Program Information
Investigation of Volume Adjustment Techniques in Low Dose CT Lung Densitometry
H Chen-Mayer1*, S Fain2 , P Judy3 , (1) National Institute of Standards & Tech, Gaithersburg, MD, (2) University of Wisconsin, Madison, WI, (3) Brigham & Women's Hospital, Boston, MA
Presentations
TU-G-204-8 (Tuesday, July 14, 2015) 4:30 PM - 6:00 PM Room: 204
Purpose:
Low dose CT for lung densitometry has shown utility in evaluating severity and progression of pulmonary emphysema in patients with chronic obstructive pulmonary disease (COPD). Effort is underway at QIBA to standardize measurement protocols based on published longitudinal studies on non-diseased subjects. The purpose of this report is to assess an important component of the densitometry CT data interpretation: the volume adjustment necessary to reduce the effect of respiration on density.
Methods:
The 15 percentile density scores (Perc 15), defined as a threshold HU containing 15% lowest density in the attenuation histogram of the whole lung, were compiled for 30 subjects obtained from the National Lung Cancer Trial database with baseline and repeat scans one year apart. A model assuming an adapted sponge model, where the relationship between volume (V) and density (rho) is given by m=rhoV^s (Staring et al, Med. Phys. 41, 021905 (2014)). s can be obtained for each subject, and for the entire sample, which allowed adjustment of the Perc15 value at the repeat scan both at the sample level and for each subject.
Results:
The average bias (change between scan 2 and 1) and standard deviation (SD) of the bias for a) unadjusted, b) sample-level adjusted, and c) subject-level adjusted Perc 15 (HU) values are 2.0 (13.7), 1.5 (3.6), and 2.4 (9.8), respectively, suggesting that the method of volume adjustment affects the assessment of the accuracy and precision of the measurement. Method b) improves both more than c). The distribution of the value s indicates that less than 1/5 of the subjects followed the sponge-like behavior (unity s).
Conclusion:
Volume adjustment has a major impact on the assessment of accuracy and precision of quantitative lung density CT, and therefore the interpretation directly influence the decision making in a clinical setting.
Funding Support, Disclosures, and Conflict of Interest: Sean Fain receives grant funding from GE Healthcare.
Contact Email: