From Screening to Therapy: How AI Is Transforming Breast Cancer Detection and Treatment Decisions
Artificial intelligence is rapidly transforming breast cancer care, from improving early detection through advanced imaging tools to supporting more personalized treatment decisions. In this discussion, Matthew Kurian, MD, and Jasmin Hundal, MD, explore how AI is being applied in radiology, pathology, and emerging clinical models to better stratify patient risk and guide therapies such as chemotherapy, radiation, and targeted treatments. While adoption is still evolving, these innovations are laying the groundwork for more precise, individualized care.
How is artificial intelligence currently improving early detection and risk stratification in breast cancer?
Dr. Kurian: I think, when you look at the use of artificial intelligence within oncology, amongst all of the current FDA-approved machine learning devices, radiology is really the first one that you really think of in terms of early detection within mammography with lung cancer screening, other things. That’s really where the majority of these devices are created at this point in time, which is really important as well, too. And you know, we now have tools like Clairity, you know, coming in onto the market and more to come, I think within the breast space, too.
What’s your thought?
Dr. Hundal: Yeah, no I agree with that. And also, like how radiology and pathology, like in the integration of both things in terms of like prognostic and risk stratification, since you’ve seen like Artera AI kind of like combining both in the prostate cancer world and also in breast cancer now, right? It got recently in FDA program, I believe, right? So like not yet an NCCN, but like hopefully, down the road with more of like group projects and like research that goes on, go see more integration.
Dr. Kurian: I think personalized screening, I think, is gonna become very important. So, the WISDOM trial was one that really looked at this first approach of personalized screening, and there’s still things that we need to work out from that, but I think it really laid a foundation for what we can accomplish in the future, and attaching artificial intelligence to an approach like that, I think, will be quite powerful.
Dr. Hundal: Yeah, and it can also help our patients like who needs every six months an MRI, mammogram, and kind of like what kind of path they’re on.
Dr. Kurian: So, I think in terms of how some of these artificial intelligence tools can utilize the practice in terms of surgery, in terms of radiation, in terms of chemotherapy decisions, if you look again at the number of machine learning approved devices, it’s really radiology, radiation oncology, and then actually pathology in terms of the number of devices approved as well, too. And actually, the number of tools for personalized decision-making is actually a small minority of this at this point in time.
But now, we have, you know, Ataraxis is one that’s coming, you know, that will help make treatment decisions about chemotherapy. We have Artera AI now, you know, recently FDA approved as another option. And I think there’s tons and tons of models that are being presented at ASCO, ESMO, San Antonio, too, that are now gonna create different models that I think are gonna become more personalized, too, to help make our decisions really about chemotherapy.
But I think we need to step beyond that and really start to think about, you know, how can we choose CDK4/6 inhibitors, for example, of whether patients are gonna benefit from those or not. How can we choose whether or not someone benefits from an oral SERD versus a TT 846 in the adjuvant setting, potentially, if RA is approved, right?
So, I think that there’s a lot of implications where I think AI can be quite helpful in empowering some of these decisions, too. But it’s still a little bit early to say that they’re going to be changing our practice at this point in time.
Dr. Hundal: Yeah, I agree. I think the one thing I’m super excited about is making treatment decision in our like adjuvant or like setting, especially in the second line setting in our metastatic, for example, where we decide like what are we gonna do like with the and whatnot. So I feel like having predictive markers, like using AI tools will be super helpful. And I think like we just need more, again, more studies and more pattern, but I feel like I’m also excited that AI accelerates those things. So, hopefully, we can take, you know, opportunity of those and like perform modeling.
Dr. Kurian: I think we’ve always looked at clinical pathologic features when making some of these decisions where we’ve learned, you know, in disease states like invasive lobular disease and other things that these things are important, but also, we need to really understand the biology, what’s driving those decisions, too, and whether patients benefit from chemotherapy or not. And I think this is where AI can really fill that gap exactly, and help build models that will tell us exactly when patients may benefit from chemotherapy, and when they may not. And then, you know, other tools, oral CDK4/6, targeted treatments, you know, and what the right combination may be is, too.
And I think it’s really gonna enhance drug discovery in a very rapid fashion that our days of waiting for 10 years for a drug to come to market. Hopefully, we’ll be overusing artificial intelligence and be able to speed up the drug discovery process. And hopefully, cut that in half, or even more.
Dr. Hundal: Yeah, and especially also escalating or de-escalating care for our patients, ’cause we have all these great drugs in the adjuvant setting. But like choosing like who’s the perfect patient for what, I think that’s key as well.
Dr. Kurian: Correct, talking about de-escalation, I think that, you know, there is no current NCCN-endorsed FDA model within breast cancer at this point, but I always like to point in terms of what is approved. There’s solid tumor types, too. And you look at the Artera AI test in prostate cancer really serves as an excellent model for us to make future decisions, future models, future tools that help us de-escalate hormonal therapy for patients in that setting. And also, help us understand exactly how do we validate these exactly. The Artera AI test was retrospectively validated, using cooperative group trials as well, too. And how do we build these tools in a way that is safe and that physicians trust as well, and patients trust most importantly, too.
And that’s really the key part of building these tools is we have to build them with good intention and make sure that, you know, that we have proper validation. The data sets are very representative of the population that we see, and they’re built not on legacy data sets, but data sets that really reflect the current standard of care is really important.
Dr. Hundal: And I feel like a physician oversight is so important. Cause I feel like we need the physician’s voice in these things as well.
Is TCHP Still the Standard? New Data Reshaping Early-Stage HER2+ Treatment
ESR1 Testing in ER+ mBC: From Trial Evidence to Real-World Practice
From Screening to Therapy: How AI Is Transforming Breast Cancer Detection and Treatment Decisions
180 Mount Airy Rd., Suite 205
Basking Ridge, NJ 07920
About Us
© 2026 Formedics, LLC. All Rights Reserved.
Leave a Reply