Invention paves the way to better ovarian cancer diagnosis (w/ Video)

July 31, 2015

Overseen by her thesis supervisor and in collaboration with B.C. Cancer Agency scientists, a Simon Fraser University doctoral student is developing a faster, more reliable way to diagnose ovarian cancer.

Computing science student Aïcha BenTaieb is designing a computer-aided diagnostic system at SFU's Medical Image Analysis Lab that will more accurately and more quickly identify ovarian 's many subtypes than the current technique.

Currently, before treatment can begin, pathologists use a microscope to identify tissue samples from cancerous ovarian tumours. According to OVCARE, a B.C.-based ovarian cancer research team, every three-and-a-half-hours in Canada a woman is diagnosed with —one of the most fatal and least understood cancers. Time is of the essence.

With research help from SFU computing science professor Ghassan Hamarneh and B.C. Cancer Agency pathologists Drs. David Huntsman and Hector Li Chang, BenTaieb's automated system will speed up decision-making on appropriate chemo- or radiation therapies.

"Traditional classification can be easily impaired by technical factors such as lighting and the pathologist's experience," explains BenTaieb. "Despite having guidelines for diagnosis, each pathologist has his or her own understanding of the disease, and the act of visually classifying something remains subjective. Classification can also be costly, as non-expert pathologists require long hours of training and additional testing might be needed."

The video will load shortly

BenTaieb's computer prototype is programmed to classify the appropriate cancer subtype using expert-annotated tissue-sample images. They have an agreed-upon subtype verified through biological testing.

The system's built-in artificial intelligence is then able to interpret tissue samples and identify the appropriate cancer subtype by comparing the sample's visual appearance to annotated images.

So far, the results have been promising. The system is already correctly classifying the subtypes in more than 92 per cent of cases.

BenTaieb will present her research paper on this development at the 18th International Conference on Medical Imaging Computing and Computer Assisted Intervention (MICCAI) in Munich, Germany, Oct. 5 to 9. Her paper is called "Automatic Diagnosis of Ovarian Carcinomas via Sparse Multi-resolution Tissue Representation".

Explore further: New research from Lawson uncovers important molecule in ovarian cancer

Related Stories

Survival improving for women with ovarian cancer

May 12, 2015

(HealthDay)—Women diagnosed with ovarian cancer are now much more likely to survive the disease than they were several decades ago, according to research published online May 6 in Obstetrics & Gynecology.

Mesothelial cells promote ovarian cancer metastasis

September 9, 2014

Less than half of the women diagnosed with ovarian cancer will survive beyond 5 years. Ovarian cancer readily spreads to abdominal organs, which are covered by a layer of cells called the mesothelium. Ovarian cancer cells ...

Spotting ovarian cancer, before it's too late

December 19, 2012

(Medical Xpress)—At just 28 percent, the five-year survival rate for women diagnosed with ovarian cancer is much lower than in other cancer cases. And, the disease can easily go unnoticed, making it difficult to find effective ...

Recommended for you

Cancer hijacks natural cell process to survive

June 26, 2017

Cancer tumours manipulate a natural cell process to promote their survival suggesting that controlling this mechanism could stop progress of the disease, according to new research led by the University of Oxford.

Targeted drug shows promise in rare advanced kidney cancer

June 23, 2017

Some patients with a form of advanced kidney cancer that carries a poor prognosis benefited from an experimental drug targeted to an abnormal genetic pathway causing cancerous growth, according to research led by a Dana-Farber ...

0 comments

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.