Encrypted login | home

Program Information

Evaluation of Dose Reduction with Philips IDose Reconstrutcion In Relations to Image Quality in a Phantom Study


Yun Liang

T Hameed1, Y Liang2*, J Sheng3, J Rydberg4, (1) Indiana University, Indianapolis, IN, (2) ,,,(3) Indiana University, Indianapolis, IN, (4) Indiana University, Indianapolis, IN

PO-BPC-Exhibit Hall-1 Saturday  Room: Exhibit Hall

Purpose: To investigate the use of a commercial designed hybrid iterative reconstruction technique iDose (by Philips Medical Systems) for CT radiation dose reduction and its effect on image quality.

Methods: A catphan504 phantom was scanned using a 64-slice CT (Ingenuity, Philips Medical Systems) with two different tube voltages (100 kVp, and 120 kVp) at three different doses, a reference dose (CTDIvol) of 42 mGy and two reduced doses of about 50% and 75% of the reference. Images were reconstructed with standard filtered-back-projection (FBP) and with iDose algorithms. Six different iDose levels were employed. Quantitative evaluation of spatial resolution, image noise, noise power spectrum (NPS), and low-contrast detectablity were carried out.

Results: For any given dose level, there was a static noise reduction with increased iDose level over the FBP. To match the standard noise resulted from the FBP at the reference dosea minimum iDose level of 4 and 6 was required for 50% and 75% dose reduction respectively. NPS showed moderate shift towards the lower frequency as iDose level increased. The NPS shift was consistent with the observed subtle change of noise texture. This shift also correlated with the change in low-contrast detectability among images with the same noise level, the higher the iDose level that was used in image reconstruction, the lower the low-contrast detectablity.

Conclusion: The iDose algorithm clearly demonstrates effectiveness in noise suppression over the FBP. The low-contrast detectability depends on noise but also on NPS, which is shifted by iDose algorithm. At very low-dose levels, greater iDose levels would be needed to reduce the image noise but may not improve the low-contrast detectabilty. This findings indicate that for any given specific clinical task, the lowest dose limit in combination with an optimal iDose level may be established by considering both noise and low-contrast detectability desired.


Contact Email