Pathology imaging informatics for quantitative analysis of whole-slide images

Abstract

Objectives: With the objective of bringing clinical decision support systems to reality, this article reviews histopathological whole-slide imaging informatics methods, associated challenges, and future research opportunities.

Target audience: This review targets pathologists and informaticians who have a limited understanding of the key aspects of whole-slide image (WSI) analysis and/or a limited knowledge of state-of-the-art technologies and analysis methods.

Scope: First, we discuss the importance of imaging informatics in pathology and highlight the challenges posed by histopathological WSI. Next, we provide a thorough review of current methods for: quality control of histopathological images; feature extraction that captures image properties at the pixel, object, and semantic levels; predictive modeling that utilizes image features for diagnostic or prognostic applications; and data and information visualization that explores WSI for de novo discovery. In addition, we highlight future research directions and discuss the impact of large public repositories of histopathological data, such as the Cancer Genome Atlas, on the field of pathology informatics. Following the review, we present a case study to illustrate a clinical decision support system that begins with quality control and ends with predictive modeling for several cancer endpoints. Currently, state-of-the-art software tools only provide limited image processing capabilities instead of complete data analysis for clinical decision-making. We aim to inspire researchers to conduct more research in pathology imaging informatics so that clinical decision support can become a reality.

Keywords: cancer prediction; computer-aided diagnosis; decision support systems; pathology imaging informatics; whole-slide images.

Figure 1

An example clinical decision support system for quantitative analysis of whole-slide images (WSI) of tissue biopsy samples. This system has the following key components: quality control to ensure only high-quality data are processed, image description to convert WSI into quantitative features, prediction modeling to develop quantitative diagnostic models, and exploratory analysis to interpret the image feature space. We include two case studies as examples of predictive modeling and exploratory analysis. ROI, region of interest.

Figure 2

Eliminating tissue-fold artifacts and pen marks in a whole-slide image of a NIH Cancer Genome Atlas ovarian serous carcinoma biopsy.

Figure 3

Normalization of color batch effects in ovarian samples provided by the NIH Cancer Genome Atlas.

Figure 4

Representation of a (A) NIH Cancer Genome Atlas whole-slide image (WSI) of a kidney renal clear cell carcinoma biopsy using various quantitative features extracted from (B) a single image tile. Quantitative features include pixel-level features, ie, (C) color histogram and (D) Gabor filter response; object-level features, ie, (E) segmented shapes and (F) graph-based topology; and semantic-level features, ie, (G) percentage of high-level clinical properties.

Figure 5

Role of region of interest (ROI) selection on the performance of whole-slide image (WSI)-based prediction models. (A) An example WSI. (B) Tiles in the tumor region (ROI) of the WSI highlighted with green boxes. Scatter plots between the prediction performance (area under the curve; AUC) of inner and outer loop of nested cross-validation for (C) models based on features from tissue tiles, including tumor and non-tumor (normal, necrosis, and stroma) regions; and for (D) tiles in the tumor region only.