Publications

Transforming Healthcare with Generative AI: the Big Picture

by
Serdar Uckun
July 16, 2024
Health

All healthcare venture ideas could be boiled down to three categories:

  • Access
  • Value
  • Personalization

Access to Healthcare

Health care is a fundamental human right. In principle (politics aside), an unemployed coal miner in West Virginia should have the same access to healthcare as the CEO of a Fortune 500 company in Manhattan. In this context, access means:

  • Equal opportunities to recognize and diagnose similar health conditions
  • Timely access to healthcare
  • Availability of similar treatments given similar ailments
  • Same quality of care regardless of wealth or insurance status.

This vision is a utopia in the United States as of 2024. Even in countries with socialized/universal health care such as Canada and the United Kingdom, health care favors the wealthy. In the UK, a factory worker with lung cancer might be assigned a surgery date 9 months ahead by NHS whereas the “noble” would be operated on the same week. AI is not going to solve all problems of humanity, but to the extent that AI could help equalize access to healthcare, it would help humanity take a forward step. It is the goal of Leap AI to recognize opportunities to further the goal of “access to healthcare” and to build profitable, successful companies that help average citizens access the same level of health care as the wealthy.

Value of Healthcare

In this context, value of healthcare could be conceived as quality divided by cost — that is, value for money. In other words, a specific treatment that costs much less in the Finland versus the US could be considered a “better value” in Finland. Similarly, if the same amount of money buys someone better diagnostics and treatment in Lebanon compared to Italy, the treatment in Lebanon has better value. Value of healthcare is “low-hanging fruit” that AI technology is expected to have the great impact on the lives of people over the next decade. Therefore,

AI solutions that either increase the nominator (quality) or decrease the denominator (cost) are of interest to Leap AI.

Personalization

With over 8 billion people on the planet, personalization of healthcare (or personalization of anything else including Spotify playlists) sounds like a pipe dream. Yet, it is not. Let’s take cancer as an example. Humans’ germline DNA sequences (what they are born with) are unique like their fingerprints. Certain disorders such as cancer modify cell reproduction mechanisms and insert errors in the genetic code. For instance, all human cancers have multiple families of cancer cells with genetics that diverge from the germline. Some mutant cell families respond to certain treatments; some develop resistance. Darwin (aka population statistics) wins in cancer as well — most cancers evolve to generate new genotypes that are resistant to past treatments. Therefore, effective treatment of cancers is like a whack-a-mole game: treat the cancer; sample surviving tissue; find what new cancer cells are susceptible to; whack them with a treatment that kills new cancer cells while sparing healthy tissues; repeat until the cancer is cured or the patient dies.

Given a market of 8+ billion individuals, personalization of healthcare is the grand calling for AI technologies in service of humankind. Even for those with unrestricted access, there are grand challenges in personalized / precision healthcare. Current treatment paradigms are based on clinical trials, which, by definition, require large cohorts of patients to participate and register statistically significant responses to the treatment in question. When treatments are laser-focused on the genomics of a single individual and their cancer cells, the clinical trial paradigm breaks down. Pharma companies are already having difficulties enrolling 20+ patients on their third-line cancer drugs. Currently, they cannot support “n of one” — one patient only, no clinical trials, entirely customized treatments. Yet, that is the future and we will get there no matter how much we resist it through policy and scientific inertia.

Leap AI is not in the healthcare policy business, but the “n of one” issue will be the most significant impediment to personalized healthcare in our lifetimes.

Leap AI is dedicated to developing technologies that assist in personalization of healthcare, and to engage in policy discussions that enable novel treatments to break the “n of one” barrier.