Are radiologists and AI mammography the one-two punch needed in breast health screening? Part 4

Breast Cancer Awareness

Breast Health Series: Advocating for Equity in Breast Care

In this 4-part series, I sit down with Dr. Amy Patel and discuss the importance of prioritizing breast health.

We explore the challenges women face in managing their breast health, potential solutions for these challenges and we look to the future and the role technology can play in optimizing breast care.

Part 1: It takes a village… the breast cancer community

Part 2: What’s driving the rising incidence of breast cancer?

Part 3: Can outreach and education solve the problem of inequity in breast care?

Are radiologists and AI mammography the one-two punch needed in breast health screening? Part 4

October is breast cancer awareness month, but in truth breast health and early breast cancer detection should always be a priority. I recently sat down with Dr. Amy Patel, Medical Director of the Breast Care Center at Liberty Hospital, a true champion and advocate for women’s health.

This week, in Part 4 of the Breast Health Series: Advocating for Equity in Breast Care, I talk with Dr. Patel about the use of technology in breast health screening and early cancer detection.

A quick Google search highlights several recent studies focused on the use of technology to support the detection of breast cancer. Two articles of interest include Improving Breast Cancer Detection Accuracy of Mammography with the Concurrent Use of an Artificial Intelligence Tool and Can AI be a second reader in breast cancer screening?, both focus on AI as a second read, running as a safety net in the background.

In my discussion with Dr. Patel, I ask her about the use of AI in breast screening, read on to get her thoughts on the subject.

Breast Health Series: Advocating for Equity in Breast Care

Part 4: Are radiologists and AI mammography the one-two punch needed in breast health screening?

Kathleen: As we close our discussion on breast health and equity in breast care, I wanted to ask you about the use of technology, specifically artificial intelligence (AI) in the management of breast health.

I’d like to get your input on the use of AI in both mammography and ultrasound. Can you tell me how it’s currently being used? What do you think the future holds for AI in breast care? Can AI help address the bias in the diagnosis and treatment of breast cancer in women of color?

Dr. Patel: I’m in the camp that I think that AI won’t replace us but make us better at what we do, but I do think it will take some time.

When you look at AI, particularly AI mammography, there are still so many in the field that are skeptical because of computer-aided detection (CAD).  We thought CAD was going to be the next best thing since sliced bread, and it just didn’t live up to the expectations in terms of cancer detection. So, we’re really going to need to distinguish AI mammography from CAD. And that starts from the baseline, from the foundation of the AI tools being created. We have machine learning algorithms that look across race so that we can try to minimize bias and have diversity in race, in gender, in age, in geographics. From the beginning, we’re going to need to have that in place to try to eliminate that those biases.

We’re also going to need a system that is reproducible or that can be utilized across all practice types. A lot of these AI tools that are being formed at institutions are just essentially applicable to those institutions. We’re going to need to be able to use systems and tools that can be used in any practice type, whether you’re in a small, rural practice in Alabama, or you’re at a large ivory tower, academic institution on the West Coast. These tools will need to be utilized across any space and still be highly accurate.

I think that given that we are seeing a huge rise in imaging volumes, burnout is arguably as high as ever, physician suicide, unfortunately, is also something that we are facing right now as a medical community. I think that AI has a lot of potential to mitigate that burnout, that additional stress component, and improve our accuracy.

 

Looking to the future, I would hope that artificial intelligence is the mainstay of every breast imaging practice across the country.

Kathleen: Tell me how AI is being used in your practice.

Dr. Patel: With our practice, we’re currently using artificial intelligence breast ultrasound, and we started using this in the summer of 2019. And to be completely honest, it’s very slow-growing. And the gains, the positive gains you see, happen very incrementally. It really takes a persistent, incremental approach. You’re not going to see results overnight. In our practice, we have seven radiologists who read breast imaging and everyone, across the board does feel that this is a helpful tool for them. It helps in terms of, it acts as a second opinion consult, particularly on something that we see on ultrasound that we’re kind of on the fence about. Do we want to give it benign findings or do we want to recommend shorter interval follow-up, or do we want to biopsy it?

AI can help provide that additional diagnostic confidence based on what we are already seeing and thinking, our pretest probability. It can help provide that reassurance. And it can also, particularly for cases when we feel something is not suspicious and AI is flagging it as suspicious and so we go after it, we do a biopsy, and it ends up coming about back as cancer. We do clinical utility, and it has been improving our workflow because we’re not perseverating over so many diagnostic cases. We’re able to get about two to three extra diagnostic cases in a day as a result of this, but it took time to get there and we’re still in the very early stages of AI use.

What we are seeing in the preliminary data that I’ve looked at is that we are maintaining our cancer detection rate while we are reducing our unnecessary biopsies. And that is consistent with others who are using AI breast ultrasound in the field, regardless of practice type.

