Compound computer vision workflow for efficient and automated immunohistochemical analysis of whole slide images

Abstract

AimsImmunohistochemistry (IHC) assessment of tissue is a central component of the modern pathology workflow, but quantification is challenged by subjective estimates by pathologists or manual steps in semi-automated digital tools. This study integrates various computer vision tools to develop a fully automated workflow for quantifying Ki-67, a standard IHC test used to assess cell proliferation on digital whole slide images (WSIs).MethodsWe create an automated nuclear segmentation strategy by deploying a Mask R-CNN classifier to recognise and count 3,3′-diaminobenzidine positive and negative nuclei. To further improve automation, we replaced manual selection of regions of interest (ROIs) by aligning Ki-67 WSIs with corresponding H&E-stained sections, using scale-invariant feature transform (SIFT) and a conventional histomorphological convolutional neural networks to define tumour-rich areas for quantification.ResultsThe Mask R-CNN was tested on 147 images generated from 34 brain tumour Ki-67 WSIs and showed a high concordance with aggregate pathologists’ estimates (n=3 assessors; y=0.9712x−1.945, r=0.9750). Concordance of each assessor’s Ki-67 estimates was higher when compared with the Mask R-CNN than between individual assessors (ravg=0.9322 vs 0.8703; p=0.0213). Coupling the Mask R-CNN with SIFT-CNN workflow demonstrated ROIs can be automatically chosen and partially sampled to improve automation and dramatically decrease computational time (average: 88.55–19.28 min; p<0.0001).ConclusionsWe show how innovations in computer vision can be serially compounded to automate and improve implementation in clinical workflows. Generalisation of this approach to other ancillary studies has significant implications for computational pathology.

Automatic CD30 scoring method for whole slide images of primary cutaneous CD30+lymphoproliferative diseases

“…The application of artificial intelligence algorithms in IHC and clinical biomarker scoring has always been a research hotspot 12 13. However, this approach is still challenging in clinical practice.…”

“…Digital pathology assists pathologists in making diagnoses, simplifying complex and time-consuming tasks, and reducing the risks and biases caused by various intraobserver and interobserver factors in the pathological diagnosis process 10–12. Several methods for the automated evaluation of immunohistochemical markers by machine learning and deep learning have been widely used13–22 (online supplemental table 1).…”

HAVOC: Small-scale histomic mapping of cancer biodiversity across large tissue distances using deep neural networks

“…In later analyses, we also incorporated a companion automated deep learning approach for quantification of cellular density across multiple HAVOC partitions less reliant on manual delineation of regions of interest. Briefly, we optimized Mask R-CNN ( 48 ) using the Detectron2 library to serve as a nuclear instance segmentation tool similar to what we previously described ( 49 ). We extracted 128 × 128 pixel image patches from different WSIs containing approximately 20 to 30 cells per patch.…”

An overview of the detection of bovine respiratory disease complex pathogens using immunohistochemistry: emerging trends and opportunities

“…Methodologies such as these enable identification and subcellular antigen location. 64 , 89 To quantitatively measure intensity, 84 , 119 as well as target protein quantification and distribution, 59 high-resolution images of IHC-stained tissue sections can be captured using digital imaging technology and analyzed with computational algorithms. Furthermore, computer-assisted detection systems using machine-learning algorithms are being developed to improve IHC detection accuracy and efficiency by identifying patterns and features that may be overlooked by the human eye.…”