MDC Scientists Create AI Tool for Personalized Cancer Therapy
Picture created by IDEOGRAM
Berlin-based researchers have created an intelligent tool that could transform how doctors choose the best cancer treatments for their patients, making personalized medicine more accessible than ever before.
With nearly fifty new cancer therapies approved annually, selecting the right treatment has become increasingly complex for both patients and physicians. Dr. Altuna Akalin from the Max Delbrück Center's Berlin Institute for Medical Systems Biology (MDC-BIMSB) has developed a solution: Flexynesis, an advanced toolkit that uses deep learning to analyze multiple types of medical data simultaneously.
According to information from the MDC press release, Flexynesis represents a significant leap forward in precision medicine. Unlike previous tools that were often inflexible or difficult to use, this new system can evaluate multi-omics data, processed texts, and medical images like CT or MRI scans all at once. "This approach helps doctors make better diagnoses, prognoses, and treatment strategies for their patients," explains Akalin, who leads the Bioinformatics and Omics Data Science technology platform.
The toolkit, published in Nature Communications with Dr. Bora Uyar as first author, addresses multiple medical questions simultaneously. It can identify cancer types, determine which medications work best, predict survival chances, and even locate the origin of metastases when unclear. The system is available through various platforms including PyPI, Docker, and Galaxy, making integration into existing workflows straightforward.
What sets Flexynesis apart is its user-friendly design. Healthcare professionals don't need deep learning expertise to benefit from its capabilities. "I hope it lowers barriers for hospitals and research groups to perform multimodal data integration without having AI experts on their team," Akalin notes.
While multi-omics data isn't yet routinely collected in German clinics, this is changing. The tool complements Akalin's previous creation, Onconaut, which uses clinical guidelines and biomarkers.