Highly Accurate In-Silico Clinical Trials


Establish Proof of concept for a system that would enable conducting highly accurate in-silico clinical trials to be conducted.


A highly accurate system for conducting in-silico clinical trials would provide accurate forecasts about which interventions will succeed and in what way. It would enable a massive reduction in the cost of drug development and discovery, thereby reducing prices and ensuring greater accessibility of treatments.

Relevant Technologies and Techniques

    Artificial intelligence Digital twins DNA sequencing Epigenomic mapping

@mkaeberlein, @Fossel, this is one of the breakthroughs that was discussed at the Future of Longevity Lab. Do you agree it would be a breakthrough in terms of longevity? And is this something XPRIZE needs to do, or do you think this is likely to be achieved anyway whether we organize a prize competition around it or not?

I think the specific phase of a trail or therapeutic area should be selected. Because some of those techniques are already here. In 2013 FDA launched a CiPA initiative with a goal to create computational models of human cell that will be used for drug safety testing. Some of the companies now reduced significantly their animal testing and replaced them with simulations because simulations are humanised and so way more accurate in predicting human response to drugs. (2017 paper with the same results: https://www.frontiersin.org/articles/10.3389/fphys.2017.00668/full). So I would say on a preclinical level, it is already employed. And FDA wants computer models to be mandatory at the first stage clinical trials because they are fast in providing results and again accurate. At the same time, there is a reasonable assumption that computer models have even bigger potential and can already be used in the clinic. For examples, this lab from Johns Hopkins used personalized in silico heart model to select the best sites for the heart procedure in real patients https://www.nature.com/articles/s41551-018-0282-2. Of course, more validation is required. But maybe it will be good to design a prize around validation. There are plenty of useful computer models out there. There is also a really cool video that was done using computer simulations of human organ and cells: