Imagine worrying about a second stroke after surviving the first one. That fear is real for many people with carotid atherosclerosis, a condition where arteries in the neck harden. Doctors need better tools to spot who is at highest risk so they can act fast to help.
Researchers examined records from 162 patients, mostly men around 69 years old at the time. They tested a new way to look at CT scans using radiomics, which measures texture and patterns in the plaque and surrounding fat. A combined model using these details beat the standard clinical model based on stenosis degree.
Over a median of one year, 63 patients experienced a stroke on the same side. The new combined model showed an AUC of 0.87, meaning it had strong predictive ability compared to the clinical model at 0.77. The radiomics-only model scored 0.80 in the testing.
This study looked back at past data rather than following patients forward. It shows a predictive association, not a cause. They did not report safety data, and the source text truncated p-values for model comparisons. We cannot say this method is ready for use in every clinic yet.