Premium
SU‐E‐T‐273: Beyond Utilizing Functional Imaging for Contours and Visualization, Towards Physiological Radiotherapy to Improve Patient Outcome
Author(s) -
Zhang H,
Mistry N,
DˈSouza W
Publication year - 2011
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 - 10.1118/1.3612224
Subject(s) - voxel , radiation treatment planning , computer science , dosimetry , medical imaging , image registration , radiation therapy , nuclear medicine , artificial intelligence , medical physics , mathematics , mathematical optimization , medicine , image (mathematics) , radiology
Purpose: To reformulate the treatment planning optimization problem by directly incorporating tissue geometry and spatial function heterogeneity derived from anatomical and physiological images, respectively.Methods: Conventional CT has been the gold standard for anatomical definition of treatment targets and OARs for radiotherapy. Functional images derived from imaging modalities have been utilized to date to draw more accurate contours or define “regions of avoidance” or “regions of dose concentration” to aid the treatment planning process. This framework assumes homogeneity among all the voxels of a region of interest, thereby, neglecting functional heterogeneity among voxels. We propose a treatment planning framework that optimizes voxel dose by explicitly incorporating tissue geometry and heterogeneous physiological function. We explore the use of varying penalty functions applied to organs for which functional information is available. Our approach formulates the dose optimization problem as linear program or quadratic program and solves for pencil beam or aperture weights. We illustrate the reformulated treatment planning framework using ventilation function images from ten locally advanced lung cases. Ventilation function images were obtained from 4DCT imaging. Plan quality and lung function preservation were compared with conventional tissue geometry alone‐based plans.Results: When comparing the plans with and without functional information via traditional DVHs, acceptable V5, V15 and V20 of the lung were achieved by both plans. However, when taking the spatial function heterogeneity into account and evaluating via functional DVH (fDVH), a significant 20% reduction of fV20 (p‐value < 0.001) was achieved by our approach Conclusions: Throughout our retrospective study, we notice that without organ function heterogeneity information available at the time of planning process, very frequently high doses were delivered to organ high‐functioning area, which may be the significant reason many of those patients suffered from complications. With our proposed approach, it is very likely such complications could be avoided.