Luckily, we have already a lot of AI tools which are commercially available. And these AI tools are available for different biomarkers. So we have for HER2, HER2-Low and HER2-Ultra-Low. So it has to be an AI tool that was validated for HER2-Low and HER2-Ultra-Low identification, not just HER2 in general, because during validation and the machine learning process, you have to have a data set enriched with HER2-Low and HER2-Ultra-Low breast cancers...
Luckily, we have already a lot of AI tools which are commercially available. And these AI tools are available for different biomarkers. So we have for HER2, HER2-Low and HER2-Ultra-Low. So it has to be an AI tool that was validated for HER2-Low and HER2-Ultra-Low identification, not just HER2 in general, because during validation and the machine learning process, you have to have a data set enriched with HER2-Low and HER2-Ultra-Low breast cancers. So this is challenging. So when purchasing or evaluating AI tools, you have to make sure that you are acquiring one that was validated specifically for HER2-Low and HER2-Ultra-Low breast cancers, not only HER2-positive. So we have different algorithms available commercially for HER2, HER2-Low and HER2-Low breast cancers. We also have for hormone receptors, estrogen and progesterone receptors, Ki67, PD-L1, and for not only biomarkers, for detection of lymph node metastases, for prostate cancer diagnosis, for breast cancer diagnosis, different lesions from benign to atypical to malignant. So all of these algorithms are commercially available and they are really useful because for HER2-Low and HER2-Ultra-Low detection, we see a lot of disagreement amongst pathologists and the sensitivity in manual scoring is lower. For PD-L1, also the schemes of interpretation are really challenging, having different categorizations according to the type of cancer and according to topography. And all these AI tools help us to be more objective and also to be more accurate and sensitive. So it’s really, really helpful in being more accurate and we have a more robust test when we come together the pathologists, a really well-trained pathologist with a robust validated AI tool and along with all mechanisms of quality control each pathology laboratory has to undergo different programs for quality control and also laboratory accreditation.
This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.