I do think there’s a lot of promise with AI in the future, but I think it’s going to take us a little longer than we would like. I think it’s important that we are adopting AI in a very methodical way to ensure that this is the best thing for our patients. And at the end of the day, as I said, AI will make us better at what we do, but won’t replace us because particularly in the field of breast imaging, a lot of what we do is patient interaction and the human element and the human touch, and AI can’t take that away.

ACR members using AI

Kathleen: I know AI is used as a safety net in identifying undocumented lung nodules, and this supports early lung cancer detection. Do you think we will see AI used as a safety net with mammography, and I’m specifically thinking 3D mammography?

Dr. Patel: I most certainly think we’ll get there, there’s a lot of companies out there that are investigating that option. I think the UK has done a really good job sinking their teeth into AI mammography and now it’s sort of seeping into the United States and we’re seeing that as well. It’s going to be interesting to look at that data, particularly the way we practice is a little bit different than in the UK. They recommend a routine screening about every two years typically, and they usually have a double read, a double reader where they have two radiologists reading the mammogram. Whereas in this country we typically do annual screening, and we typically have one radiologist reading the mammogram. There are some places in this country though that do double reader, but it’s usually just one reader in the U.S.

It’s going to be interesting to see how this plays out, but I do think we eventually will get there with AI mammography. It’s just making sure that we can prove that it is more accurate, and it is more useful than CAD, I think is going to be something we have to prove.

Kathleen: When you think of technology in general, any type of technology, what is the most interesting in terms of breast health and early cancer detection?

Dr. Patel: I think right now, hands down, the hottest is, is AI. Whether it’s AI mammography, whether it’s AI breast ultrasound, AI breast MRI, and even AI to, for example, help with scheduling. I mean, even little things like that people don’t think about. At NYU, they have created an AI tool that can help them with patient scheduling. There are so many things that are, it’s not necessarily just about cancer detection, but it’s also about our workflows and improving practice efficiency, which leads to patient satisfaction. I think that AI really has the potential to improve so many areas of what we do as physicians in our day-to-day.

Kathleen: I’m going to wrap up our discussion with a couple of questions that are focused on the bigger picture. In the breast health space, if you could change one thing that you think would make the biggest difference, what would that be?

Dr. Patel: Honestly, I think that the biggest thing is if we can get on the same page for screening recommendations. I think it would change so much in terms of insurance coverage, access to care, which can lead to increased cancer detection, reduction in advanced breast cancers, improved mortality, improved morbidity. I think that is a huge thing.

Thriving Thursday: Cancer Screening Recommendations and Updates 2021

Dartmouth-Hitchcock

Kathleen: The second big picture question… We’re going to look 10 years out, it’s 2031, what does the breast health space look like? Think the Jetsons kind of future. How are we taking care of patients? What are we doing? What technology is in use?

Dr. Patel: Well, I know to me and perhaps realistically too hopefully, I would hope that artificial intelligence is the mainstay of every breast imaging practice across the country. And that’s not only in the urban centers but also in the rural centers, the rural hospitals, where they may have lower patient volumes. I would hope that we could get there with AI. We would be able to, for example, and what we’re kind of preliminarily seeing, is you have a work list and AI flags the most suspicious findings at the top of your worklist. You’re then able to go and look at those first, make sure that you get those taken care of, got those patients called back so we can get them in quickly, this helps us with workflow practice efficiency, and it helps with patient care.

I think a large part of the future revolves around technology. There’s a lot of testing happening with blood testing and breast cancer detection and that would be a wonderful way to try to find cancer at its earliest so that we can treat patients faster. But as much as it’s a Jetson situation for me, I’m a realist, and I think that for me in that picture of 2031, I think AI is something that’s attainable and will really improve the field of breast imaging.

What are your thoughts on the use of AI in both managing breast health and early breast cancer detection? Is your hospital using AI in mammography or breast ultrasound? Do you plan to in the future?

We’d love to hear your thoughts and learn from your experience, please drop us a note in the comment section below. 

This is our final segment in the Breast Health Series: Advocating for Equity in Breast Care. The Ferrum Health team would like to thank Dr. Patel for sitting down with us and sharing her knowledge and experience, as well as the amazing work she does advocating for women’s health.

You can find the previous three segments in the breast health series here. These segments cover the work being done by the breast cancer community to bring equity to breast care and address some of the challenges women, especially women of color, face in managing their breast health.

This series is based on an interview with Dr. Amy Patel and has been lightly edited for written clarity.

Kathleen Poulos

Kathleen Poulos

Kathleen is a registered nurse with a digital marketing background, a love for using technology to solve healthcare challenges and a passion for improving patient care.

Contact Us

CASE STUDY

ARA Health Specialists

Use the button below to download your free case study and learn how our approach to validation has improved the number of clinically significant findings in AI software.

CASE STUDY

Sutter Health

Use the button below to download your free case study and learn how our approach to validation has improved the number of clinically significant findings in AI software.