Face Photos Offer Window on Cancer Survival
An artificial intelligence tool could identify patients at increased risk of death after cancer treatment from their faces, research findings suggest.
The FaceAge algorithm captures dynamic changes in facial aging from photographs, offering a visible manifestation of the complex interplay between aging, cancer, and survival.
Patients with faces that aged rapidly between radiotherapy treatments according to the tool had a decreased chance of survival, according to the study in Nature Communications.
FaceAge assessments could therefore offer a dynamic, simple, accessible and non-invasive biomarker for cancer prognosis, which could be used to stratify risk and tailor treatment planning in oncology.
“Our study suggests that measuring FaceAge over time may refine personalized treatment planning, improve patient counseling, and help guide the frequency and intensity of follow-up in oncology,” said researcher Raymond Mak, PhD, from Mass General Brigham Cancer Institute.
Facial aging involves alterations in skin texture, loss of volume, and changes in bone structure and advances in deep learning have enabled age to be accurately estimated from visual biomarkers.
To investigate its prognostic value in cancer treatment, Mak and team developed the Face Aging Rate (FAR) biomarker.
This is calculated as the change in FaceAge between two time points divided by the time interval between the two facial photographs that were used, to provide a dynamic measure of aging over time.
A FAR value above one indicates accelerated aging that is faster than expected while a FAR value below one suggests decelerated aging.
The team tested the prognostic value of FAR in 2279 patients with different types of cancer, who underwent at least two courses of radiotherapy at Brigham and Women’s hospital between 2012 and 2023.
Face photographs taken in routine care during different radiotherapy courses at distinct timepoints were associated with survival, they reported. Specifically, a higher FAR is associated with an increased risk of death over different timeframes in stratified analyses.
For photos taken with a short interval in between, of 10 to 365 days, the adjusted hazard ratio (aHR) was 1.25, while for those with a long interval of between 731 and 1460 days, the aHR was 1.65.
The aHR was not statistically significant for the mid interval of photos taken 366 to 730 days apart.
“The association between accelerated facial aging and increased mortality risk may reflect deeper systemic biological processes, such as cellular senescence, DNA damage, and reduced tissue repair; all of which are hallmarks of both aging and cancer progression,“ the authors noted.
They added: “As research progresses, FAR has the potential to become an integral component of comprehensive cancer care and a tool to monitor changes in health more broadly, embodying the future of personalized medicine where serial facial photographs can provide profound prognostic insights.”