Premium
Fluence field optimization for noise and dose objectives in CT
Author(s) -
Bartolac Steven,
Graham Sean,
Siewerdsen Jeff,
Jaffray David
Publication year - 2011
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.3574885
Subject(s) - image quality , imaging phantom , fluence , dosimetry , medical imaging , signal to noise ratio (imaging) , simulated annealing , nuclear medicine , optics , computer science , medical physics , physics , computer vision , algorithm , artificial intelligence , image (mathematics) , medicine , laser
Purpose: Selecting the appropriate imaging technique in computed tomography (CT) inherently involves balancing the tradeoff between image quality and imaging dose. Modulation of the x‐ray fluence field, laterally across the beam, and independently for each projection, may potentially meet user‐prescribed, regional image quality objectives, while reducing radiation to the patient. The proposed approach, called fluence field modulated CT (FFMCT), parallels the approach commonly used in intensity‐modulated radiation therapy (IMRT), except “image quality plans” replace the “dose plans” of IMRT. This work studies the potential noise and dose benefits of FFMCT via objective driven optimization of fluence fields. Methods: Experiments were carried out in simulation. Image quality plans were defined by specifying signal‐to‐noise ratio (SNR) criteria for regions of interest (ROIs) in simulated cylindrical and oblong water phantoms, and an anthropomorphic phantom with bone, air, and water equivalent regions. X‐ray fluence field patterns were generated using a simulated annealing optimization method that attempts to achieve the spatially‐dependent prescribed SNR criteria in the phantoms while limiting dose (to the volume or subvolumes). The resulting SNR and dose distributions were analyzed and compared to results using a bowtie filtered fluence field. Results: Compared to using a fixed bowtie filtered fluence, FFMCT achieved superior agreement with the target image quality objectives, and resulted in integral dose reductions ranging from 39 to 52%. Prioritizing dose constraints for specific regions of interest resulted in a preferential reduction of dose to those regions with some tradeoff in SNR, particularly where the target low dose regions overlapped with regions where high SNR was prescribed. The method appeared fairly robust under increased complexity and heterogeneity of the object structure. Conclusions: These results support that FFMCT has the potential to meet prescribed image quality objectives, while decreasing radiation exposure to the patient. Tradeoffs between SNR and dose may not be eliminated, but might be more efficiently managed using FFMCT.