VIDEO: AI ‘here to stay,’ even in breast cancer

Editor’s note: This is an automatically generated transcript, which has been slightly edited for clarity. Please notify editor@healio.com if there are concerns regarding accuracy of the transcription.

Artificial intelligence is here to stay. And it's impacting, of course, many aspects, not only in medicine, but breast cancer is not the exception. And for several years, you know, even before the emergence of the popular, you know, ChatGPT or the bots, artificial intelligence have been heavily researched, particularly in digital imaging analysis. There have been some implementations already, actually, that assist pathologists to evaluate tissue biopsies. There are certain routine assays or biomarkers that we, you know, use in clinics such as Ki-67 that now have certain artificial intelligence-based tools that allow more automated evaluation of that. And you know, that not only saves time and effort and resources because obviously, the pathologists can make a determination much faster. But also, you know, many times machines are better than the human eye and they can identify certain features that may be invisible to the human eye and it can be a lot more homogeneous, right? So, it's decreasing the inter-observer variability of some of these assays.

So, I think we're going to continue to see a lot of efforts to leverage artificial intelligence to analyze tumor tissue biopsies, not only for proliferation assays like Ki-67 but also to evaluate features of the immune system like tumor-infiltrated lymphocytes. As many know in the field, we've also had a lot of changes in the ways that we interpret HER2 assessment, and there's some efforts on trying to use artificial intelligence tools to detect very low levels of HER2 and try to make the assessment a little more homogeneous. I think artificial intelligence is going to make a huge impact in the pathology community.

In addition to that, we've had a lot of very interesting tools analyzing radiographic images. You know, mammograms, MRIs, and even, you know, CT scans where the algorithms and the digital tools are able to, again, pick up certain features on imaging that we have traditionally not been able to identify under our normal analysis. And these tools may help optimize the suspicion. You know, if mammogram, for example, shows an abnormality, these tools may help assist to predict the risk of this being a malignant tumor versus a benign lesion. And there's even certain tools that are able to pick up certain features that are associated with certain mutations, just from the radiographic image and even predict response. So very exciting time, I think, particularly in the diagnostic area, right? So in mammograms, imaging, and pathology.

But there is also in the research side, you know, to try to identify new drugs. Something that is very exciting as well is that artificial intelligence tools are being used to more rapidly screen potential compounds or potential drugs that may be better able to predict the likelihood of success in a future clinical trial or in future studies. Previously, a lot of this had to be done somewhat manually and in a very slow process. Now, AI is really accelerating that screening process.

And then lastly, you know, we anticipate on the clinical side that in the future, there may be certain resources to help clinicians make decisions as well. As you know, we live in an era of information overload and we have so much research being developed every day that for the busy clinician, it's many times difficult to digest. And this is where, you know, kind of your work becomes really important to help digest the load of literature. But we think there's going to be artificial intelligence-based tools that perhaps will be a point-of-care decision aid resource for clinicians practicing, you know, a very increasingly complex oncology clinical practice.