A new AI tool called MISO (Multi-modal Spatial Omics) can detect detailed cancer traits by analyzing data from tiny tissue samples, some as small as five human hairs in width. Created by researchers at the Perelman School of Medicine, University of Pennsylvania, MISO uses advanced technology to examine large amounts of data, helping doctors choose the best treatments for different types of cancer.
Key Discoveries with MISO:
- Bladder Cancer: Found special cells that improve responses to immunotherapy.
- Gastric Cancer: Differentiated between cancer cells and normal tissue.
- Colorectal Cancer: Identified sub-classes of malignant cells within a single tumor.
MISO also studied non-cancerous brain tissues, showing its versatility in analyzing various conditions.
How MISO Works:
MISO is part of a field called “spatial multi-omics,” which studies how genes, proteins, and other factors interact within tissue. It combines data from multiple sources like transcriptomics (gene activity) and proteomics (proteins) to give a full picture of cellular behavior.
One remarkable feature of MISO is its ability to analyze images where each pixel contains 20,000 to 30,000 data points. This is far more complex than the single gray-scale data point in typical MRI or CT scans.
Advancing AI in Medical Imaging:
MISO builds on previous AI tools like iSTAR, which focused on finding cancer-related patterns. Together, these tools could improve imaging and provide deeper insights into diseases.
The researchers aim to expand MISO’s capabilities to analyze multiple tissue samples at once and include data types like epigenetic marks, which control DNA activity. “MISO’s AI system will continue to learn, offering even greater insights in the future,” said Dr. Mingyao Li, the study’s senior author.
This innovation shows promise in making cancer diagnosis and treatment more accurate and personalized, potentially saving more lives.
Source: newswise
