Research News

Cleveland Clinic researchers are applying quantum simulation to prostate cancer classification, aiming to better capture how aggressive a lesion is. Their results were published in the Proceedings of SPIE--the International Society for Optical Engineering. In the study, researchers from the laboratory of Jacob Scott, MD, DPhil, modeled the behavior of quantum computing to flag clinically significant cancers in prostate MRI images with higher discrimination than current methods. 

Prostate MRI is a noninvasive imaging technique that maps out the prostate using magnets and radio waves. It maps out water and fat, so it isn’t as detailed as a regular photograph, but it can go deep into your tissue.  Prostate MRI is useful for identifying suspicious lesions, but it may not always reliably distinguish cancers that are likely to require treatment from less aggressive disease, so a biopsy is often still needed to confirm the diagnosis. 

“Current computer analyses can extract subtle patterns from MRI images that may not be visible to the human eye, but they still aren't able to account for these grey areas when interpreting a scan,” says study lead author Peng Chen, PhD. “If we can improve the accuracy in image analyses, we might be able to avoid unnecessary biopsies and make sure that patients and their caregiving teams are getting maximum information for their time and money.” 

What is quantum simulation? 

Quantum simulation uses software programs to help classical computers “think” the way quantum computers do by mimicking the behavior of units called qubits. 

Classical computers usually store information in units called “bits.” Bits are like dots on a map, representing one specific location. Quantum qubits are more like the boundaries of a region, capturing multiple possible locations at once. This perspective is much better for interpreting situations that might overlap, like the differences between a stage two or stage three tumor. 

“We approached this project with the opinion that quantum in cancer AI should be more about perspective than processing speed, because cancer exists on such a spectrum,” Dr. Scott says. 

Dr. Chen used a quantum simulation tool to analyze existing prostate MRI scans in a way that accounts for overlap between different tumor types. When the model’s predictions were compared with biopsy results, it showed higher discrimination than the classical computing tools originally used to analyze the images. 

An approach made possible by Cleveland Clinic 

Dr. Chen says her project depended on a combination of resources that aren’t typically available in the same place. Cleveland Clinic brings together clinical imaging data, a large patient population and access to quantum computing tools and working groups within the same research environment. 

These resources were a major reason she came to Cleveland Clinic for her graduate training and chose to stay for her medical physics residency. Her background combines quantum theory, medical physics and imaging research, including an undergraduate capstone in quantum physics and prior work in MRI-based tumor detection.  

Even though the Scott Lab’s study used a simulator instead of the physical quantum computer, that environment made it possible to apply quantum-inspired methods to real patient data.  

Cleveland Clinic provides access to researchers across quantum computing and clinical specialties, as well as IBM’s Quantum System One, located on campus. Dr. Chen and the rest of her team plan to continue exploring quantum machine-learning models in other cancers and for more challenging clinical questions, including predicting treatment response. 

“Experts across fields are accessible if I just email them,” she says. “Some quantum-focused seminars and efforts and working groups are already underway, and people are always happy to help.”