image: Scientists have developed an innovative diagnostic technology that analyzes the full drying process of a blood droplet using machine learning. The process could transform medical and biological screening, allowing rapid analysis of blood and other biological fluids using only a single droplet and simple microscopic imaging, and bring cost-efficient, powerful diagnostics within reach to underserved communities and developing countries.
Credit: Anusuya Pal, The University of Tokyo, Japan
Not long ago, the idea of diagnosing a disease with a droplet of blood was considered a pipe dream. Today, this technology could soon become a reality.
A group of scientists led by researchers from the University of Tokyo developed an automated, high-throughput system that relies on imaging droplets of biofluids (such as blood, saliva and urine) for disease diagnosis in an attempt to reduce the number of consumables and equipment needed for biomedical testing. In the workflow, biofluid droplet images are analyzed by machine-learning algorithms to diagnose disease. Remarkably, the technology relies on the drying process of biofluid droplets to distinguish between normal and abnormal samples.
Current medical diagnostic tests typically require 5 milliliters to 10 milliliters of blood, necessitating a trip to the clinic or other phlebotomy service to draw blood with needles and tubes. Besides being painful, inconvenient and inefficient, blood draws are often a luxury of developed nations with modern health care infrastructure. By eliminating the need for phlebotomy services and other consumables, diagnostic tests could be implemented worldwide to improve disease diagnosis and cost efficiency.
“We set out to develop a simple, rapid and reliable approach to analyze what happens when a droplet of blood dries on a surface,” said Miho Yanagisawa, associate professor at the Graduate School of Arts and Sciences at the University of Tokyo. “Traditionally, researchers have focused only on the final pattern left after drying. In our study, we looked beyond that, observing the entire drying process in real time. By tracking how the droplet’s shape and internal structures evolve over time, we were able to uncover rich information about the fluid’s composition.”
By using machine learning, the team could “decode” the evolving patterns in drying blood droplets, allowing them to clearly distinguish between healthy blood and samples with abnormalities based solely on their drying behavior.
Importantly, this technique doesn’t require specialized equipment to make an accurate diagnosis. Images of drying blood samples are acquired using brightfield microscopy (transmitting white light through a specimen, which makes it appear dark against a bright background) and a common 4x objective lens, which magnifies samples four times. Images are acquired over time with a digital camera mounted on the microscope. The same workflow can also be used to analyze other bodily fluids, including saliva and urine, expanding the diagnostic capacity of the workflow without the need for additional equipment.
“The key takeaway is that every moment of the drying process holds valuable clues, not just the final pattern left behind. Each stage reveals how proteins, cells and other components move and reorganize within the droplet, capturing a dynamic ‘story’ of the sample’s internal state,” said Anusuya Pal, a postdoctoral research fellow in the Yanagisawa Lab and first author of the research paper.
By combining this time-evolving information with machine learning, the team can accurately identify subtle abnormalities in blood samples. “This approach opens up a new way of thinking about medical diagnostics, one that is simple, fast and low-cost, yet remarkably informative,” said Pal.
The current research establishes proof of concept for the team, demonstrating an effective workflow for detecting diseases such as diabetes, influenza, malaria and others, that has potential in the field. Ideally, the researchers hope to translate their methodology into a mobile and practical health-screening tool for use in developing countries.
“Such a tool could make health monitoring faster, more affordable and more accessible, especially in communities with limited access to laboratory testing. Ultimately, our goal is to bring laboratory-level insights to the point of care, enabling early detection and preventive health care for everyone,” said Amalesh Gope, assistant professor at Tezpur University in India and co-author of the study.
###
Research paper:
Anusuya Pal, Amalesh Gope, and Miho Yanagisawa, “From Droplet to Diagnosis: Spatio-Temporal Pattern Recognition in Drying Bio-fluids,” Advanced Intelligent Systems: November 16, 2025, DOI: 10.1002/aisy.202500550
Link: http://doi.org/10.1002/aisy.202500550
Funding:
UTOPIA (supported by AMED under Grant Number JP223fa627001) (M.Y.), Japan Society for Promotion of Science (JSPS), KAKENHI Grant No. 23KF0104 (M.Y. and A.P.), Japan Science and Technology Agency (JST) (FOREST, JPMJFR213Y) (M.Y.), and CREST, JPMJCR22E1 (M.Y.).
Links:
Graduate School of Arts and Sciences
https://www.c.u-tokyo.ac.jp/eng_site/
Yanagisawa Lab
https://sites.google.com/g.ecc.u-tokyo.ac.jp/yanagisawa-lab/home
About the University of Tokyo
The University of Tokyo is Japan's leading university and one of the world's top research universities. The vast research output of some 6,000 researchers is published in the world's top journals across the arts and sciences. Our vibrant student body of around 15,000 undergraduate and 15,000 graduate students includes over 5,000 international students. Find out more at https://www.u-tokyo.ac.jp/en/ or follow us on X (formerly Twitter) at @UTokyo_News_en
Journal
Advanced Intelligent Systems
Method of Research
Imaging analysis
Subject of Research
Human tissue samples
Article Title
From Droplet to Diagnosis: Spatio-Temporal Pattern Recognition in Drying Biofluids
Article Publication Date
16-Nov-2025