iSight: Towards expert-AI co-assessment for improved immunohistochemistry staining interpretation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915772965584896 |
|---|---|
| author | Leiby, Jacob S. Yao, Jialu Lu, Pan Hu, George Davidian, Anna Koga, Shunsuke Leung, Olivia Patel, Pravin Resta, Isabella Tondi Rojansky, Rebecca Sung, Derek Yang, Eric Zhang, Paul J. Lundberg, Emma Kim, Dokyoon Yeung-Levy, Serena Zou, James Montine, Thomas Nirschl, Jeffrey Huang, Zhi |
| author_facet | Leiby, Jacob S. Yao, Jialu Lu, Pan Hu, George Davidian, Anna Koga, Shunsuke Leung, Olivia Patel, Pravin Resta, Isabella Tondi Rojansky, Rebecca Sung, Derek Yang, Eric Zhang, Paul J. Lundberg, Emma Kim, Dokyoon Yeung-Levy, Serena Zou, James Montine, Thomas Nirschl, Jeffrey Huang, Zhi |
| contents | Immunohistochemistry (IHC) provides information on protein expression in tissue sections and is commonly used to support pathology diagnosis and disease triage. While AI models for H\&E-stained slides show promise, their applicability to IHC is limited due to domain-specific variations. Here we introduce HPA10M, a dataset that contains 10,495,672 IHC images from the Human Protein Atlas with comprehensive metadata included, and encompasses 45 normal tissue types and 20 major cancer types. Based on HPA10M, we trained iSight, a multi-task learning framework for automated IHC staining assessment. iSight combines visual features from whole-slide images with tissue metadata through a token-level attention mechanism, simultaneously predicting staining intensity, location, quantity, tissue type, and malignancy status. On held-out data, iSight achieved 85.5\% accuracy for location, 76.6\% for intensity, and 75.7\% for quantity, outperforming fine-tuned foundation models (PLIP, CONCH) by 2.5--10.2\%. In addition, iSight demonstrates well-calibrated predictions with expected calibration errors of 0.0150-0.0408. Furthermore, in a user study with eight pathologists evaluating 200 images from two datasets, iSight outperformed initial pathologist assessments on the held-out HPA dataset (79\% vs 68\% for location, 70\% vs 57\% for intensity, 68\% vs 52\% for quantity). Inter-pathologist agreement also improved after AI assistance in both held-out HPA (Cohen's $κ$ increased from 0.63 to 0.70) and Stanford TMAD datasets (from 0.74 to 0.76), suggesting expert--AI co-assessment can improve IHC interpretation. This work establishes a foundation for AI systems that can improve IHC diagnostic accuracy and highlights the potential for integrating iSight into clinical workflows to enhance the consistency and reliability of IHC assessment. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_04063 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | iSight: Towards expert-AI co-assessment for improved immunohistochemistry staining interpretation Leiby, Jacob S. Yao, Jialu Lu, Pan Hu, George Davidian, Anna Koga, Shunsuke Leung, Olivia Patel, Pravin Resta, Isabella Tondi Rojansky, Rebecca Sung, Derek Yang, Eric Zhang, Paul J. Lundberg, Emma Kim, Dokyoon Yeung-Levy, Serena Zou, James Montine, Thomas Nirschl, Jeffrey Huang, Zhi Computer Vision and Pattern Recognition Immunohistochemistry (IHC) provides information on protein expression in tissue sections and is commonly used to support pathology diagnosis and disease triage. While AI models for H\&E-stained slides show promise, their applicability to IHC is limited due to domain-specific variations. Here we introduce HPA10M, a dataset that contains 10,495,672 IHC images from the Human Protein Atlas with comprehensive metadata included, and encompasses 45 normal tissue types and 20 major cancer types. Based on HPA10M, we trained iSight, a multi-task learning framework for automated IHC staining assessment. iSight combines visual features from whole-slide images with tissue metadata through a token-level attention mechanism, simultaneously predicting staining intensity, location, quantity, tissue type, and malignancy status. On held-out data, iSight achieved 85.5\% accuracy for location, 76.6\% for intensity, and 75.7\% for quantity, outperforming fine-tuned foundation models (PLIP, CONCH) by 2.5--10.2\%. In addition, iSight demonstrates well-calibrated predictions with expected calibration errors of 0.0150-0.0408. Furthermore, in a user study with eight pathologists evaluating 200 images from two datasets, iSight outperformed initial pathologist assessments on the held-out HPA dataset (79\% vs 68\% for location, 70\% vs 57\% for intensity, 68\% vs 52\% for quantity). Inter-pathologist agreement also improved after AI assistance in both held-out HPA (Cohen's $κ$ increased from 0.63 to 0.70) and Stanford TMAD datasets (from 0.74 to 0.76), suggesting expert--AI co-assessment can improve IHC interpretation. This work establishes a foundation for AI systems that can improve IHC diagnostic accuracy and highlights the potential for integrating iSight into clinical workflows to enhance the consistency and reliability of IHC assessment. |
| title | iSight: Towards expert-AI co-assessment for improved immunohistochemistry staining interpretation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.04063 |