PathAI logo

PathAI Announces Research Presentations at the San Antonio Breast Cancer Symposium

Demonstrating the use & performance of two AI-powered pathology products launched in 2023: PathExploreTM and AIM-HER2TM  .

 Boston, MA – December 5, 2023 PathAI, a leading technology company which combines AI-powered pathology solutions with end-to-end central pathology and histology services, today announced it will present four research presentations underscoring the value and potential of PathExplore’s1 human interpretable features (HIFs), and one research presentation demonstrating the ability of the AIM-HER22 product to accurately predict HER2 scores, including in HER2-low cases, at the San Antonio Breast Cancer Symposium on December 5-9, 2023, in San Antonio, TX.

“This research builds on our past successes and proceeds to show that PathAI’s algorithms continue to further our understanding of cancer biology and disease mechanisms,” said Andy Beck, M.D., Ph.D., co-founder and chief executive officer of PathAI. “At PathAI, we continue to demonstrate the clinical potential for these AI pathology tools, and we’re confident that it will ultimately translate to advancements in precision medicine and patient outcomes.”

H&E whole slide image with tissue regions and cell types identified and characterized by PathExplore
H&E whole slide image of breast cancer with tissue regions and cell types identified and characterized by PathExplore

Poster #PO1-07-01: Artificial Intelligence-Based Prediction of Oncotype DX Score from whole slide images using human-interpretable features and breast biomarkers

The Oncotype DX Breast Recurrence Score assay (ODX) is a commonly used genomic test for patients with estrogen receptor (ER)-positive, HER2-negative, early-stage invasive breast cancer. While ODX predicts patients’ recurrence risk and benefit from chemotherapy, it is time-consuming and expensive.

Using human-interpretable features extracted from H&E whole slide images with the PathExplore Breast model, together with manually assessed ER, PR, HER-2 scores, and tumor stage, PathAI, in collaboration with researchers from the Cleveland Clinic Foundation, developed interpretable models for predicting ODX recurrence scores using histological features and clinical covariates. The models achieved a level of performance comparable to previous deep-learning approaches, while also revealing the contribution of features such as cancer cell density, immune cell density, and variations in cancer nuclear shape and color to the ODX score prediction. Using readily available information, the model has the potential to be a convenient screening tool for patient stratification, which may lead to better patient care, lower costs, and faster treatment.

Presenting Author: Nhat Le
Time/Location: Wednesday, December 6, 12:00 PM – 2:00 PM, Poster Session 1
Poster #PO3-07-04: Prediction of PAM50 molecular subtypes from H&E-stained breast cancer specimens using tumor microenvironment features and additive multiple instance learning models

PAM50, a 50-gene signature, classifies breast cancers into one of five subtypes and has emerged as a key prognostic indicator influencing treatment decisions. There is growing interest in bridging the gap between expression-based metrics and histopathology. In this poster, PathAI describes a computer vision-based approach to predict PAM50 classification using H&E-stained whole slide images (WSIs).

Two separate machine learning approaches were developed to predict PAM50 subtypes from WSIs. Both models performed well in predicting Basal, Luminal A, Luminal B, and Luminal (A+B) subtypes, and provide biological interpretability not found in typical black-box models. The results support the notion that AI-powered digital pathology can accurately and reproducibly perform molecular-based classification tasks such as predicting PAM50 classifications using WSIs, suggesting a more efficient path toward clinically relevant breast cancer characterization.

Presenting Author: Nhat Le
Time/Location: Thursday, December 7, 12:00 PM – 2:00 PM CT, Poster Session 3
Poster #PO1-15-01: Digital Pathology Models reveal Case-Specific Characteristics of the Tumor Microenvironment

Tumor microenvironments are complex three-dimensional structures, but often only single histological sections are used to glean biological information for research purposes or in some diagnostic practices. These conclusions can also be affected by technical factors in slide preparation and digitization.

In this research, PathAI applied PathExplore, a suite of digital pathology models to identify cell and tissue substances as well as their spatial relationships, to whole-slide images of breast cancer tumors and quantified intra- and inter-tumor heterogeneity in human-interpretable features of the tumor microenvironment.

The results reveal biological and technical variability that can inform selection and interpretation of biomarkers derived from single slides. The ability to uniquely identify slides from the same case additionally demonstrates the technical robustness of digital pathology models for yielding quantitative insights into tumor biology.

Presenting Author: Ylaine Gerardin
Time/Location: Wednesday, December 6, 12:00 PM – 2:00 PM, Poster Session 1
Poster #PO2-14-12: Accurate quantification of slide-level HER2 scores in breast cancer using a machine-learning model, AIM-HER2 Breast Cancer

The HER2 biomarker is critical in determining a breast cancer treatment course. To aid accurate interpretation of HER2-stained breast cancer specimens, PathAI developed the AIM-HER2 machine learning model, which demonstrates that the model performs comparably to pathologists in the prediction of HER2 scores (0-3+), including HER2-low. Comparable concordance between AIM-HER2 and pathologists was also observed irrespective of the HER2 antibody clone or slide scanner used.

AIM-HER2 was developed using “Additive Multiple Instance Learning,” a novel technique which yields highly explainable results as slide-level scores and heatmaps, thereby supporting future use in AI-assisted pathologist workflows.

Presenting Author: John Shamshoian
Time/location: Wednesday, December 6, 5:00 – 7:00 PM CT, Poster Session 2

For more information, visit our poster sessions. For BioPharma opportunities, reach out to meet with our team at [email protected]. To get in touch with our Digital Diagnostics team, reach out to [email protected]. For Academic Research opportunities, reach out to [email protected] and [email protected]

About PathAI

PathAI is a recognized leader in the biopharma partnering space, uniquely combining AI-powered pathology solutions with end-to-end central pathology and histology services. The company supports biopharma partners in executing clinical trials where pathology-based endpoints, biomarker classification, and/or superior histology quality are critical to successfully gauging therapeutic efficacy, accelerating drug development for complex diseases. PathAI has already supported multiple Phase 2 clinical trials in NASH, IBD, and breast cancer, as well as oncology neoadjuvant trials, and is now expanding into larger scale global Phase 3 studies, as well as additional indications.
PathAI provides a fully integrated approach to clinical trials, enabling pharma partners to leverage the power of AI without the heavy lift of implementation. This helps reduce the impact of challenges associated with clinical trials, including unreliable turnaround times, variable histology, stain or scan/digitization quality, and challenging assessment of histological endpoints. The lab offers all major immunohistochemistry staining platforms, with flexible workflows across different scanners, stains, and biopsy types, which improves the quality of clinical trials. Services include access to PathAI’s extensive network of over 500 US Board Certified pathologists to perform high quality reading with rapid turnaround time in a cost-effective manner, plus seamless integration of PathAI's advanced AI-solutions to ensure high-quality, reproducible results with every scan. For more information, please visit

1, 2 PathExplore and AIM-HER2 are For Research Use Only. Not for use in diagnostic procedures.

Get In Touch

For media inquiries, please contact [email protected]
Path AI seating area overlooking Fenway Park in Boston