TopoCellGen: Generating Histopathology Cell Topology with a Diffusion Model

TL;DR

TopoCellGen introduces a topology-aware diffusion framework for generating multi-class histopathology cell layouts by enforcing topological constraints via persistent homology. It combines a cell-count conditioned diffusion process with intra-class and inter-class topological regularizations and introduces TopoFD, a Fréchet-based metric on persistence diagrams, to quantify topological fidelity beyond traditional image-based metrics. The method is extended to layout-guided image generation and validated on BRCA-M2C and Lizard datasets, where it achieves superior FID, TopoFD, and downstream task performance, while reducing count errors. By aligning synthetic layouts with biologically meaningful topology and counts, the approach provides more realistic data augmentation and enhances interpretability for tumor microenvironment analyses.

Abstract

Accurately modeling multi-class cell topology is crucial in digital pathology, as it provides critical insights into tissue structure and pathology. The synthetic generation of cell topology enables realistic simulations of complex tissue environments, enhances downstream tasks by augmenting training data, aligns more closely with pathologists' domain knowledge, and offers new opportunities for controlling and generalizing the tumor microenvironment. In this paper, we propose a novel approach that integrates topological constraints into a diffusion model to improve the generation of realistic, contextually accurate cell topologies. Our method refines the simulation of cell distributions and interactions, increasing the precision and interpretability of results in downstream tasks such as cell detection and classification. To assess the topological fidelity of generated layouts, we introduce a new metric, Topological Frechet Distance (TopoFD), which overcomes the limitations of traditional metrics like FID in evaluating topological structure. Experimental results demonstrate the effectiveness of our approach in generating multi-class cell layouts that capture intricate topological relationships. Code is available at https://github.com/Melon-Xu/TopoCellGen.

Figures (8)

• Figure 1: Illustrations of intra-class distribution and the inter-class relationship across various cell types. Here we only highlight the tumor/epithelial, lymphocytes, and stromal cells.
• Figure 2: An overview of our method TopoCellGen. (a) denotes the overview workflow. (b) shows the details of $\mathcal{L}_{\text{count}}$, $\mathcal{L}_{\text{intra}}$ and $\mathcal{L}_{\text{inter}}$.
• Figure 3: Intuition of our proposed Topological Fréchet Distance. TCE indicates the Total Count Error.
• Figure 4: The overall pipeline of calculating the Topological Fréchet Distance. Take the lymphocyte as an example.
• Figure 5: The qualitative results of our proposed TopoCellGen. Columns $1$-$3$: BRCA-M2C dataset. Columns $4$-$6$: Lizard dataset. The cell types and their corresponding colors are shown on the right side of the figure.