Regulatory approval is a big part of getting a new drug to market but it’s a process that can be interrupted if a regulator wants more information. These delays, or PMRs (post-marketing requirements), are an added cost so companies try to avoid them. However, anticipating what a regulator might want is difficult. To remove some of the guesswork, healthcare intelligence company Icon has developed Cassandra AI, which predicts likely post-marketing requirements from major regulators such as the European Medicines Agency and the US Food & Drug Administration (FDA).

“Being able to anticipate their requirements allows us to define a mitigation strategy early in the drug development cycle which could be critical to the overall success of a new product,” says Gerard Quinn, vice-president of innovation and informatics. The problem is that it’s time-consuming to evaluate a pharmaceutical asset against the detailed guidelines provided by the regulators and difficult to know with certainty that a PMR might arise.”

Cassandra AI is trained using machine learning and data obtained from FDA and EMA post-marketing requirements and commitments databases. It can be interrogated about drugs in development and will provide “an instant and reliable verdict on the probability of a PMR and the background context”, Quinn says. “This is an invaluable planning tool and Cassandra’s accuracy is 99 per cent for an FDA PMR and 97 per cent for the EMA. This is a unique offering within the clinical research landscape.”

Cassandra AI is an Irish Times Innovation awards finalist in the Life Science & Healthcare category sponsored by Science Foundation Ireland.