UnMICST: Deep learning with real augmentation for robust segmentation of highly multiplexed images of human tissues

Yapp, Clarence; Novikov, Edward; Jang, Won-Dong; Yuan, Chen; Cicconet, Marcelo; Maliga, Zoltan; Jacobson, Connor A.; Wei, Donglai; Santagata, Sandro; Pfister, Hanspeter; Sorger, Peter K.

doi:10.1101/2021.04.02.438285

Cited by 19 publications

(27 citation statements)

References 92 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The complete preanalytical CyCIF image processing (stitching, registration, illumination correction, segmentation, and single-cell feature extraction) was performed using the MCMICRO pipeline (36) an open-source multiple-choice microscopy pipeline, versions 60929d5b82 and 7547d0c42a (full codes available on GitHub https://github.com/labsyspharm/mcmicro). For the generation of probability maps and the nuclei segmentation, a trained U-Net model UnMicst v1 was used followed by a marker- controlled watershed used for single-cell segmentation (111). A diameter range of 3 to 60 pixels was used for nuclei detection.…”

Section: Methodsmentioning

confidence: 99%

The spatial landscape of progression and immunoediting in primary melanoma at single cell resolution

Nirmal

Maliga

Vallius

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Cutaneous melanoma is a highly immunogenic disease, surgically curable at early stages, but life-threatening when metastatic. Here we integrate high-plex imaging, 3D high-resolution microscopy, and spatially-resolved micro-region transcriptomics to study immune evasion and immunoediting in primary melanoma. We find that recurrent cellular neighborhoods involving tumor, immune, and stromal cells change significantly along a progression axis from precursor states to melanoma in situ to invasive tumor. Hallmarks of immunosuppression are detectable by the precursor stage, and when tumors become locally invasive, a consolidated and spatially restricted suppressive environment forms along the tumor-stromal boundary. This environment is established by cytokine gradients that promote expression of MHC-II and IDO1 and by PDL1-expressing macrophages and dendritic cells engaging activated T cells. However, a few mm away, T cells synapse with melanoma cells in fields of tumor regression. Thus, invasion and immunoediting can co-exist within a few millimeters of each other in a single specimen.

show abstract

Section: Methodsmentioning

confidence: 99%

The spatial landscape of progression and immunoediting in primary melanoma at single cell resolution

Nirmal

Maliga

Vallius

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…5). We segmented individual cells using UNMICST 60 and ImageJ, and quantified fluorescence intensities on a per-cell basis generating marker expression data from 940,891 cells. Gaussian Mixture Modeling (GMM) was used to assign cell types, and dimensionality reduction of the single cell data was visualized using t-distributed stochastic neighbor embedding (t-SNE) (Fig.…”

Section: Multiplexed Single Cell Analysis Of Glioblastoma Tissuementioning

confidence: 99%

Single Cell Spatial Analysis Reveals the Topology of Immunomodulatory Purinergic Signaling in Glioblastoma

Coy

Wang

Stopka

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Glioblastoma develops an immunosuppressive microenvironment that fosters tumorigenesis and resistance to current therapeutic strategies. Here we use multiplexed tissue imaging and single-cell RNA-sequencing to characterize the composition, spatial organization, and clinical significance of extracellular purinergic signaling in glioblastoma. We show that glioblastoma exhibit strong expression of CD39 and CD73 ectoenzymes, correlating with increased adenosine levels. Microglia are the predominant source of CD39, while CD73 is principally expressed by tumor cells, particularly in tumors with amplification of EGFR and astrocyte-like differentiation. Spatially-resolved single-cell analyses demonstrate strong spatial correlation between tumor CD73 and microglial CD39, and that their spatial proximity is associated with poor clinical outcomes. Together, this data reveals that tumor CD73 expression correlates with tumor genotype, lineage differentiation, and functional states, and that core purine regulatory enzymes expressed by neoplastic and tumor-associated myeloid cells interact to promote a distinctive adenosine-rich signaling niche and immunosuppressive microenvironment potentially amenable to therapeutic targeting.

show abstract

“…We hypothesize that training tissue-specific models will play a bigger role in accurate segmentation than the underlying machine learning methodology. Emerging studies also show that data augmentation and the inclusion of a nuclear envelope stain, such as Lamin, can substantially improve segmentation accuracy in a method-agnostic way (Yapp et al, 2021).…”

Section: Image Processing Workflowmentioning

confidence: 99%

A community-based approach to image analysis of cells, tissues and tumors

Csbc

Vizcarra

Burlingame

et al. 2021

Preprint

View full text Add to dashboard Cite

Emerging multiplexed imaging platforms provide an unprecedented view of an increasing number of molecular markers at subcellular resolution and the dynamic evolution of tumor cellular composition. As such, they are capable of elucidating cell-to-cell interactions within the tumor microenvironment that impact clinical outcome and therapeutic response. However, the rapid development of these platforms has far outpaced the computational methods for processing and analyzing the data they generate. While being technologically disparate, all imaging assays share many computational requirements for post-collection data processing. We convened a workshop to characterize these shared computational challenges and a follow-up hackathon to implement solutions for a selected subset of them. Here, we delineate these areas that reflect major axes of research within the field, including image registration, segmentation of cells and subcellular structures, and identification of cell types from their morphology. We further describe the logistical organization of these events, believing our lessons learned can aid others in uniting the imaging community around self-identified topics of mutual interest, in designing and implementing operational procedures to address those topics and in mitigating issues inherent in image analysis (e.g., sharing exemplar images of large datasets and disseminating baseline solutions to hackathon challenges through open-source code repositories).

show abstract

UnMICST: Deep learning with real augmentation for robust segmentation of highly multiplexed images of human tissues

Cited by 19 publications

References 92 publications

The spatial landscape of progression and immunoediting in primary melanoma at single cell resolution

The spatial landscape of progression and immunoediting in primary melanoma at single cell resolution

Single Cell Spatial Analysis Reveals the Topology of Immunomodulatory Purinergic Signaling in Glioblastoma

A community-based approach to image analysis of cells, tissues and tumors

Contact Info

Product

Resources

About