23rd September 2013

Mirada Medical presents multi-center studies investigating auto-contouring efficiency at U.S Radiation Oncology meeting

A critical limitation of many previously reported auto-contouring studies is that they investigate the performance in only one clinical center or disease site. The research reported here was carried out in a multi-center setting and was performed in collaboration with Moffitt Cancer Center, the Royal Marsden Hospital (UK), Oregon Health & Science University, the University of Oxford (UK), and Mitchell Cancer Center. The study, to be presented by Dr. Mark Gooding, examined the performance of auto-contouring using multi-disease atlases defined at one center but applied and evaluated in other centers. The study found that all evaluators from all centers and for all disease sites agreed that atlas-based contouring can save time and improve efficiency. The study also examined the variability in the assessments across centers and body regions, finding that there was variability in the degree of time saving despite evaluators using the same patient cases. The most consistently high scores were obtained for thoracic and head and neck patients.  The authors conclude that a single set of atlases can save time and improve efficiency, but that further improvements may be achieved through the use of site-specific user-defined atlases.

A second but related study, also multi-centric, examined the performance of auto-contouring for re-planning using a patient’s previous plan. This is relevant for re-treatment cases or in Adaptive Therapy. The study found that all evaluators from all centers and for all disease sites consistently scored the results at a very high level and considered the results to require minimal editing.

A third study investigated the use of different contour fusion methods for multi-atlas auto-contouring in the male pelvis. It compared the performance of two multi-atlas fusion techniques, STAPLE and Majority Voting, with single atlas methods. The study found that the fusion of multiple atlases improves the result compared to single fixed atlases, and that STAPLE can improve over Majority Voting for the male pelvis.

“Automated contouring of clinical data can save valuable time, but only if the resulting contours do not require extensive manual re-working,” stated Timor Kadir, Chief Science and Technology Officer for Mirada Medical. “Mirada’s aim is to make RT workflows as efficient as possible, and we will continue to work with leading institutions to further enhance this technology,” continued Kadir. Mirada Medical will be demonstrating the latest versions of Mirada RTx and Workflow Box and highlighting how cancer centers can improve quality and efficiency at their booth (#2733) during the ASTRO meeting. Dr. Gooding and Dr. Kadir will also be available to discuss the three papers.

Mirada welcomes ASTRO registrants to join the authors for the Poster Reception Monday, September 23, 5:30-6:30PM. Multi-center clinical assessment of DIR atlas-based auto-contouring M.J. Gooding, K.Y. Chu, J. Conibear, T. Dilling, L. Durrant, M. Fuss, D. Gujral, R. Ove, C. Stevens, T. Kadir Poster 3338   Multi-center assessment of auto-contouring using deformable image registration for adaptive therapy M.J. Gooding, J. Conibear, T. Dilling, M. Fuss, D. Gujral, R. Ove, C. Stevens, T. Kadir Poster 3345

Assessment of contour fusion methods for DIR atlas-based auto-contouring of the male pelvis

A. Larrue, T. Kadir, M. Gooding

Poster 2444

For more information, please contact:

Mirada Medical

US/Worldwide: Robert Ripley

tel: (440) 591-8638

Insource Communications