Seventy per cent of cancer patients report cognitive problems in concentration, learning and memory, slowing down of thinking, reasoning planning, etc, during and after treatment, such that their quality of life is impaired, making it difficult to carry out daily activities. Overall, this symptomatological picture is called “chemobrain”. It may disappear within months, but 30% of breast cancer survivors report that it persists even 10 years after the end of treatment.
To date, the possible mechanisms are unknown, and there is therefore no prevention and/or treatment strategy. AI4ChemoBrain intends to develop an individual risk predictive model of the onset of chemobrain, based on biological determinants (cognitive function, markers of systemic inflammation, markers of intestinal dysbiosis collected in mouse models of chemobrain).
The model will exploit the tools of machine learning and artificial intelligence and will be trained with historical data sets collected over the last 15 years from preclinical models of cognitive decline, and omics data available from IRET in the form of raw data, already validated through statistical analysis and scientific publication (in accordance with the FAIR principles – Findable Accessible Interoperable Reusable).
The project will then validate the predictive model with internal and external cohorts to ensure its robustness, effectiveness and translatability in the clinical environment.
The ML/AI model will be able to guide adjuvant therapies, including personalised ones, helping to improve the empowerment of cancer patients and reduce the economic burden of treatment side effects. This will also have a significant impact on several production chains, including pharmaceuticals, biotechnology, nutraceuticals, ICT (Information and Communication Technology) and the national health service.