Summary: An automated method, based on [18F]FDOPA PET imaging and artificial intelligence, to predict the treatment response of psychotic patients to first-line antipsychotics.
What
A short imaging protocol and a machine learning algorithm designed to distinguish patients with schizophrenia who are unlikely to respond to first-line antipsychotic drugs from those who will respond at first episode.
Why
Schizophrenia and related psychotic disorders are amongst the leading causes of global disability. Antipsychotic drugs are central to treatment but about one-third of patients show limited response to first-line antipsychotics; associated with increased health burden and longer hospital stays.
Clozapine is an effective alternative for people resistant to first-line antipsychotic drugs; associated with reduced mortality, healthcare costs and functional outcomes. Its use, however, is restricted to non-responders due to the risk of side effects and the need for blood monitoring.
There is currently no way to identify non-responders: as a result non-responders are identified by protracted, empirical treatment trials, which leads to delay in starting effective treatment, disease morbidity and increased costs.
The average delay before treating the patient with an alternative antipsychotic drug, such as clozapine, is over 4 years. There is thus a clinical need for a biomarker to identify non-responders early to guide treatment choice.
Pharma companies are also seeking to develop better second-line treatments, but it is challenging to identify and enrol suitable study participants into clinical trials.
Benefits
The technology allows non-responders to be identified at the time of first prescription. Fast-tracking treatment-resistant patients to second line treatment has a potential healthcare cost saving of $4,200 per patient.
The technology can also be used to identify non-responders for inclusion in clinical trials of novel treatments or for monitoring purposes in clinical trials. For the pharmaceutical industry, the stratification method has the potential to become analogous to Beta-amyloid/Tau PET imaging for dementia.
Opportunity
The technology is protected by a European and a US patent application and is available for licensing. Suitable commercial partners may adopt it for service-based patient analyses for different applications.
The Science
The technology uses maps of PET brain images displaying the uptake of the [18F]FDOPA radiotracer and process them with a trained artificial intelligence, which returns a prediction for the patient treatment response.
Figure: The figure shows a flowchart of the automated method for predicting patient response to antipsychotic drugs.
Patent Status
Further Information
Veronese, M. et al., (2021) "A potential biomarker for treatment stratification in psychosis: evaluation of an [18F] FDOPA PET imaging approach". Neuropsychopharmacology 46, 1122–1132.. doi:10.1038/s41386-020-00866-7