Some of the most leading-edge medical breakthroughs are moving faster than the health care system can absorb, panelist after panelist described.

Why it matters: The buzziest advances — like personalized gene therapies or AI-discovered drugs — aren't going to deliver on their potential if regulatory structures or the commercial sector can't enable them.

• There's "unbelievable promise coming out of biomedical research and science," said Monica Bertagnolli, former NIH director and the president-elect of the National Academy of Medicine, who spoke on a panel about cancer care.
• But there's also "this big lag still between getting things tested, getting trials done, and — even worse, even more egregious — getting those really innovative treatments to everybody who needs them."

State of play: Experts speaking on a wide range of panels described the same tension in different contexts and with regard to different emerging technologies. Here are some examples:

1. AI-discovered drugs collide with constraints in the current drug development system.

• Bots may be able to generate all kinds of promising new targets and treatments, but the nature of biology and inefficiencies in the system can still hinder development of new therapies.
• "We are seeing a massive acceleration of our ability to discover high-confidence hypotheses on what intervention will address a particular disease biology," said Daphne Koller, CEO and founder of Insitro, a leading AI biotech.
• "The bottleneck that we really need to resolve at this point is actually clinical development."

2. It's not just new drugs. AI can — and does — allow researchers to identify new uses for old drugs.

• Panelist David Fajgenbaum's nonprofit organization, Every Cure, is doing this work, and has identified more than a dozen drugs for repurposing.
• It's a response to an enormous gap in the current drug development system, which is based on which treatments can deliver the highest financial returns.
• In other words, there's not much incentive for doing research into repurposing drugs, especially generics, despite the potential benefit to patients.

3. Gene editing technologies could revolutionize ultra-rare genetic disease treatment, but costs and regulatory hurdles are still prohibitive.

• "Baby KJ" proved last year that it's possible to make personalized genetic therapies for individual patients with devastating diseases, and the FDA has already attempted to create more of a regulatory pathway for follow-ons.
• But there's a long way to go toward making it a viable option for other patients.
• "CRISPR cures are no longer a technological challenge," said Fyodor Urnov, director of therapeutic R&D at the Innovative Genomics Institute. But at today's costs, "CRISPR has no future ... Baby KJ will be a deeply affecting N of one."

Yes, but: A drug company will never recoup the costs of developing a drug for a single patient, the Broad Institute's Winston Yan said. But there are other ways to think about making the math work.

• "In the hospital, you can get a medication or you can get a procedure. There have been a lot of things that have been focused on, how do we make the drug development faster and cheaper, but what if we think about it as a procedure?" he suggested.

4. Geography is an inherent limiting factor, and all of the progress in cancer treatment has yet to solve for physical distance as a barrier.

• The newest, most ambitious medical advances are concentrated among select institutions located in big cities. That can mean only swaths of the U.S. population actually have access to those advances or the expertise driving them.
• "It is still a zip code lottery. Where you live depends on what your outcomes are, even with that innovative science coming through," AstraZeneca's Mohit Manrao said.

There are instances where the script is flipped and commercialization is outpacing the scientific evidence required to back it up.

Why it matters: There aren't many brakes on some of the newest technologies with the potential to cause massive harms or which pose unresolved ethical dilemmas.

Between the lines: The spiciest session I attended, by far, was a panel discussion about embryo screening.

• One of the panelists was Noor Siddiqui, CEO of Orchid Health, which offers whole-genome sequencing of embryos conceived through IVF. She defended the controversial service as both scientifically valid and ethically sound, empowering parents with maximum information.
• But another panelist, Sarah Lawrence College's Laura Hercher, responded that relevant professional societies have all said "we should not be using this to test embryos right now ... there's absolute consensus in the field."
• Compared with other countries, "this country is in a very unique position where we allow everything, and it's really about the ethical assessment of what we're doing," said bioethicist Vardit Ravitsky.

In keeping with the times, there was a session grappling with the AI chatbot's role in mental health care.

• Millions of people are using chatbots for therapeutic support, but how to do so safely is still an open question.
• "The models of ChatGPT and other LLMs is actually evolving so quickly that the research on them by the time it gets published is already outdated," said Christine Moutier, chief medical officer of the American Foundation for Suicide Prevention.
• "We can't ban this or tell folks not to use it. How do we make it safer for them?" asked Mental Health America's Jessica Jackson.