AI-powered scan could detect anaemia without blood tests — Study

Researchers have developed an artificial intelligence-powered screening tool that can estimate a person’s red blood cell levels using short videos of blood vessels in the eye, potentially offering a needle-free way to detect anaemia.

The technology, described in a study published in npj Digital Medicine, identified low haemoglobin levels with about 83 per cent accuracy in a trial involving 224 participants.

Scientists say the method could serve as a screening tool in settings where laboratory testing is limited, although it is not yet ready to replace conventional blood tests.

Anaemia, which occurs when the body lacks enough healthy red blood cells or haemoglobin, affects millions of people worldwide and is usually diagnosed through laboratory analysis of blood samples.

The study, conducted at Sheba Medical Centre in Israel, involved healthy volunteers and patients with blood disorders, including some undergoing cancer treatment.

Researchers used a microscope camera with 50-times magnification to record 10-second videos of the white part of participants’ eyes. Specialised software processed the footage and converted it into detailed images of blood vessels.

An AI model known as VesselNet then analysed patterns in blood flow to predict haemoglobin levels and red blood cell counts. The results were compared with standard laboratory tests and correctly identified low haemoglobin levels in about 83 per cent of cases.

The researchers noted that the system performs similarly to existing non-invasive technologies such as the Pronto-7 device, which estimates haemoglobin levels by shining light through a fingernail.

However, unlike fingernail-based devices, which can be affected by skin pigmentation, the white of the eye contains very little pigment and appears relatively similar across different populations.

Commenting on the findings, Dr Christine Kiire, a consultant ophthalmologist at Oxford Eye Hospital and visiting researcher at University College London, said the technology could improve access to blood monitoring if it proves reliable and affordable.

Also Read: Alleged cyberstalking: Court adjourns Doris Ogala’s case over repeated absence

“If it is validated and made affordable, the system could make blood monitoring more accessible in resource-limited environments,” she said.

Dr Theodore Leng, an ophthalmologist and vitreoretinal surgeon at Stanford University who was not involved in the study, said the technique could be useful in situations where repeated blood draws are difficult, including cancer care, dialysis follow-up and paediatric settings.

The researchers cautioned that the technology remains in its early stages and currently measures only haemoglobin levels and red blood cell counts. They plan to conduct larger studies involving more diverse populations and explore whether future versions could also assess other blood components.

While more validation is needed, the findings highlight the potential of AI and medical imaging to expand access to non-invasive health screening, particularly in low-resource settings.