HistoWAS: A Pathomics Framework for Large-Scale Feature-Wide Association Studies of Tissue Topology and Patient Outcomes

TL;DR

HistoWAS tackles the challenge of quantifying tissue spatial organization in whole-slide images and linking it to patient outcomes. It introduces a two-stage framework that combines a 72-object-level feature set with 30 spatial pathomic features derived from GIS-based spatial statistics, and uses a PheWAS-inspired mass-univariate regression with stringent multiple-testing correction. Applied to KPMP kidney data (385 WSIs from 206 patients), the method aggregates WSI-level features to patient-level phenotypes and identifies spatial patterns associated with interstitial fibrosis, providing interpretable digital biomarkers. The work delivers a scalable, open-source toolkit for large-scale pathomics-based association studies with potential to guide biomarker discovery and translational kidney disease research.

Abstract

High-throughput "pathomic" analysis of Whole Slide Images (WSIs) offers new opportunities to study tissue characteristics and for biomarker discovery. However, the clinical relevance of the tissue characteristics at the micro- and macro-environment level is limited by the lack of tools that facilitate the measurement of the spatial interaction of individual structure characteristics and their association with clinical parameters. To address these challenges, we introduce HistoWAS (Histology-Wide Association Study), a computational framework designed to link tissue spatial organization to clinical outcomes. Specifically, HistoWAS implements (1) a feature space that augments conventional metrics with 30 topological and spatial features, adapted from Geographic Information Systems (GIS) point pattern analysis, to quantify tissue micro-architecture; and (2) an association study engine, inspired by Phenome-Wide Association Studies (PheWAS), that performs mass univariate regression for each feature with statistical correction. As a proof of concept, we applied HistoWAS to analyze a total of 102 features (72 conventional object-level features and our 30 spatial features) using 385 PAS-stained WSIs from 206 participants in the Kidney Precision Medicine Project (KPMP). The code and data have been released to https://github.com/hrlblab/histoWAS.

Figures (2)

• Figure 1: An overview of the computational workflow for the HistoWAS framework. The process begins with two primary inputs: feature data extracted from WSIs and pathologist assigned histopathological descriptor scores. The framework then calculates a set of conventional object-level and spatial topology features. A high-throughput association analysis is subsequently performed to test the correlations between each feature and the clinical outcome. Finally, the results are interpreted and displayed using a suite of specialized visualizations.
• Figure 2: Conceptual workflow for spatial correlation feature extraction. (Top Right) The DBSCAN algorithm is used to identify and isolate distinct tissue clusters from the raw object point cloud, defining the region of interest for analysis. (Top Left & Middle) Visual comparison between the cumulative L-function (counting all points within a radius r) and the non-cumulative Pair Correlation g-function (counting points in an annulus at a specific distance r). (Bottom) Demonstration of the distance-based function calculation. From a set of source points (S1,S2...Si), events (e.g., tubule centroids) are counted within expanding radii (e.g., 25, 50, 75 $\mu$m). These counts are then aggregated and normalized to compute a final feature value (L) for each distance, quantifying the tissue's micro-architectural pattern.