- Scientists have created an algorithm that can accurately detect anaemia with a sensitivity of greater than 70 percent.
- The inner eyelid photographs captured by a smartphone can be used to assess the haemoglobin concentration in the blood.
- Those who live in distant areas could benefit from the use of a smartphone app to screen for anaemia.
Anemia affects more than 5 percent of the population in the United States and around 25 percent of the global population, according to the World Health Organization. The symptoms of this medical illness, which is characterised by a low haemoglobin concentration in the blood, include fatigue, dizziness, headaches, shortness of breath, and difficulties concentrating.
Severe anaemia is frequently caused by starvation, parasite infections, or underlying illnesses, among other things. It is a substantial cause of death and morbidity, particularly in vulnerable populations such as children, the elderly, and those suffering from chronic conditions.
According to a new study published in the open access journal PLOS ONE, taking a picture of a person’s inside lower eyelid with a smartphone and utilising that picture to predict anaemia is approximately 72 percent accurate.
In rural areas where people lack access to healthcare, anaemia is more common than in urban areas because healthcare practitioners often identify it with a full blood count performed with sensitive lab equipment.
As the researchers state, there is a need for a noninvasive, affordable, and easily accessible point-of-care technology that can detect anaemia in patients. In an ideal world, the tool would make use of already-existing, widely available technology.
Smartphone cameras are being used as detectors
In a two-phase investigation, the researchers investigated the feasibility of employing a smartphone camera to aid in the identification of anaemia. With a smartphone, researchers were able to capture photos of the inner lower eyelids of 142 patients who were treated in an emergency department during the first phase.
The inner lower eyelid, also known as the palpebral conjunctiva, was chosen by the researchers for study because it possesses the following distinct characteristics:
- When it comes to photographing, it is really accessible.
- Blood vessels and the conjunctival surface do not have any colours that are in competition with one another.
- The distance between the surface of the skin and the blood vessels is quite small.
- This area’s blood flow is not greatly affected by temperature or other environmental conditions, according to the study.
They were able to construct an algorithm that maximises colour resolution and a predictive model by comparing the skin and whites of the eyes to haemoglobin levels by zooming in on a small section in each photo.
The second phase involves putting the algorithm through its paces on photos taken by smartphones of 202 different emergency department patients. According to the data, the model was 72.6 percent accurate in predicting anaemia in the study participants. With a greater degree of precision, it could predict severe anaemia that would demand a blood transfusion, with an accuracy of between 86 and 94.4 percent.
Doctor Selim Suner of Brown University and Rhode Island Hospital, who was the study’s lead author, noted that after being diagnosed with anaemia, people only require iron supplements, which are inexpensive and simple to use. “The most difficult thing is making the diagnosis,” Suner explained.
Doctor Girish Nadkarni, clinical director of the Hasso Plattner Institute for Digital Health at Mount Sinai Health System, concurs with this viewpoint. “Using a smartphone to screen for anaemia is useful due to the decentralised nature of the screening — removing the need to draw blood — as well as the time and effort savings that this entails,” he explained.
The study’s advantages and disadvantages are discussed
According to the findings of the study, flash photography was not required to provide adequate photos for the purpose of anaemia detection. “RAW photographs give data straight from the camera sensor, without the customary processing and compression that occurs with standard formats, such as JPEG,” the authors write in addition.
The researchers identified a number of potential constraints, including varying image quality. Although it is possible that this was caused by the subject retracting their eyelid while the image was being recorded, this is not a given. Additionally, the lighting was not uniform, and it is unknown whether or not the different levels of brightness had an impact on the image quality.
Medical applications for cellphones in the future
A total of 36 percent of the world’s population utilised cellphones as of 2019. While rich individuals are more likely to own smartphones, evidence suggests that the use of mobile devices in lower socioeconomic zones is increasing gradually over the world.
“These findings pave the way for the creation of a smartphone application that can not only gather images but also evaluate the materials included within them in order to estimate haemoglobin concentration in real time,” the authors write.
“This is an especially enticing opportunity for poorer countries, which may have scant, unsophisticated, and poorly distributed medical systems but are well-connected by existing telecommunication networks.”
The design of a user interface that makes it simple for the general public to take a proper shot — one in which the lighting, focus, and area of interest are all maximised — will be the focus of future research. They also state that imaging technologies and future algorithm innovations that are validated by model prediction are required.
According to Dr. Suner, the findings of this study demonstrate that utilising a smartphone to forecast anaemia is a feasible notion. In the long run, this research and those that follow it could have a favourable impact on vast populations, thereby improving health around the world.