No bones about it: Automated subtraction protocol removes bony pixels from CTA images

A single-source, dual-energy subtraction CT angiography (CTA) protocol can be optimized to provide fully automatic subtraction of bone and brain parenchyma from contrast-enhanced scans of patients with cerebrovascular disorders, according to a study published June 7 in Academic Radiology.

The technique demonstrated the ability to successfully remove bone from images of nearly all patients included in the study, while providing excellent delineation of bilateral ophthalmic arteries in more than 80 percent of patients, according to Qi Li, MD, PhD, of Chongqing Medical University, Chongqing, China, and colleagues.

The promising results suggest automated CTA could be an alternative to digital subtraction angiography and MR angiography in the diagnosis of central nervous system and cerebrovascular disorders. CTA has both a higher spatial resolution and shorter exam time than the other imaging techniques.

CTA is limited, however, by the fact that bony structures can hamper visualization of blood vessels, explained the authors. The matched mask bone elimination technique can allow for subtraction of bone pixels, but requires creation of the bone mask and is fairly operator dependent.

To test their automated protocol, which subtracts both bone and brain parenchyma pixels, Li and colleagues prospectively studied 100 patients who underwent automated subtraction CTA for suspected cerebrovascular disorders. In a phantom study, they found that two helical scans with matched x-ray tube start angles allowed for fine registration and subtraction for artifact-free angiograms. “Based on these observations, we optimized our CTA examination protocol by synchronizing the x-ray tube start angle with the first scan and achieved dual-energy subtraction on a single source CT system.”

Results showed that subtraction was successful in all patients, with bone removal rated as excellent in 95 patients. The remaining five had incomplete bone removal due to motion between scans, explained the authors. Arterial segments at the circle of Willis were clearly visualized in 97 patients.

“If CTA is to serve as a noninvasive replacement for [digital subtraction angiography], it must be capable of delineating major intracranial vessels as well as very small arteries,” wrote Li and colleagues. “In our study, bilateral ophthalmic arteries could be visualized in 81% of patients. The results of our study suggest that the optimized dual-energy subtraction CTA was capable of delineating very small intracranial vessels.”

The authors concluded that optimized dual-energy subtraction CTA could be widely used as the primary imaging method for patients suspected of cerebrovascular disorders.