Tooling MRIs to pinpoint prostate cancer

RUTGERS (US)—Computerized tools may soon improve the accuracy of magnetic resonance imaging (MRI) to help doctors identify prostate cancer. The technique potentially could be adapted for imaging breast cancer and other forms of cancer.

Currently, high-resolution MRIs can reveal cancerous tissue on prostate glands and pinpoint where the tissue is concentrated, but it is not always clear whether unusual-looking visual features indicate cancerous growths or benign variations.

“At this time, it’s often just a prediction or an educated guess,” says Anant Madabhushi, assistant professor of biomedical engineering at Rutgers. “Before MRI technology is ready for widespread clinical use, the medical profession will have to be confident that it can make readings accurately and consistently.”

In the new study, researchers will make magnetic resonance images of prostate glands in cancer patients and prepare tissue samples from those same glands after they are surgically removed in the course of treatment.

The scientists from Rutgers and Siemens will then develop computerized tools that align MRI views with digitized images of tissue slices to allow investigators to better identify MRI features that reveal cancerous tissue and develop pattern recognition software that will help radiologists make accurate and timely diagnoses.

“This is an immense and complicated task,” explains Madabhushi, who is also a member of the Cancer Institute of New Jersey.

“Our tools will have to account for variations in MRI and tissue image sizes, tissue that was imaged but lost in the process of sectioning, or MRI images shaped differently from tissue sample images due to gland deformation.”

Siemens will refine research tools into software packages for scientists to use as they conduct additional studies and develop diagnostic tools and techniques.

Each year, there are more than 27,000 deaths from prostate cancer in the United States and 190,000 new cases diagnosed. Diagnosis is based on PSA levels in blood, physical examination, and needle biopsies, because current imaging techniques don’t distinguish cancerous tissue.

MRI has the potential to offer a diagnosis noninvasively and, along with other information, to help physicians customize the most effective and least debilitating treatment plan.

The study is funded by the National Institutes of Health through the American Recovery and Reinvestment Act. Researchers at the University of Pennsylvania School of Medicine are contributing to the research.

Rutgers news: