> For the complete documentation index, see [llms.txt](https://www.parapathology.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.parapathology.com/clippings/artificial-intelligence-enables-quantitative-assessment-of-ulcerative-colitis-histology.md).

# Artificial Intelligence Enables Quantitative Assessment of Ulcerative Colitis Histology

Ulcerative colitis is a chronic inflammatory bowel disease that is characterized by a relapsing and remitting course. Assessment of disease activity critically informs treatment decisions. In addition

### Summary

Histologic remission is becoming a treatment target in ulcerative colitis alongside endoscopic remission. Pathologist scoring, however, is semi-quantitative and coarse.

This PathAI study develops convolutional networks that segment tissue regions and detect cells on H\&E whole-slide images of colonic biopsies. Their outputs become slide-level human-interpretable features, which are then used to reproduce the Nancy Histological Index (NHI). A random forest on 13 of these features matched pathologist-consensus NHI with weighted κ 0.91 and Spearman ρ 0.89. It predicted histologic remission, defined as absence of neutrophil extravasation, with 0.97 accuracy.

> The abstract in the `description:` frontmatter is the paper's own, verbatim. The full text is paywalled at the publisher and has no PubMed Central copy, and the medRxiv preprint could not be retrieved, so the sections below are my own-words digest of the published abstract only.

### Citation

Najdawi F, Sucipto K, Mistry P, Hennek S, Jayson CKB, Lin M, et al. Artificial Intelligence Enables Quantitative Assessment of Ulcerative Colitis Histology. *Mod Pathol*. 2023;36(6):100124. Epub 2023 Feb 15. doi: [10.1016/j.modpat.2023.100124](https://doi.org/10.1016/j.modpat.2023.100124). PMID: [36841434](https://pubmed.ncbi.nlm.nih.gov/36841434/).

* **Platform:** PathAI
* **Access:** paywalled; no PubMed Central copy
* **Preprint:** medRxiv, doi: [10.1101/2022.04.28.22274339](https://doi.org/10.1101/2022.04.28.22274339). Its abstract reports different figures (κ 0.93, ρ 0.93, remission accuracy 0.94, with remission defined as resolution of active inflammation), so cite the published version.
* **Overlap:** four authors (Najdawi, Fahy, Khosla, Jayson) also wrote the celiac disease feasibility study captured alongside this note, which uses the same approach of building interpretable features from cell and tissue models.

### Approach

1. **Tissue and cell models.** Convolutional networks segment tissue regions and detect cells directly on H\&E whole-slide images.
2. **Human-interpretable features.** Model outputs are summarised as slide-level quantitative features that describe each sample.
3. **Correlation.** Features are compared with disease severity and with pathologist-assigned NHI.
4. **Prediction.** A random forest on 13 features predicts NHI; histologic remission is predicted from the absence of neutrophil extravasation.

### Key findings

* **Model predictions.** Tissue and cell predictions agreed with pathologist annotations; the abstract gives no figures.
* **Feature correlations.** Slide-level features correlated strongly with disease severity and with pathologist NHI.
* **NHI prediction.** Weighted κ 0.91 and Spearman ρ 0.89 (P < .001) against pathologist consensus.
* **Histologic remission.** Accuracy 0.97.

### Dataset size

The abstract does not give one. A web search summary attributes to this paper 490 training and 147 validation slides plus 293 held-out slides from the PathAI Diagnostics laboratory, with about 38,000 tissue and 124,000 cell annotations. I could not check this against the full text `[unverified]`.

### On the lecture slide

From a lecture slide titled *"Top quantitative models in GI/liver path"*:

| Disease            | Model            | Output                   | Training set       | Reference                       |
| ------------------ | ---------------- | ------------------------ | ------------------ | ------------------------------- |
| Ulcerative colitis | Najdawi (PathAI) | Neutrophils, Nancy index | 512 WSIs (334 pts) | Najdawi et al. Mod Pathol. 2023 |

"512 WSIs (334 pts)" **could not be verified**: the abstract gives no dataset size, the full text is paywalled, and the preprint could not be retrieved. It also does not match the unverified split above `[unverified]`.