Bring AI to the fight against breast cancer

University Radiology Group
Dr Roger Yang

University Radiology Group (URG), the largest provider of diagnostic imaging and subspecialty radiology services in New Jersey, has over 170 Board Certified Fellowship-trained Radiologists with expertise in all modalities and subspecialties. Their focus is on providing the highest quality care and technology available to their patients and referring physicians. In operation for over 60 years, they have 23 outpatient imaging centres and are affiliated with 12 hospitals.

Dr Roger Yang, President of URG, explains why AI has become an important tool in the fight against breast cancer.

Why Breast AI?

“As Radiologists we are always looking for the highest standards and consistency, particularly important in a large group such as ours. In screening mammography, there is variability between radiologists in their recall rates. So if physician-driven results are disparate, adding artificial intelligence might make these results more consistent and reduce false positives or negatives.”

So why Transpara?

“Like everyone else we had experience with the older computer aided detection platforms (CAD). We saw that things were evolving and wanted to get into the new environment of artificial intelligence applied to mammography interpretation. We found an excellent partner in mammography AI which is ScreenPoint Medical and their AI product Transpara. This is a customisable decision support tool that we can use with our existing workstations. This is something that I think will be very successful because of that flexibility and easy integration into mammography reading workflows.”

What impact has Transpara had on your facility?

“Productivity and efficiency have improved. It’s showing improvements in false positive and false negatives, whilst still finding the cancers that we need to find. Overall, the volume of work is going up and it is becoming more complex. In the past, we would read 15 mammograms
and consider that a heavy load but now, it’s not unheard of to hear of workloads of 150 or 160 mammograms a day! Added to that is digital breast tomosynthesis which has a great deal number of images to look at. Tools like Transpara help us cope with that.”
Did you encounter obstacles in the adoption of AI?

“When the practice looks at implementation, there’s got to be a physician champion who understands what the software is capable of doing. Transpara breast AI is not trying to replace the Radiologist, it provides decision support. That’s actually one of the things that we have to say very early on. Basically, the tool is to improve quality to help Radiologists get through their work and to ensure that things are of a consistently high standard.”

Implementation expectations?

“There will be compromises and when you roll out the technology you need to set the expectations from the beginning. Working with Transpara, whose clinical team helped us to manage expectations early on which all allowed for an easier, more successful, roll out. ScreenPoint managed our expectations accurately throughout the entire process.”

“For ScreenPoint Medical, their goal is not to replace the Radiologists, they never used that language. In Europe, you read the mammograms with a second Radiologist. In the USA, if you want to second read, you ask for it, and some practices here still do have a second read, that’s up to the practice. The mission of ScreenPoint Medical was never to replace a human being, but offer a tool that instead of two human readers, you can have one human reader with Transpara enhancing that reader’s ability to find breast cancers. And finding cancers is what it’s all about.”