MEC ImpactReport 2023 flipbook - Flipbook - Page 12
Research
Predicting treatment outcomes with AI
Dr Konstantinos Balaskas, director, Moorfields Image Reading Centre, was granted a
springboard award by the charity to pilot a machine learning tool that could help clinicians
better predict how patients with age-related macular degeneration (AMD) would respond
to different treatment options.
What is macular degeneration?
Macular degeneration is a
common eye condition where
cells in the middle of the retina
(the light-sensitive layer at the
back of your eye) die or are
damaged, making it difficult
to see fine details clearly.
The most common form of
macular disease is age-related
macular degeneration (AMD),
which generally affects people
over 50.
the patient (such as their
age or their current vision),
imaging tests (such as Optical
Coherence Tomography
scans) and the genetic makeup of each patient all play a
role. However, we don’t yet
understand how to bring
all this complex information
together to predict which
treatment option an individual
patient will respond best to.
AMD is the main cause of
vision loss in the UK, and
although there are effective
treatments in the form of eye
injections, there are many
factors affecting how well
patients respond to these
treatments. For example,
clinical information about
Dr Balaskas used new
advances in artificial
intelligence (AI) and machine
learning to help make
sense of all this data and
support clinicians in making
their decisions. His team
collected clinical and imaging
information from Moorfields
patients with neovascular
AMD (nAMD), and also did
genetic testing on each
patient. All this data was then
passed to a new machine
learning tool they’d developed,
which predicted how patients
might respond to different
treatment options.
This project was a pilot, and it helped show that machine learning can integrate and
interpret many different forms of patient data. This could lead to exciting new insights into
how diseases develop and allow us to personalise patient care based on an AI’s predictions
of how patients are likely to respond to treatment.
Dr Konstantinos Balaskas, Director, Moorfields Image Reading Centre
What next?
The research team behind this project are now looking to develop a comprehensive artificial
intelligence pipeline. They plan for this pipeline to help diagnose a far wider range of
eye conditions beyond nAMD (with a focus on commonly occurring eye conditions) and
ultimately allow doctors to personalise treatment so it’s tailored to the needs of each
individual patient.