Lead PI:
Center Researchers:
No people associated with this research.
The Challenge
Intracranial aneurysms affect 1-5% of people and aneurysm rupture rates average 1% per year. Endovascular coil embolization is the most often used treatment for these ruptured aneurysms.
A key challenge with endovascular coil embolization is the risk of aneurysm recurrence—where blood flow re-canalizes a portion of the aneurysm sac over months to years. Several studies have demonstrated IA recurrence rates of 20-30% following coil embolization. Recurrence is associated with increased risk of IA rupture or re-bleeding and generally requires additional endovascular or surgical treatment.
Digital subtraction angiography (DSA, the clinical standard for assessment of treated IAs, is an invasive technique involving real-time radiography or computed tomography during intra-arterial contrast injection through an intra-arterial catheter. DSA includes several risks to the patient such as cerebral thromboembolism, contrast reactions, high levels of ionizing radiation, and hematoma or pseudoaneurysm at the arterial puncture site. It costs about $20K and is best performed in specialized centers with appropriate equipment and expertise.
Non-invasive monitoring for treated IAs would result in significantly reduced risk to the patient, increased convenience, and decreased costs compared to DSA, yet existing modalities are insufficient. The two primary non-invasive imaging modalities for assessment of the cerebral arteries are contrast-enhanced computed tomography angiography (CTA) and magnetic resonance angiography (MRA). MRA is sensitive to local variations in magnetic field, such as those induced by aneurysm coils. CTA is faster and cheaper than MRA but suffers from limited resolution and significant image artifacts related to x-ray attenuation by metallic coils. Both MRA and CTA artifacts may completely obscure the aneurysm sac and preclude accurate measurement of the coil mass.
Our Solution
We are developing novel photon counting computed tomography (PCCT)-based non-invasive imaging techniques for improved evaluation of previously treated IAs. This fills an important unmet need for a non-invasive alternative to DSA and will improve our ability to identify IA recurrence after treatment.
Photon counting computed tomography (PCCT) is a new CT technology that provides improved image contrast, increased resolution, decreased artifacts, and equal or decreased radiation dose. Compared to conventional energy integrating CT (EICT), PCCT detectors have smaller pixels, allowing higher resolution imaging (0.2 mm slice versus ~0.6 mm for EICT), and can discriminate photon energies providing spectral information. Associated benefits include artifact reduction — reduced beam hardening and streak artifacts by utilizing information from high energy photons — and improved contrast — specific image components can be accentuated by exploiting spectral differences in attenuation.
Deep learning enabled image postprocessing techniques can further augment the diagnostic utility of photon counting CTA (PCCTA) for evaluation of treated IAs. We are focused on three key image processing domains, which have not yet been fully developed for PCCTA.
- Denoising efforts focus on image-domain techniques that utilize spectral information to reduce noise while preserving anatomic details.
- Metal artifact reduction (essential for evaluating coiled IAs) focuses on intelligent interpolation of metal-affected regions in the CT sinogram.
- Vascular segmentation employs modern deep learning segmentation to isolate vascular structures from surrounding anatomy enabling improved 3D visualization of vascular pathology and generation of patient specific virtual phantoms.
Our preliminary findings show that we achieved a much higher resolution with significantly less metal artifact.
Additionally, we’ve been able to use AI to create virtual models — digital twins or digital phantoms — to make a virtual model of the patients’ arterial anatomy for further downsteam analysis. This allows us to simulate multiple CT scans of their brains and their aneurysm without having to scan the patient again.