AI can now detect cancer from photos

Scientists from Mass General Brigham developed FaceAge — an algorithm that analyzes facial photos and determines a person's biological age. This helps predict survival chances and choose optimal treatment.

The study, published in The Lancet Digital Health, showed that cancer patients' FaceAge was on average five years higher than their actual age. The higher this indicator, the worse the prognosis. This is especially noticeable in patients over 85 years old, whose survival chances significantly decrease. FaceAge also proved more accurate than doctors in estimating the life expectancy of patients receiving palliative care. When specialists made predictions without AI assistance, their results were only slightly better than random guessing. But with FaceAge data, the accuracy of assessments significantly improved.

The algorithm was trained on 58,851 images of healthy individuals and then tested on 6,196 photographs of cancer patients. The results showed that cancer patients appear significantly older, and their biological age directly affects treatment prognosis. Scientists are already testing FaceAge for predicting other diseases and assessing overall health. In the future, they plan to expand the research, study its application at different stages of cancer, and check how FaceAge indicators change over time.

This technology could become an important biomarker that helps doctors make more accurate treatment decisions. Developers hope that FaceAge will be used in clinical practice and assist in early disease diagnosis.

Source: TechCrunch