Soenke Heinrich Bartling MD, Presenter: Nothing to Disclose

Jan Kuntz, Abstract Co-Author: Nothing to Disclose

Amit Mehndiratta MD, Abstract Co-Author: Nothing to Disclose

Julien Dinkel MD, Abstract Co-Author: Nothing to Disclose

Wolfhard Semmler MD,PhD, Abstract Co-Author: Support, SMS GmbH

Marc Kachelriess PhD, Abstract Co-Author: Nothing to Disclose

Rajiv Gupta PhD, MD, Abstract Co-Author: Speaker, Siemens AG

4D angiography provides both morphological and functional information about the vasculature. Since the volume of interest (VOI) is irradiated multiple times, concerns about high radiation dose have prevented its routine clinical adoption. This first-in-man pilot study evaluates potential dose savings from prior-image constrained compressed sensing (PICCS).

This IRB approved protocol (2008-P002384/P000704) used a prototype cone-beam CT scanner consisting of a 40x30cm² digital flat-panel detector mounted on a conventional CT gantry to provide a 25x25x18cm field of view. Dynamic angiography was performed in 12 patients to image the head and neck vasculature with clips, coils, or stents (n=7), skull base tumor (n=1), and lower limb ischemia (1), vascular injury (1), and TRAM free-flap surgery (2). A non-contrast scan (140kVp, 50mA, 20s scan time, single 3600 rotation, 30fps) was first acquired as prior and to identify the correct volume-of-interest (VOI) for 4D dynamic angiography. The VOI was centered and 2ml/kg (typically 80-120ml) of Isovue-370 (Bracco, Italy) was administered at 5ml/s. Concurrent with the start of the injection, projection images were acquired for 80s at 140kVp, 50mAs and rotation time of 10s. Filtered back-projection (FBP) was used to obtain 16 time-resolved views of the VOI. VOI was also reconstructed using PICCS whereof 2/3 of the projections were dropped. All reconstructed datasets were evaluated by 3 radiologists in consensus.

The temporal resolution was deemed to be sufficient for differentiating different phases of contrast bolus evolution and to separate arterial and venous circulations. The image quality (IQ) of the full data set reconstructed with the conventional algorithm was comparable to that from PICCS with as few as 50 projections. No significant image artefacts were identified in either of these two reconstructions. FBP with 50 projections suffered significant deterioration in IQ. The total effective head dose of standard reconstruction was 30.2 mSv (729 mGy, Air KERMA), whereof PICCS reconstruction was 12.3 mSv (297 mGy).

PICCS reconstruction can significantly decrease radiation dose of 4D angiography without compromising diagnostic IQ.

This pilot study demonstrates that PICCS can significantly decrease dose without compromising IQ of 4D angiography and such making 4D angiography clinically applicable.

Bartling, S, Kuntz, J, Mehndiratta, A, Dinkel, J, Semmler, W, Kachelriess, M, Gupta, R, Prior Image Constrained Compressed Sensing (PICCS) on a Flat-Panel Cone-beam CT: A New Method for Low Dose, 4D Human Angiography.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12035851.html