The Lancet Digital Health

That is horrific.

This is exactly why recent privacy critique papers particularly those published in journals like Lancet Digital Health, are sounding the alarm.

This completely destroys the perfect privacy argument.

It shatters the illusion that synthetic is just a convenient synonym for anonymous.

And we determined that in this proof of concept study, we determined that the multi-agentic transplant selection committee could accurately identify those patients who would derive a one-year survival benefit with an accuracy of 92%.

So this was published in Lancet Digital Health a few months ago.

Additionally, we looked at the characteristics that accounted for each agent's decision.

So it was very reassuring to see that the features of the clinical vignette each of these agents focused on really made sense from a clinical perspective, as you can see here.

We also have this particular model, which is a CV risk prediction model to predict the risk of cardiovascular events post-lever transplant, which we actually implemented in the clinic and demonstrated a 52% modification in the behavior of physicians.

to then modify and optimize cardiovascular risk among our liver transplant recipients.

So just an example, and this is in final revisions in Lancet Digital Health.

We also have this model, which we published last year as a multi-center international effort, multi-class artificial neural network for multi-class diagnosis of graft pathology, which overall worked quite well within AUC of 0.866. but we realized that it didn't work so well for diagnosis of rejection.