Improvement of depiction of the intracranial arteries on brain CT angiography using deep learning reconstruction

Abstract

To evaluate the ability of a commercialized deep learning reconstruction technique to depict intracranial vessels on the brain computed tomography angiography and compare the image quality with filtered-back-projection and hybrid iterative reconstruction in terms of objective and subjective measures. Forty-three patients underwent brain computed tomography angiography, and images were reconstructed using three algorithms: filtered-back-projection, hybrid iterative reconstruction, and deep learning reconstruction. The image noise, computed tomography attenuation value, signal-to-noise ratio, and contrast-to-noise ratio were measured in the bilateral cavernous segment of the internal carotid artery, vertebral artery, basilar apex, horizontal segment of the middle cerebral artery and used for the objective assessment of the image quality among the three different reconstructions. The subjective image quality score was significantly higher for the deep learning reconstruction than hybrid iterative reconstruction and filtered-back-projection images. The deep learning reconstruction markedly improved the reduction of blooming artifacts in surgical clips and coiled aneurysms. The deep learning reconstruction method generally improves the image quality of brain computed tomography angiography in terms of objective measurement and subjective grading compared with filteredback-projection and hybrid iterative reconstruction. Especially, deep learning reconstruction is deemed advantageous for better depiction of small vessels compared to filtered-back projection and hybrid iterative reconstruction.

The effectiveness of post-processing head and neck CT angiography using contrast enhancement boost technique

“…The clear vessel delineation may sometimes be challenging owing to artifact and image noise in the posterior fossa and hinder the interpretation of the aneurysm or dissection in the vertebrobasilar system [ 25 ]. With the recent deep learning reconstruction, the image quality was significantly increased with a lower image noise and higher CNR, SNR, and spatial resolution, and the lesion detectability was also demonstrated to be better than the traditional filtered-back projection and iterative reconstruction methods [ 7 , 26 , 27 ]. The vertebrobasilar arteries were rated by both observers as having an excellent image quality in the CE-boost images than in conventional images.…”

“…To clearly visualize small intracranial vessels, various methods (deep learning reconstruction, ultra-high-resolution CT, etc.) have been introduced to accurately visualize small cerebral vessels on CT angiography [ 7 , 8 ]. Deep learning image reconstruction increases both quantitative and qualitative image analysis on CT angiography compared with other image reconstructions by training images through deep convolutional neural networks [ 9 , 10 ].…”

Deep learning reconstruction allows for usage of contrast agent of lower concentration for coronary CTA than filtered back projection and hybrid iterative reconstruction

“…Furthermore, it has been reported that DLR provides lower image noise at lower tube voltage (30). In addition, the blooming artifact reduction is significantly improved with DLR compared with FBP and hybrid IR (21). These advantages could be applied to DLR with a lower radiation dose scanning protocol.…”

“…DLR was developed to reduce the image noise and improve spatial resolution through deep neural networks and its high-quality model-based iterative reconstruction (MBIR) or filtered back projection (FBP) training images, referred to as advanced intelligent clear-IQ engine (AiCE) (14) and TrueFidelity (15), respectively. Several clinical studies have demonstrated that the lesion detection rate and image quality of DLR are higher than those of iterative reconstruction (IR) for low-dose chest CT, abdominal CT, brain CTA, and coronary CTA (16)(17)(18)(19)(20)(21); however, there has been no evidence proving the effect of different contrast media iodine concentrations on DLR in coronary CTA. Enhancing vascular attenuation through DLR decreases the iodine load in patients with renal impairment.…”

Coronary artery calcium quantification: comparison between filtered-back projection, hybrid iterative reconstruction, and deep learning reconstruction techniques

“…TrueFidelity™ is trained with high-quality FBP images while AiCE is developed using high-quality advanced model-based IR images. Both latter methods have reported accurate identification, detection, and classification of abnormal lesions in clinical settings (17,(19)(20)(21)(22). The CAC quantification with DLR (AiCE) has not been investigated to our knowledge.…”