Bladder and upper urinary tract cancer causes 16,100 deaths per year in the U.S., and it's expected that 75,200 new cases will be diagnosed in 2013. Early detection of urinary tract cancers is very important; when diagnosed at a localized stage, five-year survival reaches 92%, said presenter Lubomir Hadjiiski, PhD, a research professor at the University of Michigan.
However, only 75% of cancers are detected at that early stage, and survival rates drop significantly when there is regional and distant metastatic disease. Urinary tract neoplasms may also be multifocal, requiring all malignancies to be detected in order to individually treat them, he said.
While MDCT urography is useful for the early detection of bladder and upper urinary tract cancers, interpretation of these studies is a demanding task for radiologists.
"With the increase in radiologists' workload, the chance for oversight of subtle lesions may not be negligible," Hadjiiski said. "Techniques which would assist the radiologist in identifying areas of the urinary tract that may contain urothelial neoplasms would be useful."
The group's long-term goal is to develop an effective CAD system to assist radiologists in interpreting CT urography studies, he said. The researchers tested their system, which combines their previously proposed Combined Model-Guided Path-Finding Analysis and Segmentation System (COMPASS) with CAD analysis.
Preliminary results showed that COMPASS and CAD can track the ureter and detect ureter cancer of medium conspicuity and relatively small size, Hadjiiski said.
"CAD can potentially increase the efficacy of [CT urography] for urothelial neoplasm detection by improving the performance and reducing the variability of both the experienced and the less experienced radiologists," he said.