Investigation on the use of cone beam ct image sets for treatment planning in radiation therapy

Abstract

Clinical studies have emphasized the significance of image guided adaptive newlineradiotherapy (IGART) to enhance the therapeutic ratio. In general, acquisition of repeat newlineCT (re-CT) is the key requisite to implement IGART in the routine clinical setting. newlineSeveral attempts were made to address the practical concerns and challenges associated newlinewith re-CT based IGART, majority of them employed daily cone beam CT (CBCT) newlineimage informations either directly or indirectly for customization of treatment plans newlinewithin the planned course of treatment. However, there are no standard and newlinecomprehensive guidelines dealing with effective utilization of this in-room imaging for newlineIGART. In most of clinical scenarios, longitudinal field-of-view is found to be clinically newlineinadequate for localizing the large and complex targets (with regional nodes) encountered newlinein high precision radiotherapy. More specifically, there is no protocol in the literature for newlinelocalization of the entire craniospinal axis and adaptive plan evaluation methods with newlineCBCT. In addition, inaccuracies in Hounsfield units (HU) in CBCT image restrict its newlinepotential use in IGART. Therefore, a need arises to formulate comprehensive IGART newlineapproaches and guidelines with in-room CBCT to reduce the uncertainties related to newlinetreatment geometry and patient deformation. newlineThe research presented in this thesis describes the potential use of the CBCTbased newlineIGART in several aspects, including specific planning guidelines with customdeveloped newlinesoftware, coded in MatLab. The protocol acquires two or more CBCT images newlinewith a linear translation of treatment couch in the patient plane, allowing 1 cm penumbral newlineoverlap (i.e. cone beam abutment) and fused as a single DICOM data set (CBCTeLFOV) newlinefor extended localization. In order to improve the dose calculation accuracy with CBCT, optimum HU correction (in the high density regions) and HU mapping (from deformed newlineinitial CT) methods were also coded in our software. newline newline