Cello: A Universal Cell-wise Feature Aggregation framework for Reliable Pathology Images Analysis

Hengrui Lou, Weihan Li, Jiazhen Yang, Lingxiang Jia, Shengxuming Zhang, Linyun Zhou, Xiuming Zhang, Zhenyang Wang, Mingli Song, Zunlei Feng
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:80650-80666, 2026.

Abstract

Computational pathology has made progress in diagnosis and prognosis prediction from whole slide images (WSIs), yet pipelines still rely on patch-level feature extraction and aggregation, departing from the cell-centric reasoning used by pathologists. This gap limits sensitivity to micro-lesions and subtle changes, and current methods rarely provide a unified solution that supports both local and global tasks with trustworthy evidence. We propose Cello, a universal cell-wise feature aggregation framework for reliable pathology image analysis. Cello integrates cell-level representations into WSI modeling via protein-signal–supervised cell-wise learning, preserving fine-grained cellular cues under gigapixel constraints. For local tasks, Cello introduces a flexible prototype-based contrastive module for scalable, task-adaptive representation learning. For global tasks, Cello adopts a weakly supervised gated aggregation that can widely leverage WSI labels. Finally, a cell–local–global decision-route consistency objective dynamically aggregates cellular evidence and aligns local predictions with global outcomes, improving reliability and faithfulness. Trained with only hundreds to thousands of samples, Cello achieves performance gains of 3.0% 7.6% and outperforms SOTA pathology foundation models pretrained on tens of thousands of samples. Code is available at https://github.com/HengruiLou/Cello.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-lou26d, title = {Cello: A Universal Cell-wise Feature Aggregation framework for Reliable Pathology Images Analysis}, author = {Lou, Hengrui and Li, Weihan and Yang, Jiazhen and Jia, Lingxiang and Zhang, Shengxuming and Zhou, Linyun and Zhang, Xiuming and Wang, Zhenyang and Song, Mingli and Feng, Zunlei}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {80650--80666}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/lou26d/lou26d.pdf}, url = {https://proceedings.mlr.press/v306/lou26d.html}, abstract = {Computational pathology has made progress in diagnosis and prognosis prediction from whole slide images (WSIs), yet pipelines still rely on patch-level feature extraction and aggregation, departing from the cell-centric reasoning used by pathologists. This gap limits sensitivity to micro-lesions and subtle changes, and current methods rarely provide a unified solution that supports both local and global tasks with trustworthy evidence. We propose Cello, a universal cell-wise feature aggregation framework for reliable pathology image analysis. Cello integrates cell-level representations into WSI modeling via protein-signal–supervised cell-wise learning, preserving fine-grained cellular cues under gigapixel constraints. For local tasks, Cello introduces a flexible prototype-based contrastive module for scalable, task-adaptive representation learning. For global tasks, Cello adopts a weakly supervised gated aggregation that can widely leverage WSI labels. Finally, a cell–local–global decision-route consistency objective dynamically aggregates cellular evidence and aligns local predictions with global outcomes, improving reliability and faithfulness. Trained with only hundreds to thousands of samples, Cello achieves performance gains of 3.0% 7.6% and outperforms SOTA pathology foundation models pretrained on tens of thousands of samples. Code is available at https://github.com/HengruiLou/Cello.} }
Endnote
%0 Conference Paper %T Cello: A Universal Cell-wise Feature Aggregation framework for Reliable Pathology Images Analysis %A Hengrui Lou %A Weihan Li %A Jiazhen Yang %A Lingxiang Jia %A Shengxuming Zhang %A Linyun Zhou %A Xiuming Zhang %A Zhenyang Wang %A Mingli Song %A Zunlei Feng %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-lou26d %I PMLR %P 80650--80666 %U https://proceedings.mlr.press/v306/lou26d.html %V 306 %X Computational pathology has made progress in diagnosis and prognosis prediction from whole slide images (WSIs), yet pipelines still rely on patch-level feature extraction and aggregation, departing from the cell-centric reasoning used by pathologists. This gap limits sensitivity to micro-lesions and subtle changes, and current methods rarely provide a unified solution that supports both local and global tasks with trustworthy evidence. We propose Cello, a universal cell-wise feature aggregation framework for reliable pathology image analysis. Cello integrates cell-level representations into WSI modeling via protein-signal–supervised cell-wise learning, preserving fine-grained cellular cues under gigapixel constraints. For local tasks, Cello introduces a flexible prototype-based contrastive module for scalable, task-adaptive representation learning. For global tasks, Cello adopts a weakly supervised gated aggregation that can widely leverage WSI labels. Finally, a cell–local–global decision-route consistency objective dynamically aggregates cellular evidence and aligns local predictions with global outcomes, improving reliability and faithfulness. Trained with only hundreds to thousands of samples, Cello achieves performance gains of 3.0% 7.6% and outperforms SOTA pathology foundation models pretrained on tens of thousands of samples. Code is available at https://github.com/HengruiLou/Cello.
APA
Lou, H., Li, W., Yang, J., Jia, L., Zhang, S., Zhou, L., Zhang, X., Wang, Z., Song, M. & Feng, Z.. (2026). Cello: A Universal Cell-wise Feature Aggregation framework for Reliable Pathology Images Analysis. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:80650-80666 Available from https://proceedings.mlr.press/v306/lou26d.html.

Related Material