Automatic discrimination of neoplastic epithelium and stromal response in breast carcinoma

• Published in: Computers in biology and medicine Vol. 110; pp. 8 - 14
• Main Authors: Salvi, Massimo, Molinari, Filippo, Dogliani, Natalia, Bosco, Martino
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01-07-2019; Elsevier Limited
• Subjects: Algorithms; Annotations; Artificial neural networks; Automatic segmentation; Automation; Biomarkers; Breast cancer; Breast carcinoma; Cancer; Classification; Computer vision; Computer-aided image analysis; Datasets; Digital pathology; Enumeration; Epithelium; Histology; Invasiveness; Lymphocytes; Morphology; Neural networks; Nuclei (cytology); Quantitative analysis; Sensitivity analysis; Stroma; Tumors; Digital pathology; Breast carcinoma; Automatic segmentation; Artificial neural networks; Computer-aided image analysis
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2019.05.009

In breast carcinoma, epithelial–stromal interactions play a pivotal role in tumor formation and progression, and it must be accurately assessed for a correct extraction of predictive and prognostic biomarkers. Evaluation of preoperative (baseline) neoplasia/stroma ratio and the enumeration of tumor infiltrating lymphocytes (TIL) represent only two conditions in which precise discrimination of cancer epithelium and stromal reaction are relevant. However, subjectivity and expertise of the operators may lead to different degrees of assessment.
In this paper, we present a fully automated method for the discrimination between neoplastic epithelium and stromal reaction in breast carcinoma. Starting from cell nuclei, the proposed method implements computer vision strategies to split the neoplastic epithelium tissue from the stromal reaction.
The algorithm is tested on 100 H&E (hematoxylin and eosin) stained images representative of 10 different cases of invasive carcinoma. The algorithm performance in the detection of neoplastic epithelium (compared to manual annotations by an expert pathologist) gave a F1SCORE of 0.8894 and mean jaccardINDEX of 0.8481.
To the best of our knowledge, the proposed method is the first fully automated algorithm for the discrimination between neoplastic epithelium and stromal reaction in H&E stained images of breast tissue. The proposed approach paves the way for an automated and quantitative analysis of predictive and prognostic biomarkers in breast carcinoma.
•An adaptive and fully automated method is presented for the discrimination between neoplastic epithelium and stromal response in breast histopathological images.•The distinction between neoplastic epithelium and stromal reaction is essential in the correct extraction of predictive and prognostic biomarkers for breast carcinoma.•The proposed approach paves the way for quantitative analysis of predictive and prognostic biomarkers in breast carcinoma.