Sarabjeet Singh MBBS, Presenter: Nothing to Disclose

Mannudeep Karanvir Singh Kalra MD, Abstract Co-Author: Research grant, General Electric Company

Thomas Louis Toth, Abstract Co-Author: Employee, General Electric Company

Dushyant V. Sahani MD, Abstract Co-Author: Researcher, General Electric Company

Sjirk Jan Westra MD, Abstract Co-Author: Nothing to Disclose

Michael Austin Blake MBBCh, Abstract Co-Author: Nothing to Disclose

Sanjay Saini MD, Abstract Co-Author: Nothing to Disclose

et al, Abstract Co-Author: Nothing to Disclose

Raw data based noise projection techniques have been reported for creating low radiation dose CT images. The purpose of our study was to assess if DICOM image based noise projection software can help to generate simulated low radiation dose from higher radiation dose CT images.

An anthropomorphic chest phantom measuring 30x20x15 cm was scanned on a 16-slice multidetector-row CT using automatic exposure control (auto mA, GE Healthcare) with different noise indices (5, 10, 15, 20, 25, 30, 35) and slice thicknesses (1.25, 2.5 and 5 mm) to obtain original images at different radiation doses. A new DICOM image based noise projection software (GE Healthcare) was used to obtain simulated low radiation dose images (with 10-95% dose reduction compared to routinely used radiation dose) from higher dose CT images. The original (n= 21 series) and simulated low dose (n= 54 image series) images were de-identified and randomized. Five radiologists unaware of image types performed a pair wise comparison on a standard PACS workstation for image noise and beam hardening artifacts. Objective image noise and CT numbers were recorded using spherical region of interests. Data were analyzed using linear and logistic regression tests.

There was strong correlation for quantitative image noise (0.97-0.98, p<0.0001) between original and simulated low dose images. None of the radiologists could differentiate image noise or beam hardening artifacts between original low radiation dose and corresponding simulated low dose images. Simulated low dose images were rated as having either identical (75-90%) or slightly lower noise (10-25%) compared to the corresponding original low dose images (p<0.01).

DICOM based noise projection offers a quick, reliable and alternative way to raw data based noise projection to obtain simulated low dose CT images for research and teaching purposes.

DICOM based noise projection offers a quick, reliable and alternative way to raw data based noise projection to obtain simulated low dose CT images for research and teaching purposes.

Singh, S, Kalra, M, Toth, T, Sahani, D, Westra, S, Blake, M, Saini, S, et al, , Can DICOM Image-based Noise Projection Software Provide Simulated Low Dose CT Images? A Phantom Study.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6016194.html