[00:00:00] Adam Walker: From Susan G. Komen, this is Real Pink, a podcast exploring
[00:00:06] real stories, struggles, and triumphs related to breast cancer. We’re taking the conversation from the doctor’s office to your living room
[00:00:17] This is Real Pink, a podcast conversation where we’re digging deep into breast cancer and the realities patients and survivors face every day. Breast cancer affects everyone differently, but the struggles and challenges this community experiences are real. We’re talking openly and honestly about just how difficult a diagnosis, treatment, and living with metastatic breast cancer can be.
[00:00:38] Technology is rapidly changing the world around us, and the same is true in women’s health. Artificial intelligence holds enormous promise in understanding risk, detecting cancers earlier, and improving treatment options and outcomes for breast cancer patients. But how do we get there, and what should we be aware of along the way?
[00:00:57] Today, we’re welcoming Jillian Wright, the CEO of Onsite Women’s Health, the leading provider in convenient in-office breast health imaging services across the US, and Jane Perlmutter, a survivor of multiple cancers and devoted member of Susan G. Komen Advocates in Science program, which brings the patient voice to scientific research and clinical decisions.
[00:01:19] We’re excited to hear from Jillian and Jane about the ways AI can drive improvement in women’s health and their ideas on how it could change breast health services and patient outcomes. Jillian and Jane, welcome to the show. Jillian let’s start with you. Why don’t we, why don’t we dive in?
[00:01:35] Jillian Wright: I’m Jillian Wright, currently act as CEO of a company called Onsite Women’s Health.
[00:01:42] We partner with OBGYNs and primary care providers to implement in-office screening mammography and other breast health services today. So think of it as you go for your annual well-woman visit, you’re able to get your mammogram right then and there instead of needing to go to another facility.
[00:02:03] I am a Nashville native which is rare these days in a city ful- full of new people moving here, which I love. And I’ve spent my whole career in healthcare. Previous to being with Onsite, I worked at a company called AmSurg in the outpatient surgery center space for about 13 years and was really excited for the opportunity to come over to Onsite and really try to push the envelope a little bit in the women’s health and breast cancer screening space.
[00:02:34] I’d love to hear a little bit about you.
[00:02:37] Jane Perlmutter: Well, it’s really nice to meet you, Jillian. I’m a little bit surprised that our paths haven’t crossed, but maybe they will in the future. I’m Jane Perlmutter. I’m i- today I’m playing the role of, not just the role of, I am a actually three-time breast cancer survivor.
[00:02:55] I am a retired professional. I am trained as a cognitive psychologist. I started out in academia, but I’ve worked in R&D and other areas. But now I’m retired, but I do a lot of cancer advocacy. My first breast cancer was over 40 years ago, and other patients provided peer support, and it was so powerful to me that ever since then I’ve been somewhat involved.
[00:03:20] But I’ve been– come much more involved more recently as, number one, there are more opportunities. Number two, I’m retired, and I do a lot of work, particularly in research advocacy, especially in clinical trials. And in part, my sort of academic training sort of fits well with that. In a previous lifetime, I actually taught statistics to be-behavioral scientists, and it was my favorite thing to teach.
[00:03:44] So clinical trial design is certainly something that’s a natural. I live in Ann Arbor, Michigan, with my identical twin, who’s retired from the university. But I grew up in the Bronx, and that’s always a part of me. I left s- when I was seventeen, which was a couple of years ago. Not only am I a three-time breast cancer survivor, and that’s all, all early stage.
[00:04:08] I was diagnosed the first time in my early thirties, so I was not due to screening.
[00:04:13] Jillian Wright: I was just about to do the math on that because you just disclosed your age prior to this, and I thought that is very early to have been diagnosed.
[00:04:21] Jane Perlmutter: Right. And it– well, I have very thin skin, and it was just obvious in the shower.
[00:04:26] But it was early stage. I, I’ve since had two more early-stage breast cancers. But since then, I’ve had metastatic esophageal and metastatic lung cancer as a result of the radiation I had back in the eighties. However, I have been the recipient of lots of research and have done remarkably well given these new immune therapies and just feel really lucky.
[00:04:51] My advocacy is not about my cancers, but it certainly does give other people hope, and it motivates me even more. My hope is that everyone else is as lucky as I am.
[00:05:04] Jillian Wright: Well, no doubt there is a reason you are still with us, so I’m excited to learn more about some of the things that, that you’ve worked on.
[00:05:12] And in particular, as a young woman myself I’m– I turn thirty-nine tomorrow, so I’m below the age of forty, yet running a, a company that does six hundred thousand mammograms a year, right? Yet I’m not at mammogram age. However, I am at risk given a family history. Both my grandmother and my aunt were diagnosed with breast cancer, both survivors, very strong women.
