Researchers in Brazil and Poland have developed an AI-powered tool that predicts cancer aggressiveness by analysing protein expression – offering new insights into tumour behaviour.
As cancer cases rise globally, researchers are turning to artificial intelligence (AI) to help. A new machine learning tool developed by scientists at Poznan University of Medical Sciences and the University of São Paulo offers a new approach by using protein expression to predict the aggressiveness of tumours.
Their study, published in Cell Genomics, introduces PROTsi (Protein-based Stemness Index) a tool that was developed using data from the Clinical Proteomic Tumour Analysis Consortium (CPTAC). It generates a stemness index from zero to one, reflecting how closely a tumour resembles pluripotent stem cells. The closer the value is to one, the more aggressive and drug-resistant the tumour is likely to be.
How PROTsi works
The model was developed by analysing over 1,300 samples from 11 different cancer types, including breast, ovarian, lung, kidney, uterine, brain (paediatric and adult), head and neck, colon and pancreatic cancers. By comparing these tumour samples to 207 pluripotent stem cells, the researchers identified proteins that drive tumour aggressiveness and may serve as potential therapeutic targets.
“Many of these proteins are already targets of drugs available on the market for cancer patients and other diseases. They can be tested in future studies based on this identification. We arrived at them by associating the stemness phenotype with tumour aggressiveness,” explained Professor Tathiane Malta, from the Multiomics and Molecular Oncology Laboratory at the Ribeirão Preto Medical School of the University of São Paulo (FMRP-USP).
Strong validation and promising applications
The PROTsi tool showed strong predictive power during validation, distinguishing effectively between tumour and non-tumour samples; clearly separating stem and differentiated cells. The model performed especially well in certain cancers – such as uterine, head, neck, pancreatic and paediatric brain tumours.
“We sought to build a model that can be applied to any cancer, but we found that it works better for some than for others. We’re making a data source available for future work,” said Malta.
Toward better cancer treatment
The research highlights the broader goal of using AI in cancer treatment personalisation. By identifying stemness-driving proteins, PROTsi could lead to new general or tumour-specific therapies.
Renan Santos Simões, co-first author of the article, emphasised the collaborative nature of the work. “Science advances slowly, carefully, and is built by many hands. It’s gratifying to realise that we’re contributing to this process. That’s what motivates us: knowing that what we do today can make a real difference for patients, improving treatments and quality of life.”
What’s next?
Professor Malta and her team at USP are now working on additional computational models to enhance the predictive capabilities of PROTsi. The current version already marks a significant innovation in translating molecular data into real clinical insights.
This project reflects not only cutting-edge science but also a broader ambition to make cancer treatment more precise, targeted and effective.
Related topics
Analysis, Artificial Intelligence, Cancer research, Computational techniques, Drug Targets, Machine learning, Oncology, Precision Medicine, Protein Expression, Proteomics, Translational Science