Program Information
Estimation of Image Quality of 4-Dimensional Computed Tomography (4DCT) Via Fourier Analysis of Corresponding 1D Breathing Surrogate
M Caraveo1* , J Wei2 , J McNamara1 , A Rimner1 , E Yorke1 and G Li1 (1) Memorial Sloan Kettering Cancer Center, New York, NY, (2) City College of New York, New York, NY
Presentations
TH-E-17A-8 Thursday 1:00PM - 2:50PM Room: 17APurpose: Motion artifacts are common in patient 4DCT, leading to an ill-defined tumor volume with variation up to 110% or setting up a poor foundation with low imaging fidelity for tumor motion study. We developed a method to estimate 4DCT image quality by establishing a correlation between the severity of motion artifacts in 4DCT images and the periodicity of corresponding 1D respiratory wave-function (1DRW) surrogate used for 4DCT reconstruction.
Methods: Fast Fourier Transformation (FFT) was applied to analyze 1DRW periodicity, defined as the sum of the 5 largest Fourier coefficients, ranging in 0-1. Distortional motion artifacts of cine-scan 4DCT at the junctions of adjacent couch-position scans around the diaphragm were identified in 3 categories: incomplete, overlapping and duplicate. To quantify these artifacts, the discontinuity of the diaphragm at the junctions was measured in distance and averaged along 6 directions in 3 orthogonal views. Mean and sum artifacts per junction (APJ) across the entire diaphragm were calculated in each breathing phase. To make the APJ inter-patient comparable, patient-specific motion was removed from APJ by dividing patient-specific diaphragmatic velocity (displacement divided by the mean period, from FFT analysis of the 1DRW) and the normalized APJ was defined as motion artifact severity (MAS). Twenty-five patients with free-breathing 10-phase 4DCT and corresponding 1DRW surrogate datasets were studied.
Results: A mild correlation of 0.56 was found between 1DRW periodicity and 4DCT artifact severity. Higher MAS tends to appear around mid inhalation and mid exhalation and the lowest MAS tends to be around full exhalation. The breathing periodicity of >0.8 possesses minimal motion artifacts.
Conclusion: The 1D-4D correlation provides a fast means to estimate 4DCT image quality. Using 1DRW signal, we can retrospectively screen out high-quality 4DCT images for clinical research (periodicity>0.8) and prospectively identify poor breathers as candidates of breath coaching prior to 4DCT scan.
Funding Support, Disclosures, and Conflict of Interest: This research is in part supported by NIH (U54CA137788/132378). MC would like to thank MSKCC summer medical student research program supported by National Cancer Institute and hosted by Department of Medical Physics at MSKCC.
Contact Email: