‘Brain Bank’ Provides Wealth of Information for Studying Brain Tumors
March 17, 2025
3 minute read
Brain tumors are notoriously challenging to diagnose and treat because they can evade conventional MRI scans and the reach of the most skilled neurosurgeons. Anatomy often makes brain tumors difficult to remove surgically, and their appearance on scans can change over time with treatment.
“Brain tumors are a rare disease,” said Jennifer Connelly, MD, neuro-oncologist and Medical College of Wisconsin faculty member. “Glioblastomas are the most aggressive type. It is a devastating diagnosis that typically has a poor outcome. We always look at optimizing quality of life and extending survival as much as possible.”
Since 2010, MCW has been building a resource to help physicians in that mission. The Neuro-Oncology Brain Bank is a combination radiology lab, brain bank and think tank dedicated to researching the anatomy of brain tumors and how they are captured by imaging. The brain bank currently houses around 150 samples, acquired as donations from patients who wanted to contribute to brain research after their death.
“To the best of our knowledge, our brain bank is the largest brain cancer specific whole brain tissue bank in the world,” said Peter LaViolette, PhD, researcher and MCW faculty member. “Having the actual brain tissue is the gold standard. It helps us understand how far a tumor has invaded the brain. What we see on the MRI scans is, unfortunately, only part of the picture.”
To improve brain cancer detection and treatment, MCW researchers developed a new imaging method called radiopathomic tumor probability mapping. It uses machine learning to analyze donated brain tissue and MRI scans from people who had glioblastoma. Scans are compared side-by-side with microscope images of the brain to determine what was not visible on the MRI. Once the images are aligned, researchers have used artificial intelligence (AI) to produce models that can predict the true extent of the cancer.
Several research studies have validated the tumor mapping approach. One, published in the journal Neurosurgery, showed that mapping revealed hidden tumors not seen on traditional scans in 69% of patients before surgery. Another study in the Journal of Neuro-Oncology demonstrated that radiopathomic mapping could anticipate how the tumor would respond to bevacizumab, a cancer drug given to patients with recurrent glioblastoma to try to prevent new blood vessels from forming and feeding tumor growth.
“Applying AI algorithm imaging of patients who are about to have surgery could help guide surgeons to areas beyond what the conventional MRI sees as brain tumor extent,” Dr. Connelly said. “The hope is that we can do more effective surgery when the patient is first diagnosed, which should correlate with better outcomes in the long run.”
The brain bank will also provide tumor maps to support radiation therapy planning. “We hope that areas of cancer left behind during surgery due to maintaining neurological function in patients can be better targeted by radiation oncologists using the AI-enhanced images,” LaViolette said. “Physicians can then adapt treatment accordingly.”
The capabilities of the Froedtert & MCW Brain and Spine Tumor Program are regionally known. Neuro-oncologists and other specialists collaborate with physicians throughout Wisconsin to offer virtual and traditional consultations. Patients may come to the Froedtert & MCW Clinical Cancer Center at Froedtert Hospital campus for exams and MRI scans, but they can usually have chemotherapy close to home.
The neuro-oncologists also provide second opinions.
“Some people have been told there are no more treatment options,” Dr. Connelly said. “But because of our team’s expertise, there are often things we can do, and patients may survive additional months with good quality of life. Due to generous brain bank donations from patients after they pass away, we are helping patients of the future.”
