Local researchers are leveraging AI to assess aging through facial photos, which could significantly impact how doctors evaluate cancer treatment options and patient life expectancy. The FaceAge algorithm estimates biological age and shows promise in surpassing clinician estimates in some scenarios. Despite its potential, the technology remains in early stages and will need more study before use in clinical settings.
In an intriguing twist on age estimation, researchers have embarked on using artificial intelligence to determine how well a patient is aging rather than simply relying on the chronological age from a birth certificate. This innovative approach, spearheaded by local scientists, involves scanning facial photos to glean insights into a person’s overall aging process. The aim? To better inform doctors about cancer treatment options and possibly gauge a patient’s life expectancy more accurately.
Dr. Hugo Aerts, who leads the Artificial Intelligence in Medicine program at Mass General Brigham, expressed enthusiasm about their findings. “Our study now is showing probably for the first time that we can really use AI to turn a selfie into a real biomarker source of aging,” he noted, hinting at the potential to revolutionize assessments of patient health through photo analysis.
The study, recently published in Lancet Digital Health, analyzed a staggering 58,851 images of apparently healthy individuals. They trained an AI program dubbed FaceAge to estimate biological ages of these subjects. Once this foundation was laid, the researchers shifted focus to a group of 6,196 cancer patients, examining photos taken at the outset of their treatment. The results were quite telling: on average, FaceAge assessed the cancer patients to be nearly five years older than their actual age, alluding to the physical effects of their illness.
Interestingly, FaceAge outperformed clinicians in predicting short-term life expectancy specifically for those in palliative care. Co-author Dr. Ray Mak, also from Mass General Brigham, emphasized the value of this technology saying, “We often have, as doctors, clinical intuition or gut reactions that a patient looks a lot older than they are on paper.” He underscored that FaceAge could bolster this clinical instinct when deciding upon treatment paths.
Mak recounted a striking case: an 86-year-old cancer patient who seemed much younger. Doctors were cautious when recommending aggressive treatments typically avoided for that age group, yet the therapy worked wonders, leading him to thrive at 90. Interestingly, FaceAge estimated the biological age of the man at ten years less than his chronological age, prompting doctors to revisit their usual assumptions.
The integration of FaceAge into clinical practice isn’t aimed at replacing doctors’ judgments or personal patient decisions—far from it. It’s more of a supplemental tool, akin to the importance of vital signs or lab results, which could give physicians more comprehensive data to tailor care plans effectively.
Moreover, when reacting to whether a patient might only have six months to live, doctors often performed slightly better than random chance. However, with FaceAge in play, they noted a 20% improvement in accuracy. This alone highlights the significant potential of AI in healthcare settings.
There are complexities involved, of course. Factors like stress and environmental conditions often lead to variations in how old someone appears. Take the famous 1930s photo “Migrant Mother,” for instance. Florence Owens Thompson was 32 at the time, yet FaceAge pegged her biological age at a whopping 46.
What exactly FaceAge analyzes to reach these conclusions isn’t thoroughly understood yet. The researchers suspect it might be looking at aspects like muscle mass or eye region features, but they admit, “it is a little bit like deciphering the black box algorithm.” They are also examining how elements like skin tone, cosmetic surgery, or even makeup might influence results.
Looking ahead, the study authors plan to broaden their research and welcome other scientists to contribute, all with the aim of refining FaceAge and laying out ethical guidelines for its use. Though still in its infancy, the prospects of FaceAge are expansive. “We look at FaceAge as just the beginning of the new field of facial health recognition AI,” Mak mentioned, hinting at exciting developments that could reshape healthcare practices in the years to come.
In summary, researchers are exploring the possibility of using AI, specifically the FaceAge algorithm, to assess biological age based on facial analysis. This approach could have significant implications for cancer treatment and patient longevity assessments, according to findings from a recent study. While promising, the technology still requires thorough examination and ethical considerations before any clinical application. The implications of FaceAge could burst forth a new realm of understanding in medical diagnostics and patient care.
Original Source: www.wbur.org