[00:05:39] It’s one of the things that really drew me into this, this particular field. But because I’m so well-educated on the subject, I talked to my provider and made sure that she did a breast cancer risk assessment for me, and then I was able to go forward, get my mammogram, found a lump through that, went all the way through to biopsy, and we believe it’s a fibroadenoma, but it looks a little bit like a phyllodes tumor, so we’re kind of monitoring that every six months.
[00:06:11] It’s not as well understood yet, cancers under the age of 40, and it seems to be… The data is showing that it is becoming more frequent, and those cancers are so aggressive. So I’m really a big advocate for women understanding risk understanding that you don’t have to wait until the age of 40 to get a mammogram if you’re high risk, and just understanding those different factors.
[00:06:35] But back in the early ’80s, when you were diagnosed in your early 30s risk below the age of 40, I mean, even below the age of 50 at that time wasn’t well understood, so you probably weren’t even thinking about needing a mammogram.
[00:06:51] Jane Perlmutter: A- absolutely. There w- y- you were informed and you knew about family history, and things have changed a lot since the ’80s.
[00:06:59] Just awareness. Back then people didn’t talk, in my neighborhood, people didn’t talk about cancer. It was the C word, and it was hidden. And the word breast was not printed in The New York Times. Right. But I was, I have a very thin skin, and I felt something in the shower. I went to my doctor.
[00:07:17] We did do a mammogram at that point, but he said, “Oh, this looks like nothing. It was my wife. I would not worry about it.” the book My Body, Myself had just come out, and I r- read about it, and it just bothered me, so I went back to him and I said, ” my breast is probably fine, but my head isn’t, so let’s do a biopsy.”
[00:07:36] And that’s started the whole shebang that you know too much about.
[00:07:41] Jillian Wright: Sure, sure. Yeah, and it’s… Your story is so, unfortunately, so common of women’s concerns being pushed off as being overly worrisome. And I guess I would say I, I am hopeful that we’re turning a corner on that with as much emphasis that is happening in women’s health today, and the new things that have been discovered.
[00:08:05] I feel like women are being heard better, and women are feeling more comfortable educating ourselves so that we can feel confident speaking to our providers and advocating for ourselves. I think the patient-provider relationship’s so different than it used to be, and I think that’ll be a really good thing for all of the women generations to come.
[00:08:27] Jane Perlmutter: R- right. A- and y- I think consumerism, people are more aware, but there’s still a lot of people that don’t even have a primary care doctor. I mean I think in addition to increasing awareness of women, doctors have to be aware to say a primary care doctor about r- risk, and gynecologists, who m- more women do see.
[00:08:50] And y- being in the same facility, I think that’s probably a real advantage. But on the other hand, doctors don’t have the time to spend with people, and so much, and so many people a- at a vulnerable age are not even bothering to see a doctor regularly.
[00:09:05] Jillian Wright: Right.
[00:09:06] Jane Perlmutter: So th- there are a lot of challenges.
[00:09:09] Jillian Wright: Some of the things similar along the lines of what you just said, doctors don’t have a lot of time. That is so true. Doctors don’t have a lot of time and we’re asking them to do and know so much more. So what we do at my company as when we partner with the providers to put mammography services in the office, we’re really intentional about trying to kind of do the full service so that it doesn’t just rest on that provider.
[00:09:36] And I know our topic is AI, but we are really trying to use a lot of AI to do that. AI tools to help categorize and risk factors, AI tools to help detect breast cancer, and actually other things, which I’ll be excited to talk a little bit more about with the mammogram. AI tools to look at the quality of the mammogram, things that maybe really had to be done by the provider before, we’re now able to do or deeply supplement with technology, and AI is a big part of that.
[00:10:13] So I would be curious to hear from you as a patient that’s gone through kind of every bit of this spectrum, how do you feel about the introduction of AI into healthcare or this kind of cancer detection space?
[00:10:29] Jane Perlmutter: Well y- I am a Believe it or not, a really early advocate of AI.
[00:10:36] In the ’70s, I did a master’s thesis in computer science, and my topic was an AI topic- Oh, wow … which people don’t realize that there was AI back even in the- Sure … ’50s with Turing. I was c- just going to say, I think one thing that AI can do is just help make patients aware, remind people to get them scheduled.
