A lay summary written by Nia Georgieva and reviewed by Dr Julie Simpson, Dr Scott Allen and a Dementia Lay Panel.
Background:
Alzheimer’s disease is the most common form of dementia. It progressively damages brain cells, causing memory loss and may eventually lead to disability and not being capable of carrying out simple daily life tasks. It is a public health concern, as being a complex disorder, which does not have a particular cure. The diagnosis for it involves expensive and, in some cases, invasive tests, which are presented only in specialised services. Thus, that make it almost impossible for some people to be examined and diagnosed easily. Therefore, a model that use photos from the back of the eye could be used as a non-invasive method. This would help to diagnose people with Alzheimer’s disease even in primary care and community settings.
Why is the study important?
The aim of the study is to develop more accessible non-invasive tool to routinely screen people for Alzheimer’s disease before more brain cells have been damaged. It would allow following the progression of already diagnosed patients.
What did the authors do and how did they do it?
For this research, the authors included 11 clinical studies from 4 countries (Hong Kong, China, Singapore, the UK and the USA). The authors used and analysed photographs of the back of the eye of 648 people with Alzheimer’s disease and 3240 without the disease. In the development of this model, four retinal photographs (2 from each eye) for every participant were used.
What are the results?
The authors report that this screening tool is able to accurately differentiate between people with Alzheimer’s disease and without, even in the presence of concomitant eye disease. As the eyes are directly connected through the optic nerve with the brain, this allows a non-invasive screening of the brain. Therefore, people would be successfully screened in routine eye tests in optometry and ophthalmology settings. This would allow to identify potential Alzheimer’s disease people before more serious events in the brain have taken place.
One disadvantage of this research could be that the data the authors used may not have diversity in order for this method to be developed as a screening tool. Thus, more diverse populations are needed in order to be ruled out as an appropriate screening tool for Alzheimer’s disease. Moreover, this screening method concentrated only on people with Alzheimer’s disease related dementia.
What do the findings mean going forward for people with the disease?
This is the first learning model in artificial intelligence and machine learning that can accurately identify Alzheimer’s disease only from photographs of the back of the eye. The simple and low-cost procedure could potentially be used to diagnose people with Alzheimer’s disease earlier in their life without going through expensive and invasive tests. People can be routinely examined in community-based settings by ophthalmologists and optometrists. The patients that have already been diagnosed with Alzheimer’s disease can be frequently screened and the disease progression followed, as it will not involve any invasion or side effects.
This study can be found at: https://www.thelancet.com/journals/landig/article/PIIS2589-7500(22)00169-8/fulltext
Paper title: A deep learning model for detection of Alzheimer’s disease based on retinal photographs: a retrospective, multicentre case-control study
Author list: Carol Y Cheung, An Ran Ran, Shujun Wang, Victor T T Chan, Kaiser Sham, Saima Hilal, Narayanaswamy Venketasubramanian, Ching-Yu Cheng, Charumathi Sabanayagam, Yih Chung Tham, Leopold Schmetterer, Gareth J McKay, Michael A Williams, Adrian Wong, Lisa W C Au, Zhihui Lu, Jason C Yam, Clement C Tham, John J Chen, Oana M Dumitrascu, Pheng-Ann Heng, Timothy C Y Kwok, Vincent C T Mok, Dan Milea, Christopher Li-Hsian Chen, Tien Yin Wong
Publication details including date of publication: Published Online September 30th 2022 in the Lancet Digit Health 2022; 4: e806–15