1219P - Artificial intelligence-based breast cancer detection facilitates automated prognosis marker assessment using multiplex fluorescence immunohistochemistry

Background

Prognostic markers in routine clinical practice of breast cancer are often assessed using RNA based multi-gene panels that are depending on a fluctuating tumor purity. Multiplex fluorescence immunohistochemistry (mfIHC) holds the potential for improved risk assessment.

Methods

To enable automated prognosis marker quantification, we have developed and validated a framework for automated breast cancer detection involving three different artificial intelligence analysis steps and an algorithm for cell-distance analysis using BLEACH&STAIN multiplex fluorescence immunohistochemistry. Pan-cytokeratin (panCK) antibodies were used to detect epithelial cells and antibodies directed against Myosin and p63 were used to identify basal cells.

Results

The optimal distance between Myosin⁺ and p63⁺ basal cells and benign panCK⁺ cells was identified as 25 μm in breast cancer and used – combined with deep learning-based algorithms – to exclude benign glands from the analysis. Our framework discriminated normal glands from malignant glands with an accuracy of 98.4% (95% confidence interval [CI]: 97.4 – 99.3). The approach for automated breast cancer detection improved the predictive performance of several prognosis markers significantly (each p<0.05) and a comparison with manually assessed data using conventional brightfield immunohistochemistry showed a high concordance for a multitude of different prognosis marker such as PR, ER, GATA3, HER2, and PD-L1 (each p<0.0001).

Conclusions

The combined assessment of up to 5 markers in a prognosis score showed strong prognostic relevance (p<0.001) and was an independent risk factor in multivariate analysis (p=0.005). Thus, the data from this study show that automated breast cancer detection in combination with artificial intelligence-based analysis of multiplex fluorescence immunohistochemistry enables a rapid and reliable analysis of multiple prognostic parameters. The major advantage of this method is the analysis of malignant cells exclusively that cannot be achieved using RNA-based panel analysis.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

G. Sauter: Other, Personal, Other, The GATA3 (MSVA-450M), PD-L1 (MSVA-711R), PR (MSVA-570R), AR (MSVA-367R), ER (MSVA-564R), TROP2 (MSVA-733R), TOP2A (MSVA-802R), Myosin (MSVA-375R), panCK (MSVA-000R), Ki-67 (MSVA-267M) antibodies were provided by MS Validated Antibodies GmbH (owned by a family member of GS): MSVA. All other authors have declared no conflicts of interest.