[00:10:57] I think some patients, and a lot of patients are actually concerned that will depersonalize things. But in some ways, some of these new AI chatbots are even more personal, and some people recognize that. The, the jury is still out. There’s pros and cons and certainly there’s probably, U- using AI to analyze mammograms is one of the furthest along, but in my mind, there are so many other really exciting places that AI are going to make a huge difference in not only early detection, but in treatment, in research, in just learning so m- th- we’re going to accelerate our learning, and it’s really exciting.
[00:11:42] Even more than mammograms, and we could get back to that, but I think already we’re making progress in using AI to analyze pathology slides. And we can learn so much more than with the naked eye. And a lot of the sequencing that takes forever, that takes really very sophisticated technology, we’re learning now that we can actually do that from pathology slides.
[00:12:06] And increasingly we’re going to be doing pathology based on blood, and AI’s going to help with that. So definitely in detection, in diagnosis, and then in finding the right treatment for the right patient, Right … Both because of the biology, but then also AI is already being used for trial matching.
[00:12:30] Unfortunately, there are so many possible ways that we can learn from clinical trials, and many patients don’t know about them, or they’re don’t find a trial that matches what they, their eligibility. But with AI tools, we can connect them, and we’re beginning to do that which I think is a really, a, a great application.
[00:12:50] Jillian Wright: So- Pretty incredible the speed at which AI is really beginning to drive Progress in all of these spaces. Right. Right. A- It’s a day is like a year.
[00:13:04] Jane Perlmutter: Right. And then using AI to data mine real world evidence as well- as previous clinical trials to predict who’s going to respond, to predict who’s going to have side effects, to really d- do the best that we possibly can.
[00:13:18] Y- I’m sure you, y- you probably are using AI in some of your applications, and I have heard some of the early work in AI on mammography, and I have both my pl- pluses and minuses that we can talk about. But why don’t you tell us a little bit about what you’re doing, and then I’ll tell you some of my concerns.
[00:13:39] Jillian Wright: Yeah, I would love to hear those because it’s really important. We as the, on the kind of healthcare provider side, we study these things, we test them, we become so familiar with them that we trust them, and then sometimes we don’t recognize that the patient hasn’t had the opportunity to do all of that.
[00:13:59] So how can we make the patient feel comfortable too? So I’m excited to hear your perspective. We started at Onsite, if I’m thinking kind of with AI as it relates to the clinical journey. We started with AI to assess the patient’s breast density. So that was one of the earlier applications for us. Breast density, as you likely know is a, a big risk factor, so both for breast cancer, but also related to how well the mammogram alone will perform for detecting the breast cancer.
[00:14:32] So breast density has been commented on and categorized by radiologists forever, but it is very subjective. And so we actually spearheaded a study to look at the subjectivity across radiologists acro- between the same patient over time and even within the same time span. And even the same radiologist reading the p- the same patient year after year had a, a very different opinion of the patient’s breast density, even when the images didn’t differ that much.
[00:15:03] So AI related to assessing breast density is a great way for us to just gain a little bit kind of more standardization around that and less subjectivity, and that is so important because it leads to, one, better understanding of the patient’s risk classification and what she should be doing for imaging, and also really kind of helps with what other imaging will she qualify for from insurance related to payment.
[00:15:34] So- that’s kind of the first thing we did. The second thing we did is to use it to assess the quality of the images that we were obtaining. So your read is only as good as the images that were taken by the technologist. So I’ll put my plug in right now for compression and the mammogram is uncomfortable, and it needs to be uncomfortable to get a beautiful, crisp, clear image for the radiologist so that he or she can see everything.
[00:16:01] And we didn’t have a way before this AI solution to, to really analyze all of the images being taken to benchmark our mammo technologists against each other and against standards for how well they’re taking images and where they could maybe improve. So we implemented some AI solution around that.
[00:16:21] And then we moved into what I think everybody would most traditionally think of for AI in breast imaging or radiology in general, and that was AI for enhanced breast cancer detection. So we have partnered with a company called FeraPixel to utilize their mammo screen product. We tested everyone out there on the market, and our radiologists felt the best about this one far and wide, and what it has done for us is incredible.
[00:16:48] It has reduced unnecessary callbacks for imaging, so what we would call false positives from the screening mammo where a patient needs to go for diagnostic imaging only to learn that she’s okay return to screening. Obviously, we always would prefer to be safe than sorry and be more conservative than necessary, but it does drive a lot of cost into the system and a lot of anxiety for patients.
[00:17:13] So if we can tighten that a little bit and ensure that There are no false negatives, so we haven’t missed anything with AI helping support that radiologist read, but also reduce the number of false positives and drive up our cancer detection rate and catch it as early as possible. So we have seen those, those false positives come down, and we have seen our cancer rates, our cancer detection rates go up across our radiologists who are all fantastic breast specialized radiologists.
