No abstract

Breast cancer is the most common malignant disease worldwide, with over 2.26 million new cases in 2020. Its diagnosis is determined by a histological review of breast biopsy specimens, which can be labor-intensive, subjective, and error-prone. Artificial Intelligence (AI)—based tools can support cancer detection and classification in breast biopsies ensuring rapid, accurate, and objective diagnosis. We present here the development, external clinical validation, and deployment in routine use of an AI-based quality control solution for breast biopsy review. The underlying AI algorithm is trained to identify 51 different types of clinical and morphological features, and it achieves very high accuracy in a large, multi-site validation study. Specifically, the area under the receiver operating characteristic curves (AUC) for the detection of invasive carcinoma and of ductal carcinoma in situ (DCIS) are 0.99 (specificity and sensitivity of 93.57 and 95.51%, respectively) and 0.98 (specificity and sensitivity of 93.79 and 93.20% respectively), respectively. The AI algorithm differentiates well between subtypes of invasive and different grades of in situ carcinomas with an AUC of 0.97 for invasive ductal carcinoma (IDC) vs. invasive lobular carcinoma (ILC) and AUC of 0.92 for DCIS high grade vs. low grade/atypical ductal hyperplasia, respectively, as well as accurately identifies stromal tumor-infiltrating lymphocytes (TILs) with an AUC of 0.965. Deployment of this AI solution as a real-time quality control solution in clinical routine leads to the identification of cancers initially missed by the reviewing pathologist, demonstrating both clinical utility and accuracy in real-world clinical application.

Objective: Tumor HER2 expression is a key prognostic and treatment influencing factor in breast cancer. As with all immunohistochemistry (IHC) staining, visual interpretation of HER2 expression is subjective, which leads to intra- and inter-pathologist variability. Recent findings on the efficacy of HER2-targeted therapy on HER2-low patients raise the need for accurate and reproducible scoring. We developed a fully automated, artificial intelligence (AI) -based algorithm for HER2 scoring. The algorithm was based on ASCO/CAP 2018 guidelines and validated against rigorous ground truth (GT) established by multiple blinded expert pathologists. Methods: Algorithm development: We developed a solution that employs two steps: The first step consists of an ensemble of Deep Learning networks that process tissue regions and classify them as various tissue classes: Invasive cancer, Ductal Carcinoma In Situ (DCIS) and other morphologies. These networks were trained on slides that were automatically labeled by a separate AI system that analyzed the corresponding H&E slides and projected its findings to the HER2 IHC slides using a registration algorithm. To further enrich the training set, especially with rare and difficult cases, a team of 8 expert pathologists manually marked tissue areas and assigned them to one of the tissue classes. In total, the training set consisted of 6,400 manual annotations and 1,300 automatically-annotated slides, both collected from 9 laboratories and scanned using 3 different scanners. The second step is an ensemble of Object Detection networks that process only the regions classified as invasive cancer, detect the tumor cells within them, and classify their staining pattern (e.g., Not stained, Moderate incomplete, etc.). Finally, the detected cells are counted, and the ASCO/CAP guidelines are applied to derive the slide-level HER2 score. Validation: The validation set was comprised of 453 HER2 slides stained using the VENTANA anti-HER2/neu (4B5) Rabbit Monoclonal Primary Antibody as per manufacturer’s instructions. HER2 slides included biopsies and excisions with different breast cancer diagnoses (e.g., Infiltrating Ductal Carcinoma (IDC), Infiltrating Lobular Carcinoma (ILC), rare invasive subtypes, with and without DCIS) from 3 different laboratories. Ground truth was established by the consensus scores of a panel of 3 pathologists, who scored HER2 according to the guidelines without additional clinical considerations, such as scoring borderline 1+/2+ cases as 2+ to have additional tests performed. Results: The algorithm showed very high performance for detecting invasive cancer in HER2 tissue sections, with AUC of 0.967 (measured on 4-fold Cross-Validation classifying invasive vs. other regional classes). The algorithm demonstrated an overall accuracy of 80.3% for the HER2 scores when compared to the GT. When using different cutoffs for binary classification the resulting performance was: for 0 vs 1+/2+/3+ Kappa was 0.800; 0/1+ vs 2+/3+ Kappa was 0.728; for 0/1+/2+ vs 3+ Kappa was 0.954. The Quadratic Kappa between the AI score and the GT was 0.898, which is considered almost perfect. The performance of the AI was similar across the different laboratories and diagnoses(e.g. IDC, ILC). Conclusion: This study reports the successful development and independent validation of a fully automatic AI-based solution for accurate HER2 scoring in breast cancer. AI solutions, such as the one reported here, could be used as decision-support tools for pathologists in routine clinical practice, enhancing the reproducibility and consistency of HER2 scoring, thus enabling optimal treatment pathways and better patient outcomes. Accurate and automatic IHC scoring solutions can also contribute to the development of new prognostic, predictive and companion diagnostic tools. Citation Format: Yuval Globerson, Lilach Bien, Jonathan Harel, Giuseppe Mallel, Geraldine Sebag, Michel Vandenberghe, Craig Barker, Tsuyoshi Matsuo, Charo Garrido, Judith Sandbank, Chaim Linhart. A fully automatic artificial intelligence system for accurate and reproducible HER2 IHC scoring in breast cancer [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P6-04-05.

