[BIC-announce] seminar announcement: Today

[Bruce Pike]

>
>
>
> Speaker: Mehdi Moradi, School of Computing, Queen's University
>
> Place: Zames Seminar Room, McConnell Engineering Bldg Rm 437
>
> Time: Wednesday, June 25, 2pm
>
> Abstract:
> Prostate cancer accounted for 29% of cancer cases among American men  
> and
> terminated the lives of 31,350 North Americans in 2007. The common  
> clinical
> diagnosis method for the disease is histopathologic analysis of biopsy
> samples acquired under ultrasound guidance. However, most prostate  
> tumors
> lack visually distinct appearances on medical images. Therefore,
> pathologically significant cases of cancer can be missed during  
> biopsy,
> resulting in false negative or repeated trials. The goal of our  
> research is
> to augment ultrasound-guided prostate biopsy by adding tissue typing
> information that can be used for targeted biopsies. Recently, we have
> proposed a new paradigm in tissue typing. The core of the idea is  
> that if a
> specific location in tissue undergoes continuous interaction with
> ultrasound, the time series of echoes from that location would carry  
> "tissue
> typing" information. In other words, although variations in the  
> intensity of
> a spatial sample of RF echo over time are partly due to different  
> sources of
> noise, they depend on the tissue type as well. To create RF time  
> series, we
> continuously record RF echo signals backscattered from tissue, while  
> the
> imaging probe and the tissue are fixed in position. Samples of RF  
> signals
> collected over time from a fixed location of tissue form one RF time  
> series.
> We extract the fractal dimension and six spectral features from the  
> RF time
> series and use them with Support Vector Machines (SVM) for tissue  
> typing. We
> report an extensive clinical in-vitro study involving 35 patients in  
> which
> the performance of RF time series features for tissue typing in  
> prostate is
> evaluated. An extended version of SVM classification that provides  
> posterior
> class (cancer or normal) probabilities for regions of interest of  
> the tissue
> is used in this study. The outcomes are validated based on detailed
> histopathologic maps acquired from studied specimen. The results of  
> our
> study show that the RF time series features are powerful tissue typing
> parameters with an area under Receiver Operating Characteristic  
> (ROC) curve
> of 0.89 in 10 fold cross validation. We also present colormaps that
> accurately highlight areas of tissue with high risk of cancer on  
> ultrasound
> images.
>
> Bio:
>
> Mehdi Moradi graduated with an MSc in Biomedical Engineering from  
> University
> of Tehran in 2003 and started his PhD at the School of Computing,  
> Queen's
> University in 2004. Mehdi's research interests are in computer-aided
> diagnosis and ultrasound-based interventions. His research on  
> ultrasound RF
> time series and their potential role in computer-aided diagnosis of  
> prostate
> cancer has been praised in several meetings and was the winner of  
> the Best
> Poster Award and the Best Technical Demonstration Award at the  
> Canadian
> Intelligent Systems Conference, May 2007, Montreal.
>
>
>
>
> -- 
> Jeremy R. Cooperstock  | Tel: +1-514-398-5992
> http://www.cim.mcgill.edu/~jer | Fax: +1-514-398-7348
>
> Associate Professor
> Centre for Intelligent Machines and
> Department of Electrical and Computer Engineering
> McGill University
> 3480 University Street, Room 424
> Montreal, QC, H3A 2A7, Canada
>
> ---
> Please note my email policy: http://www.cim.mcgill.ca/~jer/email.html
>
>
>
>
>