Portal imaging is the most important quality assurance procedure for monitoring the precision of radiation therapy delivery in standard clinical practice and is even more essential for monitoring the highly customized and technically demanding fields derived from 3D treatment planning. The role of portal imaging has become more critical in recent years due to the migration of 3D treatment planning, including high- precision conformal therapy, from the research setting to routine clinical practice. Unfortunately traditional methods for acquiring and interpreting portal images suffer from a number of deficiencies which contribute to the well documented observation that many setup errors go undetected, and some persist for a clinically significant portion of the prescribed dose. Although electronic portal imaging devices (EPIDs) provide the capability for increasing the frequency of portal image acquisition, there is no practical way to efficiently utilize EPID images for on-line monitoring of treatment geometry. Based on our own clinical experience we hypothesize that significant improvement in both accuracy and efficiency of detecting setup errors can be achieved by a combined approach of 1) enhancing low contrast portal images to aid interpretation by radiation oncologists, and 2) using automatic image registration for on-line screening of EPID images and batch processing of digitized portal films. We discuss plans to continue development of two unique image processing approaches that can be combined to provide this functionality and achieve the desired improvements. One particularly exciting approach is core analysis, a fundamental computer vision method that we have applied to automatic registration of portal and """"""""gold standard"""""""" images. Core analysis is prominent among only a very few approaches that can sensibly use the full power of EPIDs by on-line detection of setup errors. The second image processing technique is SHAHE, a variation of contrast- limited adaptive histogram equalization that has been shown in clinical observer studies to improve the accuracy of port film interpretation by radiation oncologists. Core-based registration and SHAHE have been developed with guidance primarily from anecdotal clinical experience. We plan to use computed but realistic portal images with known setup errors in observer studies to further validate our approaches and to drive ongoing development with more definitive guidance that only formal clinical testing can provide.
| Phillips, Brooke L; Jiroutek, Michael R; Tracton, Gregg et al. (2002) Thresholds for human detection of patient setup errors in digitally reconstructed portal images of prostate fields. Int J Radiat Oncol Biol Phys 54:270-7 |
| Joshi, Sarang; Pizer, Stephen; Fletcher, P Thomas et al. (2002) Multiscale deformable model segmentation and statistical shape analysis using medial descriptions. IEEE Trans Med Imaging 21:538-50 |
| Catellier, D J; Muller, K E (2000) Tests for gaussian repeated measures with missing data in small samples. Stat Med 19:1101-14 |
| Boxwala, A A; Chaney, E L; Fritsch, D S et al. (1999) Comparison of computer workstation with light box for detecting setup errors from portal images. Int J Radiat Oncol Biol Phys 44:711-6 |
| Boxwala, A A; Chaney, E L; Fritsch, D S et al. (1998) Portfolio: a prototype workstation for development and evaluation of tools for analysis and management of digital portal images. Int J Radiat Oncol Biol Phys 42:455-62 |
