S3-CIMA: Supervised spatial single-cell image analysis for identifying disease-associated cell-type compositions in tissue

Sepideh Babaei; Jonathan Christ; Vivek Sehra; Ahmad Makky; Mohammed Zidane; Kilian Wistuba-Hamprecht; Christian Schürch; Manfred Claassen

doi:10.1016/j.patter.2023.100829

S³-CIMA: Supervised spatial single-cell image analysis for identifying disease-associated cell-type compositions in tissue

Patterns (N Y). 2023 Aug 17;4(9):100829. doi: 10.1016/j.patter.2023.100829. eCollection 2023 Sep 8.

Authors

Sepideh Babaei^{1

2}, Jonathan Christ³, Vivek Sehra^{1

4

2}, Ahmad Makky⁵, Mohammed Zidane⁵, Kilian Wistuba-Hamprecht^{1

6

2}, Christian Schürch⁵, Manfred Claassen^{1

4

2

7}

Affiliations

¹ Department of Internal Medicine I, University Hospital Tübingen, Tübingen, Germany.
² M3 Research Center, University Hospital Tübingen, Tübingen, Germany.
³ Department of Physics, University of Vienna, Vienna, Austria.
⁴ Department of Computer Science, University of Tübingen, Tübingen, Germany.
⁵ Department of Pathology and Neuropathology, University Hospital and Comprehensive Cancer Center Tübingen, Tübingen, Germany.
⁶ Department of Immunology, Institute of Cell Biology, University Hospital Tübingen, Tübingen, Germany.
⁷ Institute for Bioinformatics and Medical Informatics, University of Tübingen, Tübingen, Germany.

Abstract

The spatial organization of various cell types within the tissue microenvironment is a key element for the formation of physiological and pathological processes, including cancer and autoimmune diseases. Here, we present S³-CIMA, a weakly supervised convolutional neural network model that enables the detection of disease-specific microenvironment compositions from high-dimensional proteomic imaging data. We demonstrate the utility of this approach by determining cancer outcome- and cellular-signaling-specific spatial cell-state compositions in highly multiplexed fluorescence microscopy data of the tumor microenvironment in colorectal cancer. Moreover, we use S³-CIMA to identify disease-onset-specific changes of the pancreatic tissue microenvironment in type 1 diabetes using imaging mass-cytometry data. We evaluated S³-CIMA as a powerful tool to discover novel disease-associated spatial cellular interactions from currently available and future spatial biology datasets.

Keywords: disease-associated cell types; multiplexed imaging; spatial single cell data; supervised spatial enrichment analysis; tissue microenvironment; weakly supervised learning.