Online Master's in Biomedical Engineering @
Biomedical Engineering Alliance
Shedding Light on the Nervous System
Alumni Couple Commit Gift to Computer Imaging Lab
Capitalizing on Opportunities in Medical Imaging

Mining Data for Medical Breakthroughs

Researchers at the Center for Computational Imaging and Personalized Diagnostics use imaging data to create predictive models that can aid disease management and treatment.

On a picturesque evening in July, Pallavi Tiwari headed to the Sunflower Soiree in downtown Cleveland. Like other attendees, she listened to music and feasted on an array of food prepared by local restaurants. While the night was festive, it had a higher purpose — to raise money for Prayers from Maria, the Children’s Glioma Cancer Foundation, named in memory of Maria McNamara, who died from the disease in 2007 at age seven.

Tiwari, a School of Medicine  assistant professor of biomedical engineering at Case Western Reserve University at the Center for Computational Imaging and Personalized Diagnostics (CCIPD), has spent the past four years developing neuroinformatics techniques for evaluating the presence of brain tumors and response to treatment of the disease. Although professional interest drove her to the Sunflower Wine Festival, the evening quickly became personal.

“Working in the center, we don’t regularly interact with patients,” says Tiwari. “But I met patients, parents and family at the fundraising event. They told stories about living with gliomas, and it hit home for me. I need to do whatever I can to make sure there is something out there that can help these children.”

Tiwari’s contribution to fighting brain tumors utilizes automated algorithms to analyze and integrate multi-modal imaging data for disease diagnosis, prognosis and treatment evaluation. The project that has made the most progress to date involves the latter. “Brain tumor patients get aggressive chemo radiation,” says Tiwari. “When they come back for follow-up MRIs, many times clinicians see enhancements on the images that look like the tumor has come back. However, sometimes it’s just a side effect of the aggressive radiation.” Typically, clinicians perform a biopsy to determine if it’s a tumor or benign side effect.

Tiwari and a team of a dozen undergraduates, graduate students and postdoctoral researchers in the CCIPD have accrued and evaluated more than 200 retrospective studies from multiple medical centers across the United States and China. Using machine learning, statistical modeling and pattern recognition, they are close to 90 percent accurate distinguishing between tumors and benign side effects — almost twice as accurate compared with neuroradiologists who read the same studies. If clinicians know exactly what they are dealing with, they can better treat patients.

The work continues thanks to a grant from the Ohio Third Frontier Technology Validation and Start-up Fund. “The results look very promising,” says Tiwari. “What we want to do next is take the technology, deploy it in a clinical setting and see how well it does.”

Extracting Value from Imaging Data

Image captures deformations caused due to a brain tumor and its impact to the rest of the brain, ultimately affecting overall patient survival.  

Tiwari is one of five faculty working together with about 35 other staff members, research associates, students and interns at the CCIPD, which was founded in 2012 by Anant Madabhushi, the F. Alex Nason II Professor of biomedical engineering at Case Western Reserve University. “Our main goal, in the simplest terms, is to try to extract as much value as possible from imaging data to facilitate patient management and treatment,” says Madabhushi.

The CCIPD currently focuses on four broad subject areas: image-guided interventions, digital pathology, machine learning and personalized medicine, and computational diagnostics. One overarching goal is to use imaging data to better predict the aggressiveness of a disease and the appropriate treatment for individual patients — to push the needle in terms of precision medicine.

In just five years, the center has been quite successful. The three primary faculty — Madabhushi, Tiwari and Satish Viswanath, assistant professor of biomedical engineering in the Case Western Reserve University School of Medicine — have more than 50 patents issued or pending related to their research. In the two-year period between September 2015 and September 2017, researchers in the CCIPD have received in excess of $13 million in total funding.

Joining the three primary faculty members in the center are two affiliated faculty, David Wilson, the Robert J. Herbold professor of biomedical engineering, and Vinay Varadan, assistant professor of general medical sciences (oncology) in the Case Comprehensive Cancer Center. Wilson focuses cardiovascular diseases, with recent highlights of his research including coronary plaque imaging using intravascular optical coherence tomography (OCT), detection of cardiac ischemia using CT and detection of coronary calcium using low-cost chest X-rays. Varadan is developing and applying multi-scale systems biology approaches to delineate mechanisms of disease progression and discover novel biomarkers of therapy response in cancer.

