Maryellen Giger, PhD, Professor of Radiology at the University of Chicago talks about new research involving AI to help detect cancer earlier and prevent too many biopsies for patients.
Interview conducted by Ivanhoe Broadcast News in May 2019.
What exactly is 3CB?
GIGER: 3CB and keep in mind this study involved both 3CB and radio mix which is a form of AI. So, on the 3CB side, it is a x ray imaging technique that uses x rays of multiple energies so that you can characterize the tumor and tissue. And then once you obtain these multiple images you can characterize the lipid, the protein and the water content at each point within the image, and because cancer and non-cancer lesions have different content, you can use that information to help distinguish between cancers and non-cancers. But the study then also included aspects of characterizing the tumor from a typical mammogram using AI techniques to pull out characteristics such as speculation and combine those with the different components as well as the different digital characteristics of the tumor in order to help distinguish cancers from non-cancers.
What is in a cancerous tumor that’s not in a benign tumor? What characteristics? Could you give me some examples?
GIGER: Well the cancerous tumor would have – the shape of the tumor, if you look at the margins of that tumor it would have many speculations going outwards. A benign tumor would be more encapsulated, a sharper margin.
What problems was this trying to address or what problems was it looking into?
GIGER: When a woman goes for screening if something suspicious is found she then moves on to additional imaging or perhaps a biopsy. And you want to reduce the number of biopsies because they’re traumatic to the woman. They also resemble a scar that mimics a cancer. So, when she comes back again she may be sent to another biopsy. So, you would like to reduce that but not at the costs of missing cancers. The goal here was to keep roughly the cancer detection rate the same but reduce the number of cancers that would be benign lesions, those are non-cancers, that would go to biopsy. In addition to that let’s say we were able to do that, and our preliminary data shows that. It’s very promising. And we can continue on that route and do additional studies on larger populations to show that when radiologist interpret cases that have been imaged using 3CB including the AI analysis that they can detect similar or more cancers but also reduce the number of biopsies if we can demonstrate that and then have it go to FDA so it can enter clinical practice. While we’re doing that, we may be able to come up with a tumor signature that includes both the 3CB as well as the characteristics that we might be able to see at this time of screening. But going to the next step, if we can show a 3CB AI signature and use it where the people who are being screened, then maybe even at the very early screening stage we can identify more cancers and have less benign cases going to biopsy. But you have to do the research step by step.
How could this research potentially help radiologists?
GIGER: Radiologists goal is detecting cancer as early as possible because the earlier you find it the better the cure. And to do that without causing too many biopsies of non-cancers. This will help radiologists first, in diagnosing cancers better, then ultimately in finding cancers earlier.
How can the research potentially improve the patient’s quality of life?
GIGER: As a patient if you have cancer you would prefer to have your cancer found earlier rather than later because then the cure is more likely.
Is there anything else you think you know about the study, about the research?
GIGER: I would just say this is a great example of how developments in the image acquisition, which is this new 3CB technique, and in artificial intelligence can help in interpretation. Merged together so that we optimized both acquisition and interpretation to give a better overall diagnosis.
END OF INTERVIEW
This information is intended for additional research purposes only. It is not to be used as a prescription or advice from Ivanhoe Broadcast News, Inc. or any medical professional interviewed. Ivanhoe Broadcast News, Inc. assumes no responsibility for the depth or accuracy of physician statements. Procedures or medicines apply to different people and medical factors; always consult your physician on medical matters.
If you would like more information, please contact:
Matthew Wood, Manager of Communications
773-702-5894
Sign up for a free weekly e-mail on Medical Breakthroughs called First to Know by clicking here