[00:17:45] So some would think, “Oh, well, AI needs to be used to support radiologists that that need that help.” That’s not necessarily the case. It helps every radiologist, and it’s because they’re human. And so the coupling of a human eye and AI, I think is where the sweet spot is right now. So we’ve done that, and then finally this year we have launched a program called Mammo with Heart, and I am thrilled about it because it actually uses AI to detect breast arterial calcifications, which have long been seen on mammography, but they’re just an incidental finding.
[00:18:22] They have nothing to do with breast cancer risks. But it turns out they have a lot to do with cardiovascular risk. And so being able to use this mammogram, the same image, no additional radiation, no additional appointments, to also find a cardiac risk marker is going to, I believe, be pretty revolutionary.
[00:18:44] We actually just found it in a patient who otherwise had no family history, never been to a cardiologist, no symptoms, and found the breast arterial calcifications. She went to a cardiologist, had additional imaging, labs, and then a procedure, and she was diagnosed with advanced coronary artery disease and is now in a rehab program.
[00:19:09] So maybe saved her life by taking that incidental finding on a mammogram and turning it into something more. So, I mean, I know I’ve just now gone on a monologue about it, but it’s so exciting the number of things that we can do and are doing in this space with AI.
[00:19:28] Jane Perlmutter: Well the two things that typically are raised as concerns and patient concerns are, number one, how the tools have been trained and whether or not there’s an adequately diverse patient population.
[00:19:43] And I think th- that’s debatable. It depends on the particular tool. But I think particularly when we think about minority communities, we need to make sure that we have adequate data, and not just representative data, but adequate data, and we understand when there may be some differences and utilize that.
[00:20:02] And I don’t think we’re yet to a, a really good point on that. The second concern that we hear from quite a few patients has to do with privacy issues, with anything with technology. Personally, I don’t think that’s irrele- relevant. I don’t think any of us have any privacy anymore. It’s one of the-
[00:20:23] Jillian Wright: Yeah, it’s gone anyway, right?
[00:20:24] Jane Perlmutter: It’s one of the trade-offs that we have with having more information, but it definitely is a concern. And in the health area, we– I think it’s even more of a concern, and we do try very hard to protect people’s privacies. Nevertheless, I get notes from my portal fairly regularly about break-ins, and nothing has been really safe.
[00:20:47] Some of my– the concerns that I’ve seen in research that I’ve done, I can sort of share with you. One, I’ve seen some studies of sort of deploying AI to assist radiologists or radiologists, as you say, in mammograms. And interestingly, when it was first deployed in some study I read, it really improved the detection rate because the tool helped radiologists look at suspi- suspicious spots in a different way.
[00:21:19] However, over time, the radiologists became more and more enamored with the tool and stopped looking themselves, and then it fell back. So I think we need to be cautious about how these things are implemented. So that’s one concern. Another concern is, I think there’s pretty good evidence that using AI, we can detect things a little bit earlier, some cancers earlier.
[00:21:45] And early detection definitely saves lives. But I do get concerned that sometime we– there are indolent cancers that w- don’t need to be detected, and we’re going to be over-detecting, over-treating. And I haven’t seen data on that, but it is something that I have thought about a bit and thirdly something that kind of has appalled me and actually an ethicist asked me about my perspective on it, but apparently some places that are using AI in mam- with mammography give patients the option of whether or not they would like AI to assist in their mammogram, which is probably a good thing.
[00:22:30] But if the patient chooses, there’s an additional charge. One could make the argument that we, our healthcare is not evenly distributed and there are these costs, but it does it feels un-unright and it does raise the whole issue about when we have new technologies, are we actually ind-increasing disparities?
[00:22:51] This is a, one example, but there are o-other ways that AI, I think unfortunately could increase rather than decrease disparities. So it’s a general concern, and it goes beyond just AI-assisted mammography.
[00:23:08] Jillian Wright: Jane, on your third point there so we are one of the, A few, if not only breast imaging companies, and I don’t want to speak out of turn there, there likely are others, that chose to add AI across for enhanced breast cancer detection across all of our screening mammograms at no cost to the patient, so making that standard.
[00:23:32] You’re correct, most places today are offering it as an add-on and that’s because it really, it’s expensive. For us we’re investing quite a bit in those licenses and getting no additional reimbursement, but we believe the improvement in the outcomes is worth it far and wide. And to your point about further creating divides from a, a healthcare equity perspective, we just weren’t interested in participating in that.