No abstract

Objective This study aimed to clinically validate the performance of a multi-feature AI-based solution on detection of invasive versus in situ carcinomas and in situ ductal carcinoma high grade from atypical ductal hyperplasia/low grade DCIS compared to rigorous ground truth (GT) established by multiple blinded expert pathologists in breast biopsies. Design Performance of the AI solution was prospectively tested on breast biopsies from two medical institutions in different geographies. AI results were compared against the ground truth (GT) established by consensus of two subspecialist breast pathologists. The study endpoints were detection of invasive carcinoma (IDC, ILC) and DCIS/ADH, including differentiating between different DCIS grades and ADH. ADH and DCIS Low Grade were pooled together because of similar clinical management when diagnosed on a biopsy. Results Six pathologists participated in the study and reported on 436 breast biopsies (841 H&E slides), including 156 invasive (including 31 rare subtypes), 135 DCIS/ADH and 145 benign diagnoses. The AI solution demonstrated high performance when compared with the GT with an AUC of 0.99 for the detection of invasive carcinoma (specificity and sensitivity of 93.6% and 95.5% respectively) and with AUC of 0.95 for the detection of DCIS/ADH. The AI solution differentiated well between subtypes/grades of invasive and in-situ cancers with an AUC of 0.97 for IDC vs. ILC and AUC of 0.92 for DCIS high grade vs. low grade/ADH, respectively. Only 11 (7%) cases had discrepancies on invasive diagnosis, 4 of these between invasive versus benign diagnosis encompassing one case on which the invasive component was only represented by rare lympho-vascular invasion, two cases of ILC (one with a diffuse pattern and the second in a case with granulomatous mastitis with multinucleated giant cells and hemosiderin-laden macrophages) and one rare case of tubular carcinoma surrounded by flat epithelial atypia and columnar cell lesions. Fifteen cases (11%) had discrepancies on DCIS/ADH diagnosis between the two specialist pathologists necessitating a third assessment by a specialist to establish GT. Six of these cases also necessitated a review on a multihead microscope to reach a consensus decision, since there was no majority even after 3 reviews. Conclusion This blinded multi-site study reports the successful clinical validation of a multi-feature AI-based solution in detecting and automatically imparting clinically relevant diagnostic parameters regarding invasive and in situ breast carcinoma, offering an important tool for computer-aided diagnosis in routine pathology practice. Citation Format: Anne Vincent-Salomon, Guillaume Bataillon, Alona Nudelman, Judith Sandbank, Anat Albrecht Shach, Lucie Thibault, Lilach Bien, Rachel Mikulinsky, Ira Krasnitsky, Ronen Heled, Chaim Linhart, Manuela Vecsler, Daphna Laifenfeld. A multi-feature AI-based solution for cancer diagnosis in breast biopsies: A prospective blinded multi-site clinical study [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD11-04.