How AI is Transforming Detection of the Symptoms of Brain Aneurysm

There is immense value in being able to detect the symptoms of brain aneurysm at speed. Thanks to the speed at which they can cause morbidity and mortality, it has become incredibly important to find ways of detecting aneurysms and treating them even faster. In a world where practitioners are under pressure and the symptoms of brain aneurysm are hard to detect, technology is emerging as a potential game changer, particularly artificial intelligence (AI) and deep learning.

In a recent article in the American Journal of Neuroradiology (AJNR), the authors pointed out that AI has already seen worldwide uptake and recognition thanks to its proven performance in multiple locations and applications. It has provided medical practitioners with that third eye, that pair of hands that never sleep, and ‘substantially improves diagnostic accuracy while reducing physicians’ workload’. This places AI in an interesting position when it comes to unpacking its potential in the realm of the aneurysm and transforming patient care in the future.

The question is, can AI legitimately help detect the symptoms of brain aneurysm and provide medical practitioners with increased speed to diagnosis and support?

The detection algorithm

The symptoms of brain aneurysm are usually detected using a CT scan or a MRA. The challenge is that the images created by these scans are as copious as patient volumes, so radiologists are juggling immense workloads in increasingly complex environments. A brain aneurysm can be detected by these tools, but can be missed in the volumes and stresses of the working environment. This could result in diagnoses being missed or mistakes being made – completely understandable in a workplace that’s defined by speed, stress and life changing decision making.

The above analysis of AI’s role within detecting the symptoms of brain aneurysm and supporting the physician found that AI does have potential when it comes to detecting and evaluating the rupture risk of an aneurysm, in triaging clinical therapy strategies and in predicting treatment outcomes, but that it has not quite achieved the levels of sophistication required to become a fully reliant solution. Yet.

AI is rapidly evolving and companies that specialize in developing solutions within this space are already making headway, leveraging deep learning to dig into detecting symptoms of brain aneurysm.

A deep learning analysis

In an article published by Nature entitled ‘Fully automated detection and segmentation of intracranial aneurysms in subarachnoid hemorrhage on CTA using deep learning’, the authors unpacked some of the challenges that face practitioners when it comes to the detection of aneurysm. Time consuming, complex and challenging, aneurysm detection limits practitioner capabilities and impacts on patient care. The team set out to develop a deep learning model that had the ability to detect and segment aneurysms in patients with aneurysmal subarachnoid hemorrhage (aSAH). The result was a deep learning model that could provide ‘sufficient detection of aneurysms in aSAH with almost 100% sensitivity for aneurysms > 100 mm3’, but that required additional training to ensure that it could achieve the right sensitivity and scope to detect smaller ones.

In another paper, a different team set out to explore how deep learning could tackle intracranial aneurysms to see if it could improve patient care in comparison to clinician assessment. The study, entitled ‘A clinically applicable deep-learning model for detecting intracranial aneurysm in computed tomography angiography images’, found that there were multiple factors that could impact on the performance of a model but that overall, AI could potentially play a highly complementary role in supporting clinician performance and patient care.

In 2019, an article published in the JAMA Network found that deep learning models can improve the performance of clinicians in detecting the symptoms of brain aneurysm. And this is just one model among many, emerging into the proverbial light to provide patients and practitioners with improved care and support.

The solutions in play

Today, there are already trusted solutions in place that support healthcare providers in identifying intracranial hemorrhage (ICH), the most common sign of brain aneurysm. Rapid detection and notification of early ICH helps streamline patient care and can improve outcomes. Aidoc, a leading AI solution, has been successfully validated in studies at several medical centers as a highly specific and sensitive tool to identify acute intracranial hemorrhage.

For example, the 2019 study, ‘Analysis of head CT scans flagged by deep learning software for acute intracranial hemorrhage’, found that while performance varied depending on the patient visit location, the Aidoc solution meaningfully improved radiologist workflow. The platform has since gone on to prove itself in small trials in several US and European healthcare settings, demonstrating the universal value of AI in detecting the symptoms of brain aneurysms.

In January 2021, the Aidoc solution proved itself invaluable when it detected a small right frontal cortical subarachnoid bleed in a patient complaining of a headache and nausea. Once the radiologist was notified, the diagnosis was confirmed, and an investigation of the cause was initiated. The underlying aneurysm was subsequently identified and rapidly treated, all in a matter of a few hours. Although the entire healthcare team shares the credit for this impressive “save,” in retrospect Aidoc’s solution was critical in avoiding a potentially dangerous delay in diagnosis. Considering that ruptured brain aneurysms are fatal in about 50% of cases, this enhanced efficiency is potentially a major life-saver.

The future of AI in detecting brain aneurysms may not yet be fully realized, but it is evolving at pace as organizations focus on deep learning solutions that can support practitioners, improve outcomes, and thereby truly transform healthcare.

Visit the Aidoc Resources page to view clinical studies.

Explore the Latest AI Insights, Trends and Research