PathAI to Present Machine Learning-based Quality Control Tool for HER2 Testing in Breast Cancer at the American Society of Clinical Oncology Virtual Scientific Program 2021
Collaboration to produce a machine learning (ML)-based tool that automatically analyzes images of HER2 immunohistochemistry (IHC)-stained tissue slides and reports metrics that can be used to monitor the quality of HER2 IHC testing in clinical trials
Boston, Massachusetts - PathAI, a global leader of AI-powered technology applied to pathology, today announced that new data highlighting a quality control tool for HER2 testing in digital pathology images captured in clinical trials will be presented in the American Society of Clinical Oncology (ASCO) Virtual Scientific Program 2021, held from June 4-8, 2021. These results will be shared in the poster presentation, Machine learning models to quantify HER2 for real-time tissue image analysis in prospective clinical trials (Abstract #3061), in the session, Developmental Therapeutics —Molecularly Targeted Agents and Tumor Biology.
Together, PathAI, AstraZeneca (LSE/STO/Nasdaq: AZN) and Daiichi Sankyo Company, Limited have developed ML-based models for the automated quantification of HER2 IHC images in breast cancer tissue. Expression of HER2, a protein localized in the cell membrane, is typically assessed by pathologists to evaluate patient eligibility for anti-HER2 targeted therapies. ML-based models trained to identify and quantify tumor histology features can provide highly accurate and reproducible scores that are highly concordant with manual pathology.
The PathAI HER2 models were developed to generate HER2 scores consistent with the 2018 ASCO/CAP HER2 scoring guidelines. The models also produce metrics that reflect the quality of HER2 testing, such as the area and number of tumor cells, the presence of ductal carcinoma in situ (DCIS), background staining and artifact content. In a test set including diverse tissue-types across a wide range of breast cancer types, ML quantification of HER2 was consistent with manual scores from a consensus of pathologists (ICC 0.88, 95% CI 0.82-0.92). ML scores were even more closely aligned with pathologist scores after further training to learn pathologist scoring methods (ICC 0.91, 95% CI 0.89-0.94). By providing consistent, automated HER2 IHC image analysis, PathAI ML models can provide real-time QC read-outs enabling identification of drifts or inconsistencies in HER2 testing data and images captured during clinical trials.
PathAI's broad approach towards integrating AI-powered tools into oncology clinical trial workflows is also represented by a separate study that PathAI is presenting at ASCO (Abstract #106). Both presentations are examples of how AI can enhance pathologist performance by generating accurate and reproducible clinically relevant scores that can be scaled to levels that are currently unachievable.
PathAI is a leading provider of AI-powered research tools and services for pathology. PathAI’s platform promises substantial improvements to the accuracy of diagnosis and the efficacy of treatment of diseases like cancer, leveraging modern approaches in machine and deep learning. Based in Boston, PathAI works with leading life sciences companies and researchers to advance precision medicine. To learn more, visit pathai.com.