AI analyzes facial aging to predict cancer survival, revealing how treatment accelerates biological age

Face Aging Rate Emerges as Powerful Predictor in Cancer Care

A groundbreaking study published in Nature Communications introduces Face Aging Rate as a dynamic biomarker that quantifies how cancer and its treatment accelerate biological age 1

1

. Researchers at Mass General Brigham analyzed 2,276 cancer patients undergoing radiation therapy between 2012 and 2023, discovering that facial photographs taken at different time points reveal critical information about survival outcomes 3

3

. The median results showed patients' facial aging outpaced their chronological aging by 40%, with higher rates consistently linked to decreased survival probability 4

4

.

The artificial intelligence tool builds on FaceAge, a deep learning model trained on more than 40 million facial images that estimates biological age from a single photograph 3

3

. While previous research established that cancer patients typically appear five years older than their chronological age, this new approach measures the rate of change between two facial photographs, divided by the time interval between them 1

1

. A Face Aging Rate value above 1 indicates accelerated aging, while values below 1 suggest decelerated aging.

AI Tool Demonstrates Strongest Predictive Power Over Extended Intervals

The study revealed striking mortality risk patterns based on photo timing intervals. Among patients whose facial photographs were taken within one year, approximately 20% showed faces aging at least 20 times faster than expected, experiencing a 22% higher all-cause mortality rate 2

2

. For those with photos taken one to two years apart, the death rate jumped 54% higher in the accelerated aging group when dichotomized at 10 times faster than expected 2

2

. The effect proved strongest when intervals exceeded two years, where researchers simply divided cohorts at a Face Aging Rate of 1, comparing those aging faster than expected to those aging more slowly 2

2

.

Raymond Mak, co-senior author and radiation oncologist at Mass General Brigham Cancer Institute, emphasized the clinical implications: "Deriving a Face Aging Rate from multiple, routine facial photographs allows for near real-time tracking of an individual's health. 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" 3

3

.

Non-Invasive Biomarker Offers Advantages Over Single Time-Point Measurements

The research demonstrates that Face Aging Rate provides superior prognostic insights compared to FaceAge Deviation alone, which estimates how biologically old or young a patient looks in a single photograph relative to chronological age 4

4

. Patients with both high FaceAge Deviation and high Face Aging Rate values showed significantly poorer survival probabilities, but the dynamic measurement proved more reliable over longer intervals 3

3

. This aligns with broader medical evidence showing serial measurements outperform isolated readings, similar to how blood pressure variability better predicts cardiovascular outcomes than single values 1

1

.

The non-invasive biomarker approach leverages routine clinical workflow photographs already taken during radiation therapy courses, requiring no additional procedures or patient burden 3

3

. Hugo Aerts, director of the Artificial Intelligence in Medicine program at Mass General Brigham, noted the broader potential: "Tracking FaceAge over time from simple photos offers a non-invasive, cost-effective biomarker with potential to inform individuals of their health. We hope with continued study we can learn how FaceAge may provide prognostic information for patients with other chronic diseases and for healthy individuals" 4

4

.

Future Applications Extend Beyond Oncology

The research team has launched a web portal at faceage.bwh.harvard.edu where the general public can submit facial photographs for FaceAge assessment and participate in advancing this technology 3

3

. A complementary study published in JNCI: Journal of the National Cancer Institute tested FaceAge on more than 24,500 cancer patients over age 60 receiving radiation therapy, finding that 65% had FaceAge estimates older than their chronological age 3

3

. Those with estimates 10 or more years older faced significantly worse survival outcomes, while those within five years showed better outcomes.

The study cohort included patients with varying cancer types, with approximately 63% presenting metastatic disease at baseline 2

2

. Ongoing prospective trials will investigate Face Aging Rate across different cancers and chronic diseases, potentially establishing facial photographs as standard prognostic tools in multiple medical specialties. The integration of Face Aging Rate with baseline FaceAge Deviation could provide increasingly nuanced measures of evolving health status, enabling earlier interventions before disease manifestation becomes clinically apparent 1

1

.