[OA040] Quantifying the effect of increased tissue definition on inter-observer variation in contouring of organs at risk in lung cancer patients using motion-compensated imaging. (August 2018)

• Record Type: Journal Article
• Authors: Lee, Joseph; Brook, Kayleigh; Green, Andrew; Osorio, Eliana Vasquez; McWilliam, Alan; Owen, Hywel; van Herk, Marcel
• Is Part Of: Physica medica. Volume 52(2018)Supplement 1
• Page Start: 16
• Publication Date: 2018-08
• Subjects: Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2018.06.112
• Languages: English
• ISSNs: 1120-1797

Abstract

Abstract : Purpose: The development of 4DCT imaging allows clinicians to account for movement of both the target tumour and organs at risk, although the choice of method of image reconstruction can impact how accurately observers can contour nearby organs. Motion-compensated (MC) image reconstruction is a technique that reduces blurring compared to standard averaged reconstruction. This study aims to quantify the presence of sharper boundaries by taking the value of the image gradient at contour lines, and link this increase in tissue definition to lower inter-observer variation. Methods: Eight patients with early stage non-small lung cancer were imaged with 4DCT, and these scans were reconstructed using both motion-compensated and averaged techniques. These two sets of scans were then each delineated by five clinicians, who contoured nearby organs at risk. The mean distance between observers' contours was calculated and compared to the image gradient value at each point on the delineation. Results: Drawing on results by McWilliam et al.[1], the use of motion-compensated CT image reconstruction results in a reduction in inter-observer variation across all sites delineated. The use of MC increases definition of boundaries between organs, represented by an increase in the image gradient at the delineation boundary. This result was particularly pronounced around the trachea, where preliminary results show the use of motion-compensated reconstruction reduced delineation separation by 50%; across all patients, the heart saw very little improvement between imaging techniques. Conclusions: The use of motion-compensated CT reconstruction produces images with more well-defined boundaries between tissue, and higher values of image gradient. When delineating organs at risk, the presence of large gradients is a statistically significant (p < 1E−4) predictor of inter-observer agreement. Regions where this effect was not seen—specifically the heart—is thought to be due to additional organ motion not accounted for when reconstructing the scan.