CARGO: A Cytometry Analysis framework via Regularized Graph Optimal-transport

Our plain-English summary of CARGO: A Cytometry Analysis framework via Regularized Graph Optimal-transport, by A.S. Shemonti, G.B. Gmyrek, K.L.A. Quintelier, et al (PLoS Computational Biology, 2026).

The study

• Authors: A.S. Shemonti, G.B. Gmyrek, K.L.A. Quintelier, S. Van Gassen, Y. Saeys, M. Willemsen, J.G.J.V. Aerts, E.V.E. Madsen, J.P. Robinson, A. Pothen, B. Rajwa
• Journal: PLoS Computational Biology
• Published: 2026;22(6):e1014358
• Cited by: 1 article, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Conventional data visualization techniques in single-cell analysis (such as two-dimensional dot plots, SPADE, PCA, t-SNE, or UMAP) often fall short in enabling an intuitive understanding of high-parameter flow cytometry data. These methods tend to oversimplify complex biological relationships, lack biologically meaningful interpretations, and offer no principled framework for downstream quantitative analysis. To address these limitations, we present a graph-based (network-based) visualization framework grounded in optimal transport theory. In this framework, cell populations are defined by their marker-expression profiles, and inter-population similarity is quantified using an efficiently computable optimal transport formulation known as the Sinkhorn distance. Our approach produces biologically consistent two-dimensional graph layouts using a phenotype-aware Hamming distance. Structural differences between sample graphs are characterized through a customized graph-edit distance that captures changes in population size, marker expression, and relationships between populations. We demonstrate our methods on two flow cytometry datasets: one from a clinical trial of dendritic cell-based immunotherapy in malignant peritoneal mesothelioma, involving 14 patients sampled at three time points with 14-color panels, and another from FlowCAP-II, which involved 43 acute myeloid leukemia patient samples analyzed with 7-color panels. Our framework produces robust, quantitative visual summaries of cell populations and supports statistical