Hiroshi Iimura RT, Presenter: Nothing to Disclose

Kazufumi Suzuki MD, Abstract Co-Author: Nothing to Disclose

Shuji Sakai MD, Abstract Co-Author: Nothing to Disclose

The low-dose condition with increased image noise may result in decreased reliability of quantitative data, especially in MTT. The low-dose condition also leads to a loss of contrast between normal and ischemic tissues in MTT functional mapping. Our results suggest that low-dose condition of CT perfusion may decrease clinical diagnostic reliability of MTT measurements for detecting ischemic lesions.

The relationship between quantitative stability and radiation dose in CT perfusion analysis has been debated, but no consensus was reached. It is difficult to verify multiple conditions in patients because CT perfusion is an invasive procedure with radiation exposure and contrast administration. We developed a new digital phantom to verify the effects of image noise on the quantitative reliability of an ultra-low-dose CT perfusion method.

A digital phantom was constructed from sequential Digital Imaging and Communication (DICOM) images, whose perfusion model is designed to show certain values of cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit time (MTT). Image noise was generated with a scan image of a water phantom and was embedded into the artificial DICOM images. Target MTT was set at 4, 8, and 12 s and target CBV was 2, 4, and 6 ml/100 g, respectively. CBF was estimated as CBV/MTT. CT perfusion analysis was done with commercially available console software (Aquilion ONE 4D perfusion; Toshiba medical systems, Inc., Otawara, Japan). Arterial input function data was obtained from 30 patients with brain tumors with the permission of the institutional ethics board and documented informed consent.

In models with more image noise, which simulates the low-dose condition, slightly smaller CBV, shorter MTT, and larger CBF were observed. With longer MTT models, which simulate ischemic lesions, MTT was more likely to be underestimated. Therefore, the differences with normal tissue were decreased.

Iimura, H, Suzuki, K, Sakai, S, Effect of Image Noise on Low-dose CT Perfusion Analysis— Verification with a Digital Phantom.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14003856.html