image: A digital image of breast tissue from biopsy can be analyzed by computer to reveal features that indicate cancer. view more
Credit: Ajay Basavanhally
NEW BRUNSWICK, N.J. – The National Institutes of Health (NIH) has awarded a Rutgers-affiliated company $207,000 to develop a quick and economical analysis of tissue from breast cancer biopsies. The technology is designed to predict how aggressive a common form of breast cancer is likely to be, helping physicians and patients plan effective therapies that minimize side effects.
The NIH made its grant to Ibris, Inc., through the Small Business Innovation Research program, which federal funding agencies use to encourage American businesses to engage in research and development with commercial potential. Ibris, Inc. is a Piscataway, N.J.-based start-up company that is negotiating a license for computerized image recognition technology developed by Anant Madabhushi, professor of biomedical engineering in the Rutgers School of Engineering.
"Each year, about 120 thousand women in the U.S. and a million women worldwide are diagnosed with ER positive breast cancer," said Madabhushi. Those with less aggressive forms of the disease may respond solely to hormonal therapy, typically tamoxifen, but those with more aggressive forms will also require chemotherapy.
"By determining how aggressive the tumor is, we can help those with less aggressive cancer avoid chemotherapy with its side-effects and expense, and at the same time, we can help those with more aggressive cancer get immediate access to the treatment they need to fight the disease."
Madabhushi's technology examines a high-resolution digital image of biopsy specimens, using computerized image analysis tools to reveal difficult-to-discern characteristics that indicate the disease's grades of severity. It builds on earlier research to detect and grade prostate cancer by analyzing magnetic resonance images of the gland.
The current diagnostic procedure is for a pathologist to examine samples of suspicious tissue to determine whether it is cancerous, and if so, classify the type of cancer and predict its severity.
"The problem is, anytime you have a human being in the equation, there's always a question of variability," he said. "What one observer calls aggressive, another may not."
Madabhushi noted that until less than a decade ago, oncologists would be cautious and prescribe both hormone therapy and chemotherapy because the severity was often hard to determine. In 2004, California-based Genomic Health introduced a test that examined gene expression. That test accurately pegged ER+ breast cancers that fell at the low and high ends of the severity scale. Women who scored low could be treated confidently with hormonal therapy alone and forgo chemotherapy.
While the test, known as Oncotype DX, became an important planning resource for oncologists and patients, it remains expensive – upwards of $4,000. It also takes two weeks to produce results, which for severe cancers represents a loss of valuable treatment time.
"In 2007, an oncologist at The Cancer Institute of New Jersey, Shridar Ganesan, saw our work on prostate cancer imaging and asked if my lab could apply it to ER-positive breast cancers," said Madabhushi, who is also a member of The Cancer Institute of New Jersey. "So we applied these techniques and started to get very exciting results. We found that our image-based risk score, which we dubbed 'Ibris,' was in agreement with the Oncotype DX score almost 90 percent of the time."
Work funded by the NIH grant, made through the agency's National Institute for Biomedical Imaging and Bioengineering, will establish how well Ibris scores predict long-term patient outcomes. If successful, an Ibris test could deliver results to patients almost immediately and cut costs to patients and their insurers.
The benefits could be even more dramatic in developing countries, claims Madabhushi, where the Oncotype DX cost is prohibitive and sending samples to California can add weeks to the testing interval.
"With Ibris, a doctor in Mumbai could prepare a tissue sample, digitize it, and upload it to a server where the Ibris technology analyzes it and prepares a report within hours."
Ibris, Inc. was formed in 2010 by Madabhushi and James Monaco, an assistant research professor in Madabhushi's lab. Scott Doyle, who earned his doctorate under Madabhushi in April of 2011, recently joined the company as vice president of research. Ganesan is serving as the clinical principal investigator from the University of Medicine and Dentistry of New Jersey. Doctoral student Ajay Basavanhally contributed to Ibris technology development. Pathologists Michael Feldman and John Tomaszewski from the Hospital of the University of Pennsylvania have helped with data and annotations.
Madabhushi credits the Rutgers Office of Technology Commercialization for assistance in licensing the technology and creating Ibris, Inc. He also credits the university's Center for Innovative Ventures of Emerging Technologies (CIVET) for helping to arrange early-stage funding for the start-up company and business plan assistance from the Rutgers Business School.