Developing image analysis pipelines of whole-slide images: Pre- and post-processing

Abstract

Deep learning has pushed the scope of digital pathology beyond simple digitization and telemedicine.. The incorporation of these algorithms in routine work flow is on the horizon and may be a disruptive technology, reducing processing time and increasing detection of anomalies. While the newest computational methods enjoy much of the press, incorporating deep learning into standard laboratory workflow requires many more steps than simply training and testing a model. Image analysis using deep learning methods often requires substantial pre-and post-processing order to improve interpretation and prediction. Similar to any data processing pipeline, images must be prepared for modeling and the resultant predictions need further processing for interpretation. Examples include artifact detection, color normalization, image subsampling or tiling, removal of errant predictions, etc. Once processed, predictions are complicated by image file size-typically several gigabytes when unpacked. This forces images to be tiled, meaning that a series of subsamples from the whole slide image are used in modeling. Herein we review many of these methods as they pertain to analysis of biopsy slides and discuss the multitude of unique issues that are part of the analysis of very large images.

Artificial intelligence as a tool for diagnosis in digital pathology whole slide images: A systematic review

“…Moreover, despite an academic consensus on certain approaches to color treatment or data augmentation, 47 many studies did not perform color normalization or data augmentation, for example, and most did not use model ensembles for inference, probably due to the high computational cost during inference when using multiple models.…”

Automated scoring of total inflammation in renal allograft biopsies

“…Artificial intelligence tools allow integration of digital methods into research and clinical pipelines. Applying these methods towards the analysis of histopathology has unique challenges such as image size 14 that distinguish this image analysis from the analysis of other biomedical images. Nevertheless, a few groups have created methods to compartmentalize and automatically score immunohistochemistry slides of kidney biopsies 15–18 .…”

Whole Slide Image Quality in Digital Pathology: Review and Perspectives

“…The pipeline is as generic as possible as it encapsulates any histology and cytology routines. A similar pipeline was proposed in [226]. This is illustrated in Figure 13.…”