Explainable Pathomics Feature Visualization via Correlation-aware Conditional Feature Editing

TL;DR

This paper introduces a manifold-aware diffusion framework that enables controllable, biologically plausible editing of pathomics features in digital pathology images, improving interpretability and realism.

Contribution

It proposes a novel regularization method within a VAE to account for feature correlations, enhancing feature editing realism in pathomics analysis.

Findings

Outperforms baseline methods in feature editing accuracy

Maintains structural coherence during feature manipulation

Ensures edited features stay within the biological data manifold

Abstract

Pathomics is a recent approach that offers rich quantitative features beyond what black-box deep learning can provide, supporting more reproducible and explainable biomarkers in digital pathology. However, many derived features (e.g., "second-order moment") remain difficult to interpret, especially across different clinical contexts, which limits their practical adoption. Conditional diffusion models show promise for explainability through feature editing, but they typically assume feature independence**--**an assumption violated by intrinsically correlated pathomics features. Consequently, editing one feature while fixing others can push the model off the biological manifold and produce unrealistic artifacts. To address this, we propose a Manifold-Aware Diffusion (MAD) framework for controllable and biologically plausible cell nuclei editing. Unlike existing approaches, our method regularizes feature trajectories within a disentangled latent space learned by a variational auto-encoder (VAE). This ensures that manipulating a target feature automatically adjusts correlated attributes to remain within the learned distribution of real cells. These optimized features then guide a conditional diffusion model to synthesize high-fidelity images. Experiments demonstrate that our approach is able to navigate the manifold of pathomics features when editing those features. The proposed method outperforms baseline methods in conditional feature editing while preserving structural coherence.