Source: The American University in Cairo, written by Rani Chor; research by Yehea Ismail, Reda Abdelbaset (The American University in Cairo) (October 17, 2026). AI-generated summary by biochip.com, published . Not independently reviewed.
Key takeaways
- AUC’s biochip combines microelectronic sensors, microfluidics, and AI to screen samples for diseases including leukemia, liver disorders, hepatitis C, and malaria.
- The team reports about 91% accuracy for blood serum samples with AI, compared with about 80% without it.
- The platform remains a proof of concept and still needs reliability testing, expanded datasets, repeatability studies, and regulatory testing.
Researchers at The American University in Cairo have developed an artificial intelligence-enabled biochip that reads electrical signals from cells to screen for diseases including leukemia, liver disorders, hepatitis C, and malaria. The small platform combines microscopic sensors with a chamber for a biological sample, while machine-learning software looks for patterns that may distinguish healthy from diseased cells. The team reports about 91% accuracy for blood serum samples when it used artificial intelligence, compared with about 80% without it. The device is still a proof of concept, but its compact design could eventually support preliminary screening outside major laboratories.
Reading cells through electrical signals
Led by Yehea Ismail, chair of AUC’s Department of Electronics and Communications Engineering, and co-principal investigator Reda Abdelbaset, the project combines microelectronic sensors, microfluidics, and artificial intelligence. Microfluidics handles very small amounts of liquid through tiny channels, allowing a biological sample to interact with sensors on the chip.
Rather than looking for tumors in medical images, the biochip measures physical and electrical features of cells, including their size, movement, mass, and conductivity. As cells pass through microscopic structures, those features change the electrical signals the sensors detect. A computer model then analyzes the signals for patterns linked to disease.
The team has also investigated dielectrophoresis, a technique that uses electric fields to move or manipulate cells. The researchers’ work has involved tens of thousands of samples and collaborators at medical and research institutions, including Ain Shams University Hospital, Children’s Cancer Hospital 57357, Zewail City of Science, Technology and Innovation, Khalifa University, and Carnegie Mellon University Africa.
A preliminary screen, not a diagnosis
For blood serum samples, Abdelbaset reported about 91% accuracy with artificial intelligence, compared with about 80% without it. The team has also explored whether the system can distinguish healthy samples from those associated with different stages of metabolic dysfunction-associated steatotic liver disease, a liver condition linked to increased risk of liver cancer.
The researchers intend the system to flag samples for further evaluation, not replace a clinical diagnosis. Its models are designed to favor sensitivity over specificity: in practice, that means accepting more false alarms to reduce the chance of labeling a potentially diseased sample healthy. A positive or uncertain result would still need follow-up testing.
From laboratory prototype to accessible testing
AUC researchers envision a small, disposable chip that could be manufactured inexpensively and used nearer to patients. They estimate that mass production might bring the chip’s manufacturing cost to between 10 and 50 cents, with packaging bringing the total to about $1. The setup would also need a separate reader to process the electrical signals and send results to a computer, phone, or physician.
That cost and portability could matter in places where diagnostic equipment is concentrated in large hospitals. The team points to the need for timely screening for conditions such as malaria, which can worsen quickly, and liver disease. But the proposed price is an estimate, not the cost of a commercially available test.
Why This Matters
The biochip brings together cell sensing and machine-learning analysis in a format designed for small samples and potentially low-cost manufacture. Before it can serve patients, researchers still need to test its reliability across temperatures and manufacturing conditions, demonstrate repeatable results, expand the datasets, and complete regulatory testing. The current results establish a proof of concept, not a production-ready medical device.
The next step is to strengthen the evidence and engineer the system for consistent use outside the laboratory. A separate reader, expanded testing, and manufacturing development will all be part of determining whether the chip can become a practical screening tool.
Topics: Biochip · Cancer · Diagnostics & precision medicine
Start here: What is a biochip?