Premium
Poster ‐ 51: A tumor motion‐compensating system with tracking and prediction – a proof‐of‐concept study
Author(s) -
Guo Kaiming,
Teo Peng Troy,
Kawalec Philip,
Pistorius Stephen
Publication year - 2016
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4961825
Subject(s) - computer science , actuator , imaging phantom , tracking (education) , motion system , tracking system , motion control , match moving , motion (physics) , computer vision , motion estimation , slider , optical flow , position (finance) , control theory (sociology) , artificial intelligence , simulation , engineering , physics , control (management) , optics , mechanical engineering , kalman filter , psychology , pedagogy , finance , robot , economics , image (mathematics)
Purpose: This work reports on the development of a mechanical slider system for the counter‐steering of tumor motion in adaptive Radiation Therapy (RT). The tumor motion was tracked using a weighted optical flow algorithm and its position is being predicted with a neural network (NN). Methods: The components of the proposed mechanical counter‐steering system includes: (1) an actuator which provides the tumor motion, (2) the motion detection using an optical flow algorithm, (3) motion prediction using a neural network, (4) a control module and (5) a mechanical slider to counter‐steer the anticipated motion of the tumor phantom. An asymmetrical cosine function and five patient traces (P1–P5) were used to evaluate the tracking of a 3D printed lung tumor. In the proposed mechanical counter‐steering system, both actuator (Zaber NA14D60) and slider (Zaber A‐BLQ0070‐E01) were programed to move independently with LabVIEW and their positions were recorded by 2 potentiometers (ETI LCP12S‐25). The accuracy of this counter‐steering system is given by the difference between the two potentiometers. Results: The inherent accuracy of the system, measured using the cosine function, is −0.15 ± 0.06 mm. While the errors when tracking and prediction were included, is (0.04 ± 0.71) mm. Conclusion: A prototype tumor motion counter‐steering system with tracking and prediction was implemented. The inherent errors are small in comparison to the tracking and prediction errors, which in turn are small in comparison to the magnitude of tumor motion. The results show that this system is suited for evaluating RT tracking and prediction.