[00:24:04] But it is tough. You– when you’re at these point where these things are so new that they’re not yet reimbursable and you need more real-world evidence to support it to get reimbursed for it, I understand how companies find themselves at a cross– at the, at this crossroads. But we just we couldn’t do it.
[00:24:22] And so really proud of our decision despite maybe the financial implications of that because I feel like it’s the right thing and it has borne itself out in the outcomes. But you hate for companies providing healthcare to have to make those kinds of decisions. It’s tough.
[00:24:40] Jane Perlmutter: Right.
[00:24:40] The last concern I have about AI-assisted mammography is I actually believe that mammography is going to be going away. We have all of these multi-gene early detection methods, blood-based things and other even imaging modalities. MRI has many people think, advantages over mammography, but that’ll take quite a while before that happens.
[00:25:08] We’ve talked a lot about, though, early detection, and I think we might not be having very much time left. But to me, as I said before, the more exciting areas of AI and healthcare have to do with pathology, finding the right drug for the b-right patient, and learning drug development i- designing new molecules, and then data mining from all of the clinical trials and real-world evidence that we have s-so much and we can learn so much that could really help patients rapidly.
[00:25:42] And I think we probably don’t have enough time today to discuss all that, but I hope some of the future talks on AI will get into those topics.
[00:25:52] Jillian Wright: Yeah, I think that’s I’d like to learn more from you around that side. We… I’m very tunnel visioned around the early detection side of it. But once detected and in treatment how can we make that journey as effective and impactful and safe for the patients?
[00:26:13] And it seems like tons of work is happening there.
[00:26:17] Jane Perlmutter: Right. I mean, the good news is there are more and more long-term survivors, but the bad news is cancer is still the second leading cause of death, and we have so much to go.
[00:26:28] Jillian Wright: That’s right.
[00:26:29] Adam Walker: I want to thank you both for being on the show today for sharing your personal experiences and your perspective.
[00:26:36] It’s really, it’s genuinely impactful What excites you about what you’re doing to drive positive changes in women’s health and the future of breast cancer detection?
[00:26:47] Jane Perlmutter: Well I very little of the work that I do has to do with detection, but I do quite a bit of work in breast cancer in two spheres.
[00:26:57] The probably the thing that I’m most passionate about is the I-SPY trial. That is a platform trial that’s been going on for almost 20 years. I was th- involved in it well before it got started. I’m the lead advocate of about 30 advocates that we have involved in it. We’re involved in it in every single way and it’s just such an innovative trial.
[00:27:19] We’ve learned so much in so many different ways. We could definitely do a whole podcast or- a whole series of them on that and I’m most passionate about that. There’s also a, a little over 10 years ago the Metastatic Breast Cancer Alliance was formed- which combi- w- works together of all different breast cancer advocacy groups, but focuses on metastatic breast cancer.
[00:27:43] And in addition to the advocates, the, the drug companies are involved, and they’ve been doing some really great work, and I’m proud to say that I was also actually one of the first people involved in that. And in fact later this month, I’ll be going to a, a meeting that they have on that topic.
[00:28:01] So Jillian, what about you?
[00:28:04] Jillian Wright: Well, let’s see. I’m going to probably cheat a little bit too and not just say one thing. But playing off of some of the other discussion that we had, I think the opportunity to continue to I guess democratize expertise and I want any woman receiving a mammogram in a community OBGYN or family medicine practice that would benefit from the same level of technology and expertise that she would receive at a major academic medical center, and that’s what we are doing every day when we’re able to sort of implement our program with our tools and our radiologists and our AI into practices that may be in rural areas that need better access.
[00:28:49] So that’s just kind of bread and butter, something I am really proud of and excited about. And then really the ability, a- as I mentioned earlier, to use the mammogram to find even more risk factors for the patient. So and that’s two fact, twofold. One even risk factors not relevant to s- breast cancer, but other risks like cardiovascular risks but also just more personalized risk for the patient.
[00:29:16] So there’s now some FDA-cleared AI technology to use the breast composition and images to determine the patient’s five-year risk of breast cancer in a very different way than we ever have before. And so I think soon Guidelines will change to be more personalized and tailored versus every woman gets a mammogram a- at this cadence in this way.
[00:29:43] It might very much change and be very different person to person, as it should. So just lots to be excited about in this space and the pace of change that’s happening.
[00:29:55] Adam Walker: Thanks for listening to Real Pink, a weekly podcast by Susan G. Komen. For more episodes, visit realpink.komen.org, and for more on breast cancer, visit komen.org.
[00:30:04] Make sure to check out @SusanGKomen on social media. I’m your host, Adam. You can find me on Twitter @AJWalker or on my blog, adamjwalker.com.