Boston researchers at Mass General Brigham say the thymus, a small triangle-shaped gland tucked behind the breastbone, might be doing a lot more for adults than medicine has given it credit for. By using artificial intelligence to re-read routine chest CT scans, the team linked a "healthier" thymus to lower risks of major diseases and stronger responses to cancer immunotherapy, challenging the long-held notion that the organ basically retires after childhood.
According to WCVB, the researchers reviewed routine CT scans from more than 2,500 participants in the Framingham Heart Study and roughly 25,000 adults in a national lung cancer screening program. Their analyses suggest that thymic features visible on everyday scans could help predict risk for cancer, cardiovascular disease, and other major outcomes. The coverage describes the finding as early but potentially practical, a possible window into immune fitness in middle-aged and older adults.
AI turned routine CT scans into a thymus health measure
The team built a deep-learning model that automatically measures thymic size, shape, and internal structure on chest CTs, then rolls those characteristics into a thymic health score. ASCO Post reports that the score was applied to nearly 3,500 real-world cancer patients treated with immune checkpoint inhibitors, where higher thymic health was tied to substantially better outcomes. In non-small cell lung cancer specifically, better thymic health corresponded to about a 35% lower risk of progression (HR 0.65) and a 44% lower risk of death (HR 0.56) after adjustment.
What that could mean for patients
If future studies hold up these findings, thymic health on standard CTs could serve as a noninvasive, tumor-agnostic biomarker of adaptive immune competence, helping doctors gauge who is most likely to benefit from immunotherapy. EurekAlert! notes that study authors see thymic health as adding host-immune information that PD-L1 expression or tumor mutational burden do not fully capture. The approach could let clinicians squeeze more insight out of scans patients are already getting for lung screening or cancer care, without extra imaging or blood work.
Experts urge caution and more testing
Study investigators and outside experts stress that the current work is retrospective and that clinical decisions should not change until prospective validation is in hand. They are calling for randomized trials to show that acting on thymic health scores actually improves outcomes. EurekAlert! quotes researchers saying thymic health "has the potential to enhance patient stratification in precision oncology," while underscoring that this remains a promising hypothesis rather than a new standard of care.
How the technology already exists
AI tools that can automatically segment and quantify the thymus on routine chest CT are already in use in research settings, cutting down on manual reading time and improving consistency across studies. BMC Medical Imaging describes a 2025 model that pulls multiple thymic features from CT images and performs reliably across external test sets. That technical foundation is what made the larger population-level analyses from Boston possible.
Mass General Brigham has spent years exploring ways to apply AI to routine imaging and screening, and the new thymus work fits into that broader push to mine existing scans for more prognostic clues. Mass General Brigham highlights the potential to repurpose low-dose CT scans used for lung screening to surface additional risk markers, while experts caution this is not a reason to seek extra imaging. For Boston-area patients, the message is more about curiosity than panic: an organ long treated as background scenery after childhood may turn out to be a quiet indicator of immune resilience.
