Researchers studied whether computer analysis of medical scans could help predict outcomes for patients with a small number of lung metastases. They looked at 126 patients from a trial who received stereotactic ablative radiotherapy (SABR), a precise form of radiation. The team developed 'radiomics' models that analyze patterns in CT scans that the human eye can't easily see.
The study found these imaging-based models were better at predicting progression-free survival and early widespread progression than models using only basic clinical information. For example, one model successfully separated patients into higher and lower risk groups. The models showed good accuracy in identifying patients at risk of their cancer spreading more widely within six months.
This research is from a phase II trial, which is an early stage of testing. The study did not compare SABR to other treatments, and the models were not tested in actual clinical decision-making. The main reason for caution is that these promising results come from one group of patients and need to be confirmed in different settings before doctors could use them to guide care. Readers should understand this is a step toward better prediction tools, not a ready-to-use test.