Computer aided detection in medical screening: Potential and current state of art

Speaker: Nico Karssemeijer

Affiliation: Radboud University Medical Centre Nijmegen

Time: Tuesday 07/12/2010 from 14:00 to 15:00

Venue: Access Grid UWS. Presented from Campbelltown (26.1.50), accessible from Parramatta (EB.1.32) and Penrith (Y239).

Abstract: Early detection of disease by screening can dramatically increase treatment options and prognosis. In screening, medical imaging plays a prominent role. For instance, breast and lung cancer screening are based on imaging. Evaluation of images in screening puts a heavy burden on medical professionals, because many images have to be reviewed while only few are abnormal. Subtle signs have to be recognized. Therefore, screening exams are often judged by two readers to avoid errors. To reduce workload in screening and to help readers, large efforts are made to develop intelligent programs to automatically recognize abnormalities in screening images. In breast cancer screening, computer aided detection (CAD) systems are already applied on a large scale. However, despite huge investments in development of CAD, none of the developed systems yet matches the performance of human readers. Most experts believe that the technology is still in its infancy. In the presentation, perceptual issues related to use of CAD in practice and recent advances in CAD research will be discussed. As huge digital imaging archives are becoming available, there are opportunities to improve computer assisted screening beyond the level of experts.

Biography:

Nico Karssemeijer obtained a masters degree in Physics from Delft University of Technology. He graduated at the Radboud University Nijmegen, where he performed research on model guided analysis of medical images and developed a probabilistic framework for segmentation of abdominal CT scans. In 1989 he joined the Department of Radiology of the Radboud University Nijmegen Medical Center, where he formed a research group in computer aided diagnosis (CAD). Much of his work is focused on improvement of breast cancer detection in screening programs.

He is Associate Editor of IEEE Transactions on Medical Imaging, and member of the Editorial Boards of Physics in Medicine and Biology and Medical Image Analysis. He has been chair of the first Computer-Aided Diagnosis conferences of SPIE Medical Imaging from 2007 to 2010 and organized and chaired IWDM98 and IPMI 2007. He (co)authored over 75 peer reviewed scientific journal articles. Nico Karssemeijer was closely involved in the development of the R2 ImageChecker, the most widely used CAD system to date, and is co-founder of MÄtakina, Ltd. (Wellington, New Zealand), a company that develops technology for quantitative mammography.