Nature Medicine puts lung digital twins inside the donor decision
Donor lungs under ex vivo perfusion alongside a digital function model.📷 AI-generated image / TECH&SPACE
- ★The models were developed from data collected from hundreds of donor lungs during ex vivo perfusion.
- ★The goal is to predict organ function and therapeutic response before a transplant decision.
- ★The work points toward computational support for donor-lung assessment, but not a replacement for clinical judgment.
Ex vivo lung perfusion is already an important intermediate step in evaluating donor organs. Instead of accepting or rejecting lungs on the basis of a limited record, the process maintains and tests them outside the body under controlled conditions, giving clinicians a clearer view of function. In the context of lung transplantation, that additional assessment matters because available organs are scarce and a wrong call carries a high cost for the recipient and for the broader allocation system.
Multimodal data from hundreds of donor lungs under ex vivo perfusion enabled models that assess organ function and can simulate therapeutic efficacy.
Sensor data from the perfusion system feed the donor-lung model.📷 AI-generated image / TECH&SPACE
What makes this report notable is the combination of scale and data type. The source points to hundreds of lungs and a multimodal dataset, which means the model is not built on one isolated measurement. A digital twin here is not a generic patient avatar, but a computational representation of a specific donor organ during testing. If the model can connect changes during perfusion with lung function and potential therapeutic effect, it opens a sharper question: can an organ be improved, or assessed with more confidence, before the final transplant decision?
Caution is necessary. The supplied report does not establish that the system is already routine clinical practice, and it does not provide numerical performance results that should be repeated without the full paper. The direction, however, is clear. Transplant medicine is looking for tools that do more than record organ status. It needs systems that can model what may happen after an intervention. That is especially relevant when the organ is being evaluated outside the body, because the clinical team has a rare window to measure it and potentially improve it before implantation.
For hospitals and transplant teams, the important question will not be whether the term digital twin sounds advanced. It will be whether such a model can consistently reduce uncertainty, help guide therapy during perfusion and support decisions for organs that would otherwise remain in a clinical gray zone. In allocation networks such as the OPTN, any reliable method that improves donor-lung assessment could have real clinical value, but only if it is rigorously validated and clearly integrated into physician responsibility rather than treated as an automatic verdict.
