
While it’s only relatively recently that AI became a household term, the technology is already having a major impact on society. But perhaps AI’s greatest impact has been in the medical field, specifically in the areas of diagnostics and devices.
Machine learning and other AI-based technologies have enabled researchers to gather data and undertake studies that are yielding an explosion of information on conditions and diseases of all kinds, as well as providing guidance on ways to prevent or treat them. Such information was either impossible to gather in the pre-AI era – or was so cumbersome for humans and existing technology to review that any conclusions would have taken years to draw.
One of the forces fueling the development and use of AI in medicine is the growing bridge between universities and other research institutions with the commercial sector. Many of today’s AI capabilities for medicine and clinical sciences are the results of decades of computational and bioinformatics research that have made their way to industry via translational science programs. Such translational cooperation is vital to meeting the needs of a growing and aging population and advancing healthcare in general. Thanks to that synergy, novel medical decision support systems can harness AI to integrate multiple sources of data to provide better diagnostics and treatment planning.
Today’s advances in AI and machine learning are fueling a digital revolution in pathology and radiology, enabling researchers to go far beyond what was possible just five to ten years ago. Here’s a glimpse at some of the most promising advancements in cancer treatment heading into 2025:
AI in Medical Imaging for Cancer
One of the key aspects of diagnostics is being able to identify anomalies, and advanced analysis of images using machine learning and AI algorithms are increasingly helping to detect patterns that can make that process faster and far more accurate. There are thousands of publicly available datasets for images of all types, as well as many more proprietary datasets, all of which can help researchers better understand the image they are looking at. AI algorithms are being used to more effectively analyze CT scans, MRIs, and X-rays for abnormalities such as fractures, hemorrhages, and tumors, assisting medical staff in more quickly identifying what treatment is required, as well as assisting emergency medical staff in triaging serious cases. AI can also assist researchers in detecting and grading cancer from histopathological images, critically enabling them to catch cancer and other diseases at much earlier stages and allowing for earlier and more effective intervention.
Computational Pathology and Multi-Omics for Personalized Healthcare
Recent advances in computational pathology and -omics technologies have made it possible to generate complex, personalized health data on a large scale, improving the prediction of treatment outcomes. Images are only part of the puzzle. The field of Multiomics Molecular Pathology, which includes RNA sequencing, DNA methylation analysis, and protein expression, is rapidly advancing.
AI Elucidating the Relationship Between -Omics and Histology Data
However, practical and reliable methods for integrating multi-omics and image-based data remain underdeveloped. One exciting development is the emergence of an artificial intelligence approach to understanding the relationship between -omics and histology data, using the omics signature to predict outcomes for patients with only histological data. This method will enhance histology-based prognosis and allow us to predict the marginal value of omics data for each patient.
In one example from this quickly rising domain, my lab is currently pursuing research that addresses this issue by inferring connections between histology and omics data, thereby improving the prognosis for patients with only histological biopsies. Additionally, our approach enables the personalized prioritization of broad omics measurements, further enhancing its cost-effectiveness.
Going forward, these areas and others will continue to depend on cooperation between basic research and industry, bringing solutions from the lab to commercialized spaces. This needs to be a priority not just for researchers and entrepreneurs but for those funding the process, including donors to universities’ translational science programs. Ensuring the future of basic science—and that it reaches real-world applications–is just as important as venture capital in transforming modern medical care.
About Professor Yoni Savir
Prof. Yoni Savir is an expert in systems biology and AI for health applications at the Technion–Israel Institute of Technology and Zimin Institute for AI Solutions in Healthcare. He is a co-founder of several startups, and consults biotech companies.