It greatly influences how drugs are prescribed in most nations, because a physician’s justification for writing a prescription must typically be documented in terms of an ICD code describing the diagnosis. As such, the addition of an extension code to denote aging may have a really huge impact on the financial rewards that drug developers can expect to reap if they bring treatments for age-related ill-health to the market.”

One of the strongest conclusions I reached in working towards the classification of aging as a disease is the incredibly pressing need not just for biomarkers of aging, but the need for actionable biomarkers that could be put into practice today to measure the effectiveness of healthspan-extending interventions today. It is biomarkers of aging and longevity, which can serve as a proxy measure of the effectiveness of longevity-focused therapies, that will pave the way to getting approval for drugs not on the basis of single, narrowly-classified diseases, but on their effects on aging itself, and will lay the necessary regulatory infrastructure needed for the rapid industrialization of longevity to scale, as well as to allow governments to measure the effectiveness of their efforts to increase national Healthy Longevity, a goal that a number of longevity-progressive governments are now keenly moving forward towards.

Now, through our work with Dr. Richard Siow and the Longevity AI Consortium at King’s College London, we are striving to fill this unmet gap and transforming this goal into reality, through the development of AI-empowered technologies and solutions to the development of actionable biomarkers of aging, health and longevity to scale. One special function of the Longevity AI Consortium will be the development of an optimal panel of biomarkers of aging - a specific niche where the implementation is lagging behind the science.

The present article, which is an excerpt from my guest chapter in the upcoming book, “Longevity Industry 1.0: Defining the Biggest and Most Complex Industry in Human History”, co-authored by Deep Knowledge Group Co-Founders Dmitry Kaminskiy and Margaretta Colangelo, aims to outline the need for actionable biomarkers of aging, health and longevity, how AI will be the major driver for achieving it in the coming years, and Deep Knowledge Group’s plans and activities focused on turning these potentials into concrete realities.

1st Edition of Report (16_9) - Biomarker

What are Biomarkers of Aging?

How do we know when a biomarker is a biomarker of aging? It depends on how it is sourced. The current approach to biomarkers is to take them from people at various stages of a disease’s known progress, which in practice means sourcing them from hospital patients. Isolating biomarkers of aging, however, means collecting data which marks the difference between healthy people with as little trace of officially recognised disease as possible. This presents a challenge because whereas hospital patients remain in dedicated areas, and are available for analysis at the doctor's convenience, collecting biomarkers of aging means collecting vast amounts of data from the daily lives of people who have no reason to be in a hospital. There are however options available for aggregating such data, such as the devices which come under the AgeTech umbrella, which will find their user base among those suffering from aging but not officially-recognized disease per-se.

It is important in technology never to let the perfect be the enemy of good, especially when the technology is of great humanitarian significance. For example, in the early 2000s, enthusiastic proponents of the application of regenerative medicine to aging were urging governments, entrepreneurs and thought-leaders to make this a priority. They argued that technology was ahead of the science and the funding, and that while a great deal remains to be discovered about the mechanisms of aging, we already know enough to optimize the existing toolkit of regenerative medicine to address the damage of aging, which is already thoroughly researched. And thus out of this paradigm shift arose the now rapidly-rising Longevity Industry.

Teaser - Biomarkers of Longevity_ Most C

Now once again, the technology is ahead of the science and the funding. And, once again, a paradigm shift is due. Presently the necessary biotechnologies for the implementation of P4 medicine technologies and therapies are already in place. What is needed now is big data analytics to develop optimal panels of biomarkers of aging and to determine how to optimize their implementation with maximum precision.

There is however a risk that governments and governmental or political and strategic bodies may make one or both of the following errors:

They might assume that the missing bridge on the road to Health Adjusted Lide Expectancy-extending P4 medicine is still largely a scientific problem, rather than one of practical technological implementation.
They might assume that, because the current scientific quest for ever more precise biomarkers is not slowing down, we don’t yet have a set of biomarkers precise and sufficiently actionable to take immediate action.

As such, government strategic bodies therefore risk limiting their strategic ambitions with regard to time frames.

For example, in the UK, Theresa May’s government has announced a commitment to adding 5 extra years on the nation’s HALE by 2035, and Aging Analytic Agency has advised the UK’s recently-formed All-Party Parliamentary Group for Longevity that actionable biomarkers of aging will be necessary to meet this goal, and to measure the effectiveness of their development plans to extend National Healthy Longevity.

It is therefore desirable that such bodies have access to a panel of biomarkers which are not only comprehensive but also actionable. A panel of less precise but easily implementable biomarkers of aging would be much more useful much sooner, than an extremely precise and comprehensive panel of biomarkers of aging that is too hard or expensive to translate easily into widespread practical and clinical use across nations.

The Need for Actionable Biomarkers of Aging and Longevity

The development of precise and actionable biomarkers of aging is one of the most prevalent and important areas of longevity research today, as well as an area where practical implementation is lagging behind the science. This is one of the most important diagnostic services that could be offered, and yet it does not receive the attention it deserves compared to the amount of tangible benefits it can deliver.

For example, a panel of aging biomarkers was developed recently which is based on Deep Learning analysis of standard blood biomarkers, which is less precise than the more precise available biomarkers of aging (e.g., DNA Methylation clocks), but which is nonetheless precise enough, and can be implemented by any researcher, doctor and clinician that has access to routine blood tests.

As a further example of actionability, consider that biomarkers of aging have been constructed using Deep Learning-based analysis of photographs of mice, which could quite easily be extended to humans. Their accuracy alone is not enough to make them a research priority, but the increasing video capabilities of smart-phones means that this rapid development of photographic biomarkers of aging (e.g. of the face or the eye) could now be a very actionable area of research whose practical level of precision and accuracy will develop quite rapidly in coming years.

However, the use of AI in longevity R&D is lagging behind in its application to geroscience. While there is a small handful of companies that are working at this frontier, the overall proportion in comparison to the total size of the longevity industry is still quite small. Deep Knowledge Ventures has been identifying and supporting companies working on the frontlines of AI for Longevity since 2014, when it provided the seed funding for Insilico Medicine, now a leader in the application of AI for Longevity research, drug discovery and biomarker development.