Often, when a patient hears the prognosis of a doctor that he or she has cancer, it is a moment of heartbreak for the patient.
The problem that often arises in the late stages of cancer, especially the often-occurring is “How long do I have to live?”
Doctors usually rely on experience and medical tests to make the best educated guess. A series of treatments based on predictions Recommended.
However, researchers at the hospital they belong to Harvard Medical School We remove the speculation from this important prognosis.
They developed facean artificial intelligence system that analyzes photos A more accurate estimate of patient biologicals instead of age. For example, a healthy 75-year-old may have physiological properties of about 60 years old.
“On average, we found that cancer patients appear to be about five years older than their age and have a statistically higher face compared to a clinical cohort of cancer-free patients. It will be handled For conditions that are benign or pregnant,” the researchers wrote to them paperwhich of It has been published At Lancet Digital Health.
By properly assessing your body truth The prediction of age, survival is more accurate, which determines which treatment to be given to the patient.
“We showed when clinicians’ predictive survival performance improved. Face risk model predictions have been made The paper said.
Face, that Developed Using deep learning techniques, patients older than their actual age were shown to have a higher likelihood of achieving worse outcomes even after controlling for traditional clinical risk factors.
According to the authors, including Canadian and European researchers, “age was correlated with worse overall survival.”
For patients, Faith represents a future in which a single photograph can provide personalized insight into the health, risks and treatment decisions that may come with lab tests and medical scans.
For healthcare providers, The face can complement clinical judgment. This is especially important when treating critically ill patients.
this In many cases, narrow windows of survival are particularly relevant to cancers in which doctors force doctors to make difficult decisions about aggressive treatment based on their own prognosis.
AI is increasing It is being used In a healthcare setting, the key role that doctors and other caregivers play cannot be replaced, healthcare experts told PYMNT.
However, the AI tools are: important According to PYMNTS Intelligence Report, it complements patients with a seamless digital experience.Digital Healthcare Gap: Streamline Patient Travel. ”
How the face works
Faceage AI model trained We tested 6,200 cancer patients in the US and the Netherlands with almost 59,000 images of healthy people over the age of 60. Using a two-stage neural network system, the algorithm detects the faces of the photographs, extracts key features, and generates the estimated biological age.
This tool was better at predicting survival than looking at patient age among the three groups. People who are undergoing curative radiation therapy, patients with chest cancer, and those who are receiving palliative care for metastatic disease.
Cancer patients had an average face of 5 years older than their actual age, with a statistically significant gap.
Additionally, the survival predictions for terminal patients have been improved. When used with TeachH clinical model, a tool used to estimate the average life expectancy of patients receiving palliative radiation therapy, Faith increased the accuracy of the model for prediction, the paper states.
Doctors also performed better to predict 6 months of survival when face aidedAccording to the paper. this You may have major The impact on treatment decisions helps clinicians in near-terminal patients weigh the advantages and disadvantages of treatment.
Despite that promise, the system rises important Ethical considerations. Researchers acknowledged risks, including potential misuse by insurers and advertisers, and racial or socioeconomic bias in the model.
Although Faceage showed minimal bias among ethnic groups in its preliminary trials, researchers sought further verification using more diverse data sets and careful regulatory oversight, saying in their paper that “further assessment of performance bias in different populations is essential.”
Although Faceage is not yet ready for routine clinical use, its success illustrates the steps to integrate AI-based biomarkers into healthcare. It suggests that something as simple as a patient’s face may immediately hold a more accurate and humane key and personalization care.
from now onIn the paper, the researchers stated that further research is needed to “establish whether a larger group and further research “extends to patients with other diseases.”
Subscribe to our daily AI newsletter for all PYMNTS AI coverage.