While the team has worked on a variety of diseases, ranging from epilepsy to digestive diseases, the mainstay of its research pertains to cancer for two primary reasons — one that’s altruistic, and the other more practical. “Cancer conjures up a fear that perhaps no other disease does in quite the same way,” says Madabhushi. “On a more practical note, the fact is for what we do, data is king. And the cancer space has a long history of carefully recording data and capturing outcomes, which we can use to create predictive models to figure out how these patients are going to do and respond to treatment.”

For data scientists, says Madabhushi, that’s a good place to be. “A critical piece in all the models we create is how did the patient do?” he says. “If you don’t have that information, then you can’t really do much.” Knowing patient outcomes is key to evaluating predictive models.

Although cancer organizations lead the way on the big data front, others are catching up. For instance, the National Institute of Diabetes and Digestive Diseases and Kidney Disorders has called for precision medicine initiatives inspired by the cancer community. “As the infrastructure evolves and the data sets and outcome information become available,” says Madabhushi, “there’s no doubt in my mind that we will start to take our bag of tools and apply it to other indications.”

Collaborating with Clinicians

Collaboration among disciplines is key to the work conducted at the CCIPD. “We need to bring together information from across multiple modalities — MRIs, genetic tests and so on,” says Viswanath. “And the best way to get an understanding of what each modality is capturing is by engaging different people.”

Viswanath’s research on colorectal cancer illustrates the importance of interdisciplinary teamwork. In 2016, he was awarded a three-year, $569,000 grant from the Department of Defense’s Congressionally Directed Medical Research Programs for a colorectal cancer project. “This project is unique because it brings together a variety of disciplines from three different medical institutions – University Hospitals, Cleveland Clinic and the VA Medical Center,” says Viswanath.

A team of pathologists, radiologists, biostatisticians and colorectal surgeons are partnering with Viswanath and Madabhushi, co-investigator on the project, as well as Dr. Sanford Markowitz, who holds the Markowitz-Ingalls professorship of Cancer Genetics at Case Western Reserve University School of Medicine. “Individually, everyone involved does great work. They are experienced researchers and clinicians,” says Viswanath. “But the sum of the parts is greater than any individual.”

The group is developing new image analytics tools to guide treatment and therapeutic interventions in colorectal cancer patients using standard-of-care imaging. “There’s an opportunity to build new tools and provide people with the information they need to potentially avoid surgeries and plan follow-up treatments,” says Viswanath. This would be particularly helpful for colorectal cancer, where it’s difficult to identify the extent and aggressiveness of the disease — or if it’s regressed after treatment.

The project includes two components. The first is a surgical “GPS” plan using analytic tools that provide a map — or navigation system — for honing surgery. If clinicians had a more accurate visualization of the extent of the tumor, they could potentially customize the surgery to the patient rather than perform a complete colorectal excision. The second component is development of a risk scoring system to identify the magnitude of a patient’s disease, then provide a risk stratification score to determine if treatment or a “wait and watch” approach is necessary.

Madabhushi calls the CCIPD’s close proximity to major medical centers in Cleveland “a game changer.” He often schedules one-on-one meetings with leading physicians in breast cancer — one of his research focus areas — and takes students with him. “It’s absolutely fabulous for the students,” he says. “The ability to get immediate, translatable feedback from clinicians is huge. You can’t buy that.”

Creating Collegiality in the Lab

Just as the CCIPD fosters collaboration with clinicians, so too does it emphasize teamwork among colleagues in the center. “We have close to 40 people working on different projects, but because we meet so regularly everyone knows what the others are working on,” says Tiwari. “We help each other out whenever we can.”

One way Madabhushi encourages camaraderie is through the Friday Literary Guild, a practice he started as a professor at Rutgers University in 2005 and continues today at Case Western Reserve University. The CCIPD team is enticed to the weekly meeting through the promise of lunch, then they share status reports on their projects and present research papers. The Friday Literary Guild is approaching its 400th meeting.

“It’s critical to get everybody in the same room, listening to different people present,” says Madabhushi. “It’s very easy for a student to think about his or her particular thesis project and stay in that silo. I understand that, because you have to drill deep. But if you shut off all that’s going on around you, that’s counterproductive. Good ideas may come from another area.”

It’s all about creating synergies — between projects and people. “At the center, we have people in different disciplines looking at different diseases from different perspectives. We are doing disease diagnosis, prediction, prognosis, prevention and treatment,” says Viswanath. “We can attack problems from so many perspectives. When you begin to look at problems holistically, you see connections you might not have seen previously.”

Ultimately, those connections will benefit clinicians and their patients. Patients like Maria McNamara.