AI and Biomarkers: How Selfies Can Aid Doctors' Decisions

Introduction

Researchers are investigating the potential of using artificial intelligence (AI) and facial images to assist clinicians in making informed decisions about patient care. This innovative approach aims to assess a patient's biological age, which could play a critical role in treatment decisions.

The FaceAge Algorithm

An upcoming presentation at the 2026 HIMSS Global Health Conference & Exposition in Las Vegas will showcase findings from clinical trials involving a deep learning algorithm known as FaceAge. This algorithm analyzes facial images to predict an individual's biological age and survival factors.

Dr. Raymond Mak, a radiation oncologist at Brigham and Women’s Hospital in Boston and a researcher with Mass General Brigham's AI in Medicine Program, explains that facial analysis may provide a more objective measure than a doctor's visual intuition alone. "What we think is that there is underlying biology, captured in your facial tissues – the skin, the muscles, the blood flow," he stated.

Testing Doctors' Predictions

Traditionally, healthcare has assessed patient biological age through various methods, including genomics, blood tests, and imaging. Researchers are now exploring how AI, combined with simple facial photographs, could enhance these assessments. Dr. Mak noted, "We demonstrated a proof of principle that's a novel prognostic factor in cancer patients, and it can be illustrated as a clinical use case."

Age is a significant factor in medical decision-making, and some physicians excel at estimating their patients' biological ages. Dr. Mak reported that a few doctors achieved accuracy rates of approximately 80% in their assessments, indicating they are picking up on subtle visual cues.

Over the past year, the researchers trained FaceAge using more than 58,000 images of healthy individuals and 6,196 cancer patients. In comparative studies involving palliative care patient data and known outcomes, the algorithm outperformed clinicians in predicting short-term life expectancy, specifically whether patients had less than or more than six months to live. "When we asked doctors to guess that, just based on the photograph – the eyeball tests – they're only slightly better than a coin flip on average," Mak explained.

While doctors' performance improved by about 10% when provided with standard clinical information, it significantly increased when they received data from FaceAge, allowing them to predict correctly seven or eight times out of ten.

The forthcoming research report, which Dr. Mak indicated will be published soon in Nature, revealed that cancer patients typically had a FaceAge five years older than their actual age, with older facial analyses correlating to poorer survival rates.

Potential Use Cases and Ethical Considerations

FaceAge, along with a second algorithm in development called FaceSurvival, captures different aspects of an individual's facial health. When integrated into prognostic models, these algorithms provide additional predictive power. Dr. Mak noted that while facial images are readily available and inexpensive to obtain, future research may also incorporate hand data or medical imaging, such as CT scans, provided that longitudinal patient images are accessible for analysis.

Despite the potential benefits, these tools face criticism. While the technology offers a non-invasive method to predict biological health, it raises ethical concerns, particularly regarding the possibility of AI predictions being used to deny life-saving treatments.

Dr. Mak emphasized that the goal is not to deny treatment but to enhance clinical decision-making. "From the very beginning of medical school, we're taught to look at a patient and document whether the patient looks older than stated age, or the patient looks younger than stated age," he explained. These assessments are crucial for making significant medical decisions, such as determining a patient's suitability for major surgeries or intensive cancer treatments.

In some cases, Dr. Mak has observed that patients undergoing chemotherapy can appear to age dramatically within weeks. He noted, "But that's not everybody." The data generated by FaceAge could help oncology patients with genetic markers navigate the uncertainties surrounding their medical decisions.

Ultimately, FaceAge and FaceSurvival may assist doctors in answering critical questions about patient prognosis and treatment options. However, some patients may prefer not to know their biological age for psychological reasons or fear that the data could affect their health plan coverage.

Conclusion

While the FaceAge algorithm has shown efficacy despite factors like cosmetic surgery, researchers aim to investigate whether weight loss or facial injuries could impact its accuracy. Collaborations with surgeons worldwide are planned for future studies exploring the applications of FaceAge analysis in various health assessments, including cardiovascular and neurological specialties.

The presentation titled "FaceAge: Using Artificial Intelligence to Decode Biological Age with a Selfie" is scheduled for Wednesday, March 11, from 2-2:30 p.m. in Level 5, Palazzo D in the Venetian, at HIMSS26 in Las Vegas.

Source: Global With AI and biomarkers, selfies could support doctors' decision-making - HealthcareITNews