Encrypted login | home

Program Information

Towards Quantitative 4D CT Lung Imaging: Variability of Density Measures Across Scanners From Different Manufacturers

T Yamamoto

T Yamamoto1*, D Pinkham2 , S Yu2 , P Maxim2 , C Kumaran Nair1 , J Ready1 , (1) University of California Davis, Sacramento, CA, (2) Stanford University, Stanford, CA


MO-RAM-GePD-I-4 (Monday, July 31, 2017) 9:30 AM - 10:00 AM Room: Imaging ePoster Lounge

Purpose: Four-dimensional (4D) CT is in routine use for assessment of respiratory motion in radiotherapy at many centers. Variability of 4D CT-measured density across scanners, clinical sites and time has not been explored. The purpose of this study was to quantify variations in 4D CT-measured Hounsfield unit (HU) values between three CT scanners from different manufacturers across two clinical sites.

Methods: A total of six 4D CT scans were acquired with two protocols: (1) clinically-used protocols and (2) protocols modified by reference to the published standardized breath-hold CT protocols for chronic obstructive pulmonary disease studies, on each of the CT scanners from three manufacturers (Siemens, GE and Philips). Modified parameters include the reconstruction algorithm, field of view, and slice thickness. Two stationary phantoms were scanned: (1) the Phantom Laboratory Catphan, and (2) Alderson Rando phantom. HU values measured with the six 4D CT image sets were compared for five materials (Catphan air, PMP, LDPE, polystyrene, and Rando lung).

Results: With the clinically-used protocols, the difference between the mean HU value of each image set and overall mean ranged from -27 to 13 HU for the Catphan air target. The mean value was -916 HU for Siemens, which was considerably lower than GE and Philips (-876 HU for both). With the modified protocols, the differences in the mean HU were greatly reduced and ranged from -3 to 5 HU. There were no marked differences between the clinically-used and modified protocols for the other materials.

Conclusion: 4D CT-measured air HU values vary widely between different CT scanners/protocols. HU variations have been greatly reduced after protocol modifications, suggesting that standardization may significantly reduce variability of 4D CT-measured density and density-based metrics. Future work will focus on investigating a wide range of imaging parameters using a phantom with density comparable to the human lung.

Contact Email: