Proteomic-Based AI Score Predicts Tumor Stemness and Cancer Aggressiveness
Cancer diagnosis and treatment face growing challenges as cases increase worldwide and tumors become more complex. Artificial intelligence (AI) is proving to be a valuable asset in improving prediction and detection methods. A new AI tool developed by Brazilian and Polish researchers offers a promising approach by predicting tumor aggressiveness through protein analysis.
This machine learning model generates a stemness index that ranges from zero to one, where zero indicates low tumor aggressiveness and one indicates high aggressiveness. A higher index correlates with tumors that are more drug-resistant and prone to recurrence. These findings were published in the journal Cell Genomics.
What Is Tumor Stemness?
Tumor stemness reflects how closely cancer cells resemble pluripotent stem cells, which have the ability to transform into nearly any cell type in the body. As tumors advance, their cells lose resemblance to their tissue of origin and adopt an undifferentiated, self-renewing state. This stem-like characteristic contributes to tumor aggressiveness and challenges in treatment.
Development of the Protein Expression-Based Stemness Index (PROTsi)
The researchers built their model using data from the Clinical Proteomic Tumor Analysis Consortium (CPTAC), covering 11 cancer types including breast, ovarian, lung, kidney, uterine, brain, head and neck, colon, and pancreatic cancers. They analyzed over 1,300 tumor samples and integrated these with proteomic data from 207 pluripotent stem cells.
This integration allowed the team to identify specific proteins driving tumor aggressiveness, presenting potential targets for new therapies. PROTsi adds a protein-level perspective to previous stemness models based on RNA and DNA data, making it more applicable to clinical treatment development.
Clinical and Research Implications
- PROTsi effectively differentiates tumor cells from non-tumor cells and distinguishes between various tumor grades.
- It showed predictive value in cancers such as uterine, head and neck, pancreatic, adenocarcinoma, and pediatric brain tumors.
- Several proteins identified are already targeted by existing drugs, paving the way for repurposing and tailored therapies.
- The tool enhances understanding of tumor progression and mechanisms behind resistance to current treatments.
While PROTsi works better for some cancer types than others, the researchers are refining computational models to improve its accuracy and applicability. The data sets used are being made available for future research to expand this work.
Why This Matters to Healthcare Professionals
Understanding tumor stemness at the protein level offers a practical avenue to predict cancer behavior more precisely. This can inform treatment decisions, helping to personalize therapy plans and potentially improve patient outcomes. As resistance to drugs remains a major obstacle, targeting proteins linked to stemness may open new therapeutic options.
Healthcare professionals can benefit from staying informed on such AI-driven tools that integrate proteomic data with clinical insights. This approach represents a step toward more data-driven and individualized cancer care.
For those interested in expanding their knowledge on AI applications in healthcare, exploring specialized courses can be valuable. Resources like Complete AI Training’s healthcare-focused courses offer practical insights into AI tools and their clinical uses.
Additional Resources
Your membership also unlocks:
