Is AI Better Than Neuroradiologists at Evaluating Aneurysm Growth on CTA and MRA Scans?

For the monitoring of patients with small aneurysms via computed tomography angiography (CTA) or magnetic resonance angiography (MRA), Jeremy Heit, M.D., Ph.D., in a recent interview with Diagnostic Imaging, pointed out that variations in scanning and contrast administration can potentially impact treatment decisions.

“ … Anyone who does imaging (knows) the patient is not always scanned in the exact same plane. There could be variations in the contrast bolus. So, you're not always apples to apples, and that can lead to some uncertainty in the (aneurysm) measurement size,” noted Dr. Heit, the chief of neuroimaging and neurointervention at the Stanford University School of Medicine.

Accordingly, Dr. Heit and his colleagues compared neuroradiologist assessment to automated evaluation with the Rapid Aneurysm software (RapidAI) in a retrospective study involving 44 patients with a total of 78 unique intracranial aneurysms.

According to the study findings, recently presented at the International Stroke Conference (ISC), there were 28 cases of maximum linear dimension growth with aneurysms. While the Rapid Aneurysm software detected 27 of these cases, neuroradiologists detected this growth in only 14 of the cases. Overall, the researchers found that Rapid Aneurysm software provided a 96 percent sensitivity and 94 percent specificity for aneurysm growth in contrast to 50 percent sensitivity and 95 percent specificity for unassisted neuroradiologist evaluation.

“Rapid Aneurysm found substantially more growth, substantially for 50 percent increased catchment in aneurysms that are growing. Now you could say, well, you know, is that wrong? Is that right? How much growth really matters? Those are things we still have to sort out,” said Dr. Heit, an associate professor of radiology and neurosurgery at the Stanford University School of Medicine.

“But it's very interesting that the automated Rapid Aneurysm is finding growth that I am not seeing, and it makes me wonder in all my patientsthat I don't have this information for, what am I missing?”

(Editor’s note: For related content, see “Study Shows Viability of Enhanced 3T MRI for Predicting Intracranial Aneurysm Instability,” “Multicenter Study Assesses Intrasaccular Flow Disruption for Treating Sidewall Aneurysms” and “Study: AI Model Significantly Enhances CTA Workflow Efficiency and Detection for Cerebral Aneurysm.”)

References

1. Snyder S, Rhodes S, Iv M, et al. Artificial intelligence augments radiologist capabilities in tracking aneurysm growth over time. Presented at the International Stroke Conference, February 4-6, 2026, New Orleans. For more info, visit https://professional.heart.org/en/meetings/international-stroke-conference .
2. Sahlein DH, Gibson D, Scott JA, et al. Artificial intelligence aneurysm measurement tool finds growth in all aneurysms that ruptured during conservative management. J Neurointerv Surg. 2023;15(8):766-770.