A new smartphone app designed by scientists in the USA and Kenya makes it possible to measure hemoglobin levels in blood without actually drawing blood. A photograph of the inner eyelid is all that is needed for artificial intelligence to come up with an accurate measurement. Recently scientific magazine Optica published an article on this innovative research.
The intelligent software and the smartphone app were developed by a team of researchers at Purdue University and Indianapolis University (USA) along with scientists from the Vanderbilt University School of Medicine in Nashville, and the Moi University School of Medicine in Eldoret (Kenya). The objective of the researches was to come up with a blood testing method that can be used in places where laboratories may not always be available.
Research leader and main author is Young Kim, professor in Biomedical Engineering at Purdue University. He sees the COVID-19 pandemic as a driving force behind technological advances in the field of e-health. “Our new mobile health solution may open the door to new testing methods for diagnosing anemia, but also kidney failure, bleeding and hematological disorders such as sickle cell anemia.”
Using software, scientists have expanded the functionality of ordinary smartphone cameras in order to transform the phone into a reliable portable image scanner that can provide accurate hemoglobin measurements without the need for a blood draw or a sophisticated laboratory. The new technology can be very beneficial in the developing world, professor Kim states. “Diseases like sickle cell anemia pose a great challenge in developing countries such as Kenya.”
The American-Kenyan team designed software that utilizes super-resolution spectroscopy. Spectroscopic analysis is a common method to measure hemoglobin levels, but it requires large and expensive equipment. The new software is able to translate the low-res images from a smartphone camera into hi-res images.
All that is necessary are photographs of the inner eyelid. Not only is this an easily accessible location, but it is also not affected by skin colour and looks the same in every human, thus bypassing the need for calibrating the software for each individual. The patient merely needs to pull the eyelid down, and have the inside of his eyelid photographed. Afterwards, the super-resolution algorithm kicks in, and using a computer the images are compared to an image database, leading to an analysis. The app also comes with enhanced image stabilisers and flashlight synchronisation so that images will always be uniform and consistent.
A small scale clinical trial has taken place in Kenya, where 153 volunteers were included that had to call in at the hospital for a blood draw. Of 138 of them, the data was fed to the learning algorithm, and then the app was tested on the remaining 15 persons. It turned out that the app findings were between 90 and 95% consistent with the data from conventional analysis of blood samples. A larger clinical trial is currently underway in Indianapolis, where focus lies on monitoring hemoglobin levels in cancer patients. And further trials are prepared in Rwanda and India, to test how well the app copes performs with diagnosing sickle cell anemia and nutritional deficiencies.