Inteligência artificial para descobrir a próxima geração de nanopartículas personalizadas para a terapia do cancro da mama triplo-negativo.

The project is multi and interdisciplinary. It aims at developing new technology benefiting society and TNBC therapy, the most hard-to-treat BC type since these cancer cells lack the estrogen and progesterone receptors, as well as sufficient HER2 protein, to make hormone treatment or targeted HER2 therapies effective. Given the entrepreneurial mindset of the team, we will strive to translate findings to the clinic while further positioning Portugal at the forefront of innovative healthcare. Each task entails well-defined goals, milestones, deliverables towards the envisioned breakthroughs. KPIs were designed to monitor execution. For feasibility, we focus on preclinical validation and propose specific aims: Aim 1. implement active ML/AI that is traceable/retrainable/interpretable and interfaces human intuition. The ML/AI will suggest the next batch of meaningful experiments and formalize expert knowledge. It will compress a vast search space and allow the discovery of effective lipid NPs, minimize experimentation and promote research sustainability. Our approach presents benefits relative to design of experiments, which would require a larger fraction of assays to gauge the NP bioactivity landscape. The ML/AI will augment expert perception through interpretation pipelines and optimize conflicting objectives (e.g. cardiotoxicity). Aim 2. conduct screening assays for the NPs. We generate the dataset to train an initial ML/AI. This is obtained via synthesis/profiling of NPs with random compositions within an expert-defined search space. The iterative learning will be tuned to deliver at least one lipid NP tailored towards TNBC and minimize potential cardiotoxicity. Aim 3. understand the key physicochemical properties and structural motifs that drive efficacy and safety of the best NPs. In vitro safety, cellular uptake, survival, proliferation/migration and cell viability assays will be critical to understand key PK/PD, efficacy/toxicity profiles and motivate transition from in vitro to in vivo studies. Aim 4. perform in vivo studies to evaluate the NPs in TNBC model. Together with formulation development studies, these will enable an informed selection on NPs for advanced preclinical assays. Lipid NPs with good in vitro activity will be evaluated in vivo to scrutinize efficacy, liabilities and potential benefits for large scale in vivo testing. These studies are designed to demonstrate modulation of disease by the selected NPs. Aim 5. provide first rate training to junior staff and strengthen collaborations between chemists, biologists, data, formulation scientists and clinicians. A new breed of scientists will emerge displaying a unique skill-set that is relevant to next generation drug development in academia and industry. Aim 6. disseminate results to the scientific community, engage the general public, patient groups, stakeholders and policy makers to raise awareness towards ML/AI for healthcare, nanotechnology and TNBC as a global unmet medical need.