(Tuesday, 7/31/2018) 10:00 AM - 10:30 AM
Room: Exhibit Hall | Forum 1
Purpose: To quantify the impact of limited-angle (180 - 270 degree) and sparse data on image quality in cone-beam breast CT for faster scanning compared to the conventional 360 degree acquisition.
Methods: Existing projection raw data of a human breast CT were used for image reconstructions. The breast was imaged on a cone-beam breast CT scanner with 300 projection angles (views) in full scan (360 degrees). Projection raw data was uniformly selected for simulating 180 and 270 degree scan with varying angular sampling rate (ASR=projections/degree). Three ASRs (0.83, 0.41, and 0.21) were considered here. Image quality was quantified by cropping two 32 x 32 ROIs (adipose and fibroglandular areas), yielding quantitation of signal difference to noise ratio (SDNR). Two image reconstruction algorithms were considered, including Feldkamp-Davis-Kress (FDK) with Parker weights and Maximum Likelihood Expectation Maximization (MLEM). FDK image with 0.83 ASR in 360 degree scan was chosen as reference. Performance was assessed by measuring SDNR as a function of the ASR.
Results: For MLEM, the SDNR was maintained when ASR was decreased to 0.41 for both 180 and 270 degree scans. However, below 0.83 ASR there was substantial degradation of resolution with MLEM. The mean glandular dose (MGD) can be reduced to 5.4 mGy, and potentially to 2.7 mGy, from the full-scan MGD of 10.9 mGy. Unlike MLEM, FDK with Parker weights suffers from artifacts due to the limited-angle scan even at the highest ASR (0.83) used.
Conclusion: This work indicates the feasibility of using short scan for cone beam breast CT. With iterative reconstruction algorithms, shorter scanning time and less radiation dose becomes possible and the image quality can be maintained. More advanced image reconstruction algorithms will be used with a large clinical database in future investigations of more complex tasks.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by NIH/NCI grants R01 CA195512 and R01 CA199044. The contents are solely the responsibility of the authors and do not reflect the official views of the NIH/NCI.
Keywords
Breast, Cone-beam CT, Reconstruction
Taxonomy
IM- Breast x-ray Imaging: Breast CT
Contact Email
