Weakly-supervised deep learning models enable HER2-low prediction from H &E stained slides.

BACKGROUND: Human epidermal growth factor receptor 2 (HER2)-low breast cancer has emerged as a new subtype of tumor, for which novel antibody-drug conjugates have shown beneficial effects. Assessment of HER2 requires several immunohistochemistry tests with an additional in situ hybridization test if a case is classified as HER2 2+. Therefore, novel cost-effective methods to speed up the HER2 assessment are highly desirable. METHODS: We used a self-supervised attention-based weakly supervised method to predict HER2-low directly from 1437 histopathological images from 1351 breast cancer patients. We built six distinct models to explore the ability of classifiers to distinguish between the HER2-negative, HER2-low, and HER2-high classes in different scenarios. The attention-based model was used to comprehend the decision-making process aimed at relevant tissue regions. RESULTS: Our results indicate that the effectiveness of classification models hinges on the consistency and dependability of assay-based tests for HER2, as the outcomes from these tests are utilized as the baseline truth for training our models. Through the use of explainable AI, we reveal histologic patterns associated with the HER2 subtypes. CONCLUSION: Our findings offer a demonstration of how deep learning technologies can be applied to identify HER2 subgroup statuses, potentially enriching the toolkit available for clinical decision-making in oncology.

Prognostic value of Maspin protein level in patients with triple negative breast cancer.

The search for prognostic markers in breast cancer has bumped into a typical feature of these tumors, intra and intertumoral heterogeneity. Changes in the expression profile, localization of these proteins or shedding to the surrounding stroma can be useful in the search for new markers. In this context, classification by molecular subtypes can bring perspectives for both diagnosis and screening for appropriate treatments. However, the Triple Negative (TN) subtype, which is already the one with the worst prognosis, lacks appropriate and consistent molecular markers. In this work, we analyzed 346 human breast cancer samples in tissue microarrays (TMA) from cases diagnosed with invasive breast carcinoma to assess the expression and localization pattern of Maspin and their correlation with clinical parameters. To complement our findings, we also used TCGA data to analyze the mRNA levels of these respective genes. Our data suggests that the TN subtype demonstrates a higher level of cytoplasmic Maspin compared to the other subtypes. Maspin transcript levels follow the same trend. However, TN patients with lower Maspin expression tend to have worse overall survival and free-survival metastasis rates. Finally, we used Maspin expression data to verify possible relationships with the clinicopathological information of our cohort. Our univariate analyses indicate that Maspin is related to the expression of estrogen receptor (ER) and progesterone receptor (PR). Furthermore, Maspin expression levels also showed correlation with Scarff-Bloom-Richardson (SBR) parameter, and stromal Maspin showed a relationship with lymph node involvement. Our data is not consistently robust enough to categorize Maspin as a prognostic marker. However, it does indicate a change in the expression profile within the TN subtype.

The breast cancer immunophenotype of TP53-p.R337H carriers is different from that observed among other pathogenic TP53 mutation carriers.

Germline TP53 mutations are associated with Li-Fraumeni syndrome, an autosomal dominant disorder characterized by a predisposition to multiple early-onset cancers including breast cancer (BC), the most prevalent tumor among women. The majority of germline TP53 mutations are clustered within the DNA-binding domain of the gene, disrupting the structure and function of the protein. A specific germline mutation in the tetramerization domain of p53, p.R337H, was reported at a high frequency in Southern and Southeastern Brazil. This mutation appears to result in a more subtle defect in the protein, which becomes functionally deficient only under particular conditions. Recent studies show that the BC phenotype in TP53 mutation carriers is often HER2 positive (63-83%). Considering that the immunophenotype of BC among p.R337H carriers has not been reported, we reviewed immunohistochemistry data of 66 p.R337H carriers in comparison with 12 patients with other non-functional TP53 germline mutation. Although 75% of carriers of these mutations showed significant HER2 overexpression (3+), corroborating previous studies, only 22.7% of p.R337H patients had BC overexpressing HER2. These results reinforce the notion that different germline mutations in TP53 may predispose to BC via different mechanisms.