A basic assumption for full realization of conformal radiotherapy (CRT) is that accurate daily patient treatment setup localization and online treatment monitoring leads to improved targeting of tissues tagged for treatment. Validation of this concept would facilitate significant reductions in treated volumes, allowing investigations of improved local control with higher delivered doses at safe levels of normal tissue toxicity. Electronic portal imaging devices (EPIDs) were in large part developed to fulfill this promise. However, in practice current EPIDs suffer from poor image quality relative to radiographic film, high dose (for fields often hot used for treatment themselves), and lack of integration with automated treatment planning and delivery systems. Thus, daily patient setup variations are not generally accounted for at the time of treatment, but assessed off-line with respect to target margins settled upon as compromises between the desires of CRT and treatment efficiency. Such applications, while useful, do not realize dramatic reduction in treatment volumes. It is a hypothesis of this study that major factors inherent in the lack of online use of EPIDs lie in the current detector technology. Over the last decade there has been great interest in two- dimensional sensor arrays based on thin-film electronics for use in x-ray imaging. Recently variety of these area devices (generically categorized as active matrix flat panel imagers (AMFPIs) have become available for basic investigation, as have a few clinical EPID prototypes employing AMFPIs. These devices promise to |improve the quality of megavoltage portal images via low contrast detail sensitivity similar to film, doses smaller than conventional EPIDs or film, and digital image presentation in a sufficiently rapid fashion for online verification and monitoring of tumor position for CRT. It is the long term objective of the current study to investigate the basic hypothesis that AMFPIs have t e required characteristics to make online patient setup and online treatment monitoring possible; thus fulfilling the quest for minimal target volume expansion for inclusion of daily treatment variations. The investigation will include testing the hypotheses that: a) the improvement in image quality of AMFPIs will result in superior image feature localization accuracy and speed and at a lower dose than current EPIDs or film systems; b) the improvements serve as sufficient conditions for clinical implementation of online daily setup and online monitoring of treatments, and c) the improvements will reduce variations in treatment fields thus allowing reduced treatment volumes.
| Kim, Jeongtae; Fessler, Jeffrey A (2004) Intensity-based image registration using robust correlation coefficients. IEEE Trans Med Imaging 23:1430-44 |
| Balter, J M; Lam, K L (2001) Technical note: acquisition of CT models for radiotherapy applications with reduced tube heating. Med Phys 28:590-2 |
| Kim, J; Fessler, J A; Lam, K L et al. (2001) A feasibility study of mutual information based setup error estimation for radiotherapy. Med Phys 28:2507-17 |
