AI Detects Cancer and Viral Infections with Nanoscale Precision

Researchers have developed an advanced artificial intelligence tool, AINU (AI of the NUcleus), that can distinguish between cancerous and normal cells, as well as detect the early stages of viral infections within cells. The findings, published in the journal Nature Machine Intelligence, represent a significant leap in diagnostic technology, offering the potential for earlier and more accurate disease detection. The research was conducted by teams from the Centre for Genomic Regulation (CRG), the University of the Basque Country (UPV/EHU), Donostia International Physics Center (DIPC), and the Fundación Biofisica Bizkaia (FBB) at the Biofisika Institute.

High-Resolution Imaging at the Nanoscale

AINU leverages a cutting-edge microscopy technique called STORM, which captures high-resolution images of cells at nanoscale precision. A nanometer is one-billionth of a meter—about 5,000 times smaller than the width of a human hair. The AI analyzes these images to detect structural changes inside cells, as small as 20 nanometers, which are too subtle for traditional methods to identify.

Transforming Early Disease Detection

This nanoscale resolution allows AINU to detect changes in cellular structures soon after they occur. According to ICREA Research Professor Pia Cosma, a co-corresponding author of the study, “The resolution of these images is powerful enough for our AI to recognize specific patterns and differences with remarkable accuracy, including changes in how DNA is arranged inside cells, helping spot alterations very soon after they occur.” This capability could enable doctors to monitor diseases more effectively, personalize treatments, and ultimately improve patient outcomes.

How AINU Works: A New Era of Medical Imaging

AINU is built on a convolutional neural network, a type of AI specialized in visual data analysis. This technology is already used in various applications, from facial recognition to self-driving cars, and now, in medical imaging. The AI was trained using nanoscale-resolution images of cell nuclei in different states, learning to recognize specific patterns that differentiate cancer cells from normal ones.

For instance, cancer cells often exhibit distinct changes in their nuclear structure, such as alterations in DNA organization or enzyme distribution. After extensive training, AINU can now analyze new images and classify cell nuclei as cancerous or normal based on these microscopic features.

Detecting Viral Infections at the Molecular Level

One of AINU's most groundbreaking capabilities is its ability to detect viral infections shortly after they begin. The AI can identify subtle changes in the nucleus of a cell infected by the herpes simplex virus type-1, as early as one hour post-infection. This early detection is possible because the AI can detect minute differences in how DNA is packed within the nucleus—a change that occurs as the virus begins to alter the cell’s structure.

“Our method can detect cells that have been infected by a virus very soon after the infection starts,” explains Ignacio Arganda-Carreras, co-corresponding author of the study. This early detection could revolutionize how viral infections are diagnosed and treated, enabling quicker, more accurate responses in clinical settings.

Potential for Clinical Application and Challenges

While AINU shows immense potential, significant challenges remain before it can be used in clinical practice. STORM imaging, required by the AI, is currently limited to specialized biomedical research labs due to the need for advanced equipment and technical expertise. Additionally, STORM typically analyzes only a few cells at a time, which limits its efficiency in clinical diagnostics where speed and volume are crucial.

However, as Dr. Cosma notes, advances in STORM imaging technology may soon make these techniques more accessible, potentially bringing them into clinical settings. The researchers are optimistic about conducting preclinical experiments in the near future.

Accelerating Stem Cell Research

Beyond diagnostics, AINU also shows promise in accelerating stem cell research. The AI can identify pluripotent stem cells—cells that can develop into any type of cell in the body—with high precision. This capability could enhance the safety and effectiveness of stem cell therapies, which are crucial for regenerative medicine.

“Current methods to detect high-quality stem cells rely on animal testing,” says Davide Carnevali, first author of the research. “However, all our AI model needs to work is a sample that is stained with specific markers that highlight key nuclear features.” This approach could significantly reduce the need for animal testing, speeding up research while contributing to more ethical scientific practices.