December 9, 2010: Prostate IGT Presentation

From NCIGT Wiki
Jump to: navigation, search
Home < December 9, 2010: Prostate IGT Presentation

Back to Events

 Thursday, December 9, 2010, 1:30pm
 
 Sack Conference Room
 Image-based solutions for prostate interventions
 Mehdi Moradi, PhD

Abstract

Prostate cancer is the most common noncutaneous malignancy among men. The diagnosis and treatment of this disease involve several types of image-guided interventions. In all of these interventions, unsolved technological challenges present research problems for engineers. For diagnosis, the commonly accepted method is biopsy under the guidance of transrectal ultrasound. However, as a result of the multi-focal nature of the disease and the variability of the appearance of prostate tumors on ultrasound B-mode images, the sensitivity of the biopsy procedure is as low as 50%. We present two potential solutions: an ultrasound-based method that takes advantage of RF time series analysis and supervised classification to create cancer probability maps to augment the biopsy process, and an MRI-based approach that combines features extracted from dynamic contrast enhanced (DCE) and diffusion tensor (DTI) MRI with support vector machines to provide accurate cancer detection. For treatment, low dose brachytherapy (implanting the prostate with small radioactive sources) and radical prostatectomy (the surgical removal of the prostate) are both effective options. However, they often lead to complications, including incontinence and impotence. This talk presents new imaging tools that can combine accurate visualization of the prostate region and the cancerous tumors. In particular, we describe prostate vibroelastography. We show that Ultrasound vibroelastography improves the visualization of the prostate for gland contouring with significantly stronger edges. We also describe a hybrid B-mode/elastography approach for prostate segmentation. We argue that a combination of elastography and RF time series analysis for tissue typing can help in defining the surgical resection planes to avoid damaging critical anatomical structures in conventional or robotic prostatectomy. Finally, we describe the need for real-time dosimetry during brachytherapy. Solutions that include either the detection of radioactive sources based on only ultrasound data, or fluoroscopy-based detection of sources followed by registration to ultrasound are presented. While the presented techniques mainly target the detection and treatment of prostate cancer, the findings can be used to tackle similar challenges in managing other types of cancer, including breast cancer.


Speaker Biography

Dr Mehdi Moradi is a research associate at the Department of Electrical and Computer Engineering, University of British Columbia, and a lecturer of biomedical engineering in the School of Engineering Science, Simon Fraser University, Vancouver, Canada. Since early 2010, He holds a Prostate Cancer Training Award from Department of Defense, Congressionally Directed Medical Research Program and works on developing image-based techniques to improve the delivery of image-based prostate interventions such as brachytherapy and prostatectomy. Dr Moradi completed his PhD (2008) in Biomedical Computing at Queen’s University where he developed a new approach for the analysis of ultrasound RF signals for tissue typing, targeting cancer detection. In 2009, he was awarded a Postdoctoral Fellowship from Natural Sciences and Engineering Research Council of Canada. Dr Moradi has a B.Sc. and an M.Sc. in biomedical engineering. His areas of research interest include biomedical imaging, image-based interventions and medical instrumentation.

Agenda for the Day

  • 11am: Sandy Wells
  • Noon: Lunch (Sandy, and anyone else who wants to join)
  • 1pm: Clare Tempany
  • 1:30pm: Talk (In Sack Conference Room)
  • 2:30pm: Noby Hata
  • 3:30pm: Leave for airport (5:30pm flight)