A GPU-based framework for modeling real-time 3D lung tumor conformal dosimetry with subject-specific lung tumor motion

    Authors

    Y. G. Min; A. Santhanam; H. Neelakkantan; B. H. Ruddy; S. L. Meeks;P. A. Kupelian

    Comments

    Authors: contact us about adding a copy of your work at STARS@ucf.edu

    Abbreviated Journal Title

    Phys. Med. Biol.

    Keywords

    RESPIRATORY MOTION; CONVOLUTION METHOD; DOSE CALCULATIONS; BREATHING; MOTION; MONTE-CARLO; BEAM MODEL; RADIOTHERAPY; SIMULATION; FIDUCIALS; TRACKING; Engineering, Biomedical; Radiology, Nuclear Medicine & Medical Imaging

    Abstract

    In this paper, we present a graphics processing unit (GPU)-based simulation framework to calculate the delivered dose to a 3D moving lung tumor and its surrounding normal tissues, which are undergoing subject-specific lung deformations. The GPU-based simulation framework models the motion of the 3D volumetric lung tumor and its surrounding tissues, simulates the dose delivery using the dose extracted from a treatment plan using Pinnacle Treatment Planning System, Phillips, for one of the 3DCTs of the 4DCT and predicts the amount and location of radiation doses deposited inside the lung. The 4DCT lung datasets were registered with each other using a modified optical flow algorithm. The motion of the tumor and the motion of the surrounding tissues were simulated by measuring the changes in lung volume during the radiotherapy treatment using spirometry. The real-time dose delivered to the tumor for each beam is generated by summing the dose delivered to the target volume at each increase in lung volume during the beam delivery time period. The simulation results showed the real-time capability of the framework at 20 discrete tumor motion steps per breath, which is higher than the number of 4DCT steps (approximately 12) reconstructed during multiple breathing cycles.

    Journal Title

    Physics in Medicine and Biology

    Volume

    55

    Issue/Number

    17

    Publication Date

    1-1-2010

    Document Type

    Article

    Language

    English

    First Page

    5137

    Last Page

    5150

    WOS Identifier

    WOS:000280966600016

    ISSN

    0031-9155

    Share

    COinS