Jan Eckardt, Co-Founder and Chief Scientific Officer at Cancilico, made the following post on LinkedIn:
“Clinical research has a couple of problems, two of which are data sharing and the ever-increasing cost of clinical trials.
Data can often not easily be transferred between institutions: You need a lot of paperwork and a long breath before the first data table changes hands. A WEF report suggests that 97% of health data are not being used for any downstream analysis.
The costs of clinical trials are exploding. What’s more, failure rates are staggering, especially in oncology, and many trials do not meet their recruitment goals and become underpowered. That’s devastating in the age of precision oncology where we need to run evermore trials to generate robust evidence for new drugs!
A possible solution to both? Synthetic medical data!
Synthetic data can be generated by artificial intelligence to closely resemble real patient data without being exact copies. So, you get a cohort of synthetic patients that behaves exactly like the real ones without the headaches that sharing real data might cause you.
Plus, you can tailor these synthetic patients to meet eligibility criteria for clinical trials and thereby augment or fully substitute control groups. But it’s not all flowers and sunshine: Synthetic data generation has to adhere to design standards, safeguard patient privacy, and be rigorously evaluated for usability in downstream tasks.
Want to know ho to do that?”
Title: Artificial intelligence-generated synthetic data for cancer research and clinical trials
Authors: Jan-Niklas Eckardt, Waldemar Hahn, Arsela Prelaj, Martin Bornhäuser, Jan Moritz Middeke, Jakob Nikolas Kather
Read The Full Article
Other articles featuring Jan Eckardt on OncoDaily.