FaceAge, an artificial intelligence (AI) tool introduced in The Lancet Digital Health magazine, has the ability to influence decisions in cancer treatment.
This tool uses in-depth algorithms, trained on tens of thousands of portraits, it evaluates cancer patients as having an average age biologically older than healthy people. The studys authors say it can help doctors decide who can safely withstand harsh treatments and who can be healthier with a gentler approach.
"We propose that FaceAge could be used as a Biodegree in cancer care to date patients' biological ages and help doctors make these difficult decisions," senior author Raymond Mak, an oncologist at Mass Brigham Health, a health system under Harvard Boston, told AFP.
For example, a patient is 75 years old, but has a biological age of only 65 and a person is 60 years old but has a biological age or 70, in this case, strong radiotherapy can be suitable for the former but dangerous for the later.
Similar reasoning can help make decisions about heart surgery, hip replacement, and lifelong care.
There is growing evidence that people age at different speeds, shaped by genetics, stress, exercise and habits such as smoking or drinking alcohol. While expensive genetic testing can reveal how DNA is eroded over time, FaceAge promises to provide detailed information just by using self- draw photos.
FaceAge's model is trained on 58,851 portraits of adults who are considered healthy over the age of 60, taken from public data.
It was then tested on 6,196 cancer patients treated in the US and the Netherlands, using photos taken immediately before radiotherapy. Patients with malicious diseases look on average 4.79 years older biologically than their actual age.
Among cancer patients, a higher FaceAge score predicts a poor chance of survival, even after considering actual age, gender, and type of tumor.
According to the authors, FaceAge did not detect any significant differences in race perspective. However, they are also training a second model with more than 20,000 patients.
They are also learning how factors such as makeup, cosmetic surgery, or changes in light in the room can deceive the system.
Ethnic debates are heating up. An AI tool that can read biological age from self- draw photos may be a benefit for clinicians, but it is also attractive for life insurance companies or employers who want to assess risks.
Researchers are planning to open a public FaceAge portal where people can upload their own photos to register to participate in a study to authenticate additional algorithms. Commercial versions targeting clinicians may follow, but only after further verification.
