Researchers led by Hugo Aerts have developed an innovative artificial intelligence (AI) model capable of predicting a person’s biological age using just a selfie. This innovative technology has the potential to transform the way physicians evaluate patients and determine their course of treatment, especially in cancer care.
Our model was trained on a very large dataset. This new dataset includes more than 58,000 photos of people over the age of 60 collected from IMDb and Wikipedia. The ultimate aim is to better understand an individual’s biological condition and the process of aging itself. This assessment provides information that reaches well beyond their chronological age.
In real-world clinical environments, many factors will ultimately dictate how the AI model’s predictions inform treatment plans. One instance would be informing physicians of the necessary aggressiveness of a treatment appropriate to the patient. This guidance is often rooted in the patient’s biological age. It’s very subjective. Aerts stressed that the evaluation is usually quite subjective. It doesn’t have to be limited to today’s releases; it can have a profound impact on countless future decisions in medical practice.
The AI model was able to uncover some pretty compelling findings. For example, it found that cancer patients typically look an average of five years older than their real age. This difference could alert clinicians to patients’ general health and ability to undergo an intervention.
Meaningful ground truth Despite the model’s potential, a number of other factors can undercut its predictions. Factors like cosmetic enhancement, behavioral intervention, or social media filters can all heavily bias the outcome. That’s why even as these predictions are released, Aerts stresses the need for clinician judgment to help interpret them. This absolutely requires our focus,” Aerts underscored. Examples of patient-centric development So the key issue we face is: “How do we make sure that these technologies end up only helping the patient?”
Under current practices, physicians utilize subjective measures to evaluate patients’ health conditions. Nicola White emphasised that “when you’re dealing with people it’s a whole different game to dealing with metrics and statistics. This announcement highlights the importance of deliberation and caution as AI technologies are brought into the real world of patient care.
Skyler Ware, a freelance science journalist with a Ph.D. in chemistry from Caltech, is optimistic about the new AI model. He argues that we need a lot more validation before using it to guide treatment decisions. In particular, validation data needs to prove that implementing the model leads to better treatments or quality of life for patients.
As Aerts and his colleagues are quick to point out, introducing AI into healthcare isn’t as simple as creating the technology and letting it go to work. He further stressed the importance of attention, not just money, to ensure these new technologies are a boon for patients. This was notable because he stressed their commitment to the responsible use of AI.
As this powerful AI model continues to evolve, its influence on the medical field may be profound. Specifically, doctors can access more comprehensive information to understand patients’ biological ages. By understanding what’s causing their specific issue, it can allow them to treat things more precisely.
