CPathOGen: Spatially and Morphologically Controlled H&E Counterfactuals for Probing Pathology Models
Abstract
Computational pathology models infer biologically and clinically meaningful outcomes from histology, but their predictions are shaped by complex, intertwined tissue signals whose roles are important to understand. Common pixel- and feature-space perturbations can produce implausible tissue, making model responses difficult to interpret. We introduce CPathOGen, a conditional latent-diffusion framework for generating paired H\&E counterfactuals with explicit controls over cellular spatial organization, nuclear morphology, and stain appearance. Cellular maps condition spatial structure through a spatial encoder, while a morphology/appearance vector modulates denoising through blockwise feature-wise linear modulation (FiLM). On held-out H\&E tiles, CPathOGen generates visually plausible tissue with improved distributional agreement after spatially guided selection, as reflected by lower Fréchet Inception Distance (FID) and Kernel Inception Distance (KID); generated cells track requested abundance and position, and measured morphology and color vary monotonically with their controls. We use these verified interventions to probe pathology encoders with endpoint heads, task-specific classifiers, and survival models. Responses are quantified using total variation distance and prediction-flip rate. We further introduce the Biology-Nuisance Sensitivity Ratio, a metric that contrasts model sensitivity to biologically motivated morphology and spatial factors with sensitivity to non-biological, stain-related nuisance variation. CPathOGen provides a practical, fidelity-audited framework for evaluating robustness and controlled feature sensitivity in computational pathology model. \href{https://github.com/a12dongithub/PathOGen}{GitHub} and \href{https://huggingface.co/a12donhf/CPathOGen}{Hugging~Face}.
Related
- Querying Counterfactuals on Tissue Graphs with Supervised Disentanglement
- What Language Adds, and at What Cost: Four Controlled Probes of Pathology Vision-Language Models
- Investigating the Role of Different Co-receptors on T-cell Activation Through IFN-γ Secretion Using Spatially Controlled Cell Monitoring Platform
- Generalized pollination of morphologically specialized Vincetoxicum hirundinaria (Apocynaceae, Asclepiadoideae)
- Anterior part of fronto-temporal language network as the neural basis of morphologically structured word form processing: evidence from Korean morphologically complex word.
BibTeX
@misc{singhal2026cpathogen,
title = {{CPathOGen: Spatially and Morphologically Controlled H&E Counterfactuals for Probing Pathology Models}},
author = {Samarth Singhal and Varang Rai},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.00602},
eprint = {2610.00602},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.00602},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - CPathOGen: Spatially and Morphologically Controlled H&E Counterfactuals for Probing Pathology Models
AU - Samarth Singhal
AU - Varang Rai
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2610.00602
UR - https://arxiv.org/abs/2610.00602
AB - Computational pathology models infer biologically and clinically meaningful outcomes from histology, but their predictions are shaped by complex, intertwined tissue signals whose roles are important to understand. Common pixel- and feature-space perturbations can produce implausible tissue, making model responses difficult to interpret. We introduce CPathOGen, a conditional latent-diffusion framework for generating paired H\&E counterfactuals with explicit controls over cellular spatial organization, nuclear morphology, and stain appearance. Cellular maps condition spatial structure through a spatial encoder, while a morphology/appearance vector modulates denoising through blockwise feature-wise linear modulation (FiLM). On held-out H\&E tiles, CPathOGen generates visually plausible tissue with improved distributional agreement after spatially guided selection, as reflected by lower Fréchet Inception Distance (FID) and Kernel Inception Distance (KID); generated cells track requested abundance and position, and measured morphology and color vary monotonically with their controls. We use these verified interventions to probe pathology encoders with endpoint heads, task-specific classifiers, and survival models. Responses are quantified using total variation distance and prediction-flip rate. We further introduce the Biology-Nuisance Sensitivity Ratio, a metric that contrasts model sensitivity to biologically motivated morphology and spatial factors with sensitivity to non-biological, stain-related nuisance variation. CPathOGen provides a practical, fidelity-audited framework for evaluating robustness and controlled feature sensitivity in computational pathology model. \href{https://github.com/a12dongithub/PathOGen}{GitHub} and \href{https://huggingface.co/a12donhf/CPathOGen}{Hugging~Face}.
ER -