AI tool measuring ‘face age’ predicts survival in older patients with lung cancer | Managed Healthcare Executive
Key Takeaways
Deep learning–estimated face age predicted overall survival, with each 10-year increase associated with a 39% higher mortality risk after multivariable adjustment, whereas chronological age lacked independent prognostic value.
Early mortality risk rose with higher face age, and a face age ≥85 years identified elevated 2-year mortality irrespective of actual age.
Spirometry-derived lung age was markedly older (median 98 years) yet minimally correlated with face age, and face age retained independent association with survival when modeled jointly.
Face–chronological age discordance tracked outcomes, consistent with cumulative biological aging; older-appearing phenotypes fared worse, while younger-appearing phenotypes had better survival.
Operational integration appears feasible because radiation oncology workflows routinely capture identification photographs, but prospective validation and bias assessment across diverse cohorts and confounders are required.
Filters & Sorting