MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Phase Dependency of Deformable Image Registration-Based Target Volume Propagation in the Lung

B Spieler1*, F Yang1, L Young2, Y Yang3, (1) University of Miami, Miami, FL, (2) University Washington, Seattle, WA, (3) University of Science and Technology of China, Hefei, Anhui, CN

Presentations

(Thursday, 7/16/2020) 10:30 AM - 12:30 PM [Eastern Time (GMT-4)]

Room: Track 4

Purpose: Instabilities of using 4D CT aided by deformable image registration (DIR) for target volume propagation was recognized; however, its temporal dependency was little explored and rarely quantified. The current study is aimed to quantify the dependency of DIR-based target volume definition with respect to temporal phase of respiration motion.

Methods: The dataset being used included 4D fan-beam CT (FBCT) of 20 locally advanced non-small cell lung cancer patients undergoing radio-chemotherapy. The 4D FBCT images were acquired during simulation as respiration-correlated CTs with 10 breathing phases (0-90%, phase-based binning) under audio-visual biofeedback being provided to reduce irregularities in respiration. For 4D FBCT of each patient, lesion was delineated by a single radiation oncologist on all 10 phases. DIR-based target volume propagation was carried out for each of the manually contoured volumes on the 10 phases. Differences of the resulting volumes in regard to their respective intended target volume—union of manually delineated volumes on the 10 phase–were pooled according respiratory phases and assessed quantitatively.

Results: Compared to the intended targets, DIR propagated volumes are in general smaller in RECIST long/short diameters with the average centroid displacement ranging between 0.16-0.22cm among the 10 phases. DIR propagated volumes overall are geometrically consistent and intersect substantially with the intended volumes; however, they tend to underestimate the target size by up to 45% depending on the respiratory phases. Evaluation based on metrics related to distance demonstrated that the majority of differences occur at the borderline areas of the target volumes and the exact extent of observed differences varies with respiratory phases.

Conclusion: Error incurring in DIR-based lung target volume propagation is present to varying degrees and exhibits dependency on respiratory phase. These findings allow for the formation of new hypotheses towards achieving robust and more accurate RT target in the lung.

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email