Development and Validation of Fully Automatic Deep Learning-Based Algorithms for Immunohistochemistry Reporting of Invasive Breast Ductal Carcinoma

TL;DR

A deep learning-based semi-supervised trained, fully automatic, decision support system (DSS) for IHC scoring of invasive ductal carcinoma that automatically detects the tumor region removing artifacts and scores based on Allred standard and can be used to develop DSS for other cancer types.

Abstract

Immunohistochemistry (IHC) analysis is a well-accepted and widely used method for molecular subtyping, a procedure for prognosis and targeted therapy of breast carcinoma, the most common type of tumor affecting women. There are four molecular biomarkers namely progesterone receptor (PR), estrogen receptor (ER), antigen Ki67, and human epidermal growth factor receptor 2 (HER2) whose assessment is needed under IHC procedure to decide prognosis as well as predictors of response to therapy. However, IHC scoring is based on subjective microscopic examination of tumor morphology and suffers from poor reproducibility, high subjectivity, and often incorrect scoring in low-score cases. In this paper, we present, a deep learning-based semi-supervised trained, fully automatic, decision support system (DSS) for IHC scoring of invasive ductal carcinoma. Our system automatically detects the tumor region removing artifacts and scores based on Allred standard. The system is developed using 3 million pathologist-annotated image patches from 300 slides, fifty thousand in-house cell annotations, and forty thousand pixels marking of HER2 membrane. We have conducted multicentric trials at four centers with three different types of digital scanners in terms of percentage agreement with doctors. And achieved agreements of 95, 92, 88 and 82 percent for Ki67, HER2, ER, and PR stain categories, respectively. In addition to overall accuracy, we found that there is 5 percent of cases where pathologist have changed their score in favor of algorithm score while reviewing with detailed algorithmic analysis. Our approach could improve the accuracy of IHC scoring and subsequent therapy decisions, particularly where specialist expertise is unavailable. Our system is highly modular. The proposed algorithm modules can be used to develop DSS for other cancer types.

Figures (11)

• Figure 1: The stain classification model’s stain classification performance of four different stain categories. (a) The performance of a DNN model (approach A1), and (b) the performance of CMYK (approach A2) based stain classification method.
• Figure 2: Intra-observer results comparison on validation slides for (a) ER stain, (b) PR stain, and (c) Ki67 stain.
• Figure 3: Highlighting the effect of artifacts on tissue on nuclei segmentation and stain classification. (a) A snapshot of the tissue region from ER WSI. (b) The nuclei segmentation and stain classification results.
• Figure 4: Illustration of the ground-truth generation process.
• Figure 5: Examples of tissue sub-region annotations to be used for ROI deep learning model training and validation.