Rohit Bhargava, professor in Bioengineering and Bliss Faculty Scholar at the University of Illinois at Urbana-Champaign, and his research collaborators have developed a promising new predictive method to tackle prostate cancer, the second-leading cause of death (behind lung cancer) in U.S. men.
Images of tissue used to demonstrate new chemical imaging method.
The team’s new chemical imaging method will complement existing clinical indices and tools to determine more effectively how well patients with prostate cancer are likely to respond to treatment and whether the cancer is expected to recur.
Traditional predictive methods include manually examining the structure of dyed prostate cells through a microscope to see abnormalities. In contrast, Bhargava, faculty and researchers from Illinois, and physicians at the University of Illinois at Chicago are using their combined expertise in medicine, chemical imaging and computer science to examine both the molecular content and the structure of the tissue.
“This allows us to make decisions based on the chemistry of the tissue and not just its appearance,” Bhargava said. “The chemistry recorded is very complicated, however, and beyond the ability of humans to interpret. So we use computer algorithms to provide us information.”
Bhargava is working closely with Saurabh Sinha, associate professor in Computer Science and member of the graduate program faculty in Bioengineering, who also is among the University’s newest Donald Biggar Willett Scholars. The two began their University of Illinois careers at about the same time and soon discovered similar research interests and complementary skills and experience.
“Rohit is a colleague I respect a lot,” Sinha said, “and he had these data sets that needed a computer scientist’s expertise to make the most out of.”
Jin Tae Kwak, who was a graduate student in Computer Science at Illinois and is now a post-doctoral fellow at the National Institutes of Health, brought additional computational expertise to the project. Kwak also serves as first author on the paper and received a UI Computational Science and Engineering fellowship that was jointly supervised by Bhargava and Sinha.
Bhargava and many other researchers have been working to lower the mortality rate of prostate cancer for years, and this recent work “is a culmination of several technological developments in our laboratory to use novel instrumentation and experimental design to address this urgent need in prostate cancer diagnosis,” he said. "
“While people have long focused on epithelial cells, as prostate cancer arises in these cells, we considered all cells in the tumor,” he said. “Some of the predictive capability of our method arises from these other cells, making our study very distinctive from conventional analyses in this area.”
The researchers were able to address current cumbersome barriers to examining the multitude of cells in the microenvironment — “the cells, molecules and blood vessels that surround and feed a tumor” (National Cancer Institute), for which there are no existing methods. The chemical imaging developed by the UI team, however, adds another dimension to the physician’s toolbox, improving the ability to predict which cancers will recur by examining small changes in chemistry using computer algorithms and complex data sets rather than human interpretation.
“Instead of simply relying on structure of the tissue to diagnose dangerous forms of prostate cancer,” Bhargava said, “our technology uses molecular information from tissue in the hope of providing a new diagnostic test.”
The next step in this research is to use the new instrumentation to rapidly measure clinical samples (biopsies) to validate the method. More than 20-fold-faster instrumentation from Bhargava's group was reported earlier this year in Analytical Chemistry.
The research was published in the journal Scientific Reports, a publication of Nature. The work was sponsored by the National Cancer Institute, National Institutes of Health.
PUBLISHED PAPER IN SCIENTIFIC REPORTS: