Researchers have united artificial intelligence (AI) and nanotechnology to exploit the inelastic scattering of light particles or “photons” to create a highly sensitive and rapid disease detection system on a chip.
The team behind this “nano-finger” has applied it to heart attack detection, driven by the urgent need to save lives from the condition. They also adapted the system to the detection of liver cancer to demonstrate its capacity as an early warning platform beyond just heart attacks.
“Some biomarkers related to inflammation, metabolic changes, or abnormal protein levels can be found in both conditions. This system uses a multi-diagnostic approach to analyze different sets of indicators for each disease, allowing it to test for both liver cancer and heart attacks efficiently,” said Zihan Wang from the University of Southern California. “Also, both diseases are difficult to predict in their early stages, as heart attacks occur suddenly without clear warning signs, and liver cancer develops silently with no obvious symptoms.”
Not only do both conditions have high mortality rates, making early detection critical, but in both cases, early intervention can significantly reduce fatality rates, as timely treatment for a heart attack can save a patient’s life, and detecting liver cancer at an early stage greatly improves the chances of successful treatment.
“This system was designed to address these challenges by providing rapid, high-sensitivity detection before it is too late,” Wang said.
An early warning for heart attack
Heart attack cases are increasing annually, and its sudden onset and rapid progression make it one of the leading causes of death. In the U.S. alone, 1 in 4 deaths is the result of heart attack, with over 30.3 million adults at risk. However, traditional diagnostic methods fail to predict heart attacks before they occur, leading to the failure to administer potentially life-saving interventions.
For example, traditional blood tests face several limitations related to disease detection. Their low sensitivity means it is difficult to detect disease biomarkers at early stages, while long detection times, ranging from 10 minutes to an hour, can prevent timely intervention in critical cases.
Additionally, some blood tests may require multiple steps, such as sample collection, preparation, and analysis. This analysis depends heavily on medical professionals, which can result in variability in results and potential misdiagnoses. Finally, specialized equipment and test reagents can result in high costs, while reliance on laboratory facilities can limit accessibility to blood tests, especially in deprived regions.
“These limitations prevent blood tests from providing early warnings before disease onset, increasing the risk of patients missing the optimal treatment window. This issue is particularly critical in heart attack detection, where survival depends on immediate intervention,” said Zerui Liu, researcher at the University of Southern California and contributing scientist to the current study. “To address this, we developed a system capable of predicting heart attacks within just 10 seconds, enabling rapid emergency response.”
Why heart attacks and liver cancer?
Both heart attack section and cancer detection both suffer from similar limitations, underscoring the urgent need for an ultrafast early warning system for both — and beyond.
Thus, to demonstrate the system’s capacity to detect the biomarkers of diseases other than heart attacks, the team chose to apply it to the detection of liver cancer, too.
“These diseases were chosen deliberately. Heart attacks progress rapidly and have a high fatality rate. It is critical to predict a heart attack before it occurs to allow enough time for medical intervention. However, no existing methods effectively provide early warnings for heart attacks, making this a pressing issue to address,” Liu said. “Similarly, liver cancer shares key characteristics among cancers, including vague early symptoms and a high mortality rate. The selection of these two diseases highlights the platform’s ability to detect conditions where early diagnosis is crucial for saving lives.”
Wang also highlighted other similarities between these two diseases, pointing out that both liver cancer and heart attacks can be indicated by certain biomarkers — specific molecules in the blood that indicate disease.
Wang and colleagues turned to AI as the basis of a system capable of such rapid disease detection, arguing that nanotech-based, AI-powered detection systems are capable of being deployed in situ within communities and even households, providing immediate heart disease warnings without requiring a trained medical professional.
Detection time is reduced by the automation of data analysis, allowing for real-time diagnostics, while AI improves sensitivity, facilitating the identification of disease biomarkers at extremely low concentrations, which traditional methods might miss.
“Imagine our nanofinger platform as a highly sensitive micro-sensor that can capture disease-related molecules from a small blood sample. These nanofingers enhance weak molecular signals, making it possible to detect even tiny amounts of disease biomarkers. Once the sample is collected, our system analyzes it using Raman spectroscopy, a technique that reads molecular fingerprints,” Wang said. “The collected data is then processed by an AI-powered chip, which acts like a high-speed medical expert, recognizing patterns and determining if a patient is at risk of a heart attack or has another disease.”
Wang added that, unlike traditional computers that store and process data in separate parts, this chip uses special technology called memristors an electrical component that regulates the flow of current while remembering the charge that has passed through it, to do both in the same place.
“This makes it much faster and more energy-efficient, allowing it to work in small, portable devices — even in ambulances or remote areas where power and space are limited,” Wang continued. “During testing, we found that the system’s potential extends beyond heart attack and liver cancer detection. The flexibility of the nanofinger sensor design means that by modifying surface chemistry, the platform can be adapted to detect a wide range of diseases, opening doors for multi-disease diagnostics from a single chip.”
The low-power and compact nature of the memristor-based system-on-chip was even more promising than the team expected. In the future, this aspect of the system could allow the system to be integrated into wearable health monitoring devices.
“This was not part of the original design goal, but it now represents an exciting possibility for continuous, real-time health monitoring,” Wang concluded. “These surprises demonstrate that this system is not just a high-accuracy diagnostic tool but a scalable, adaptable, and energy-efficient platform that could transform point-of-care and at-home disease detection.”
Reference: Wei Wu, et al., Multi-Diseases Detection with Memristive System on Chip, Advanced Intelligent Systems (2025). DOI:10.1002/aisy.202400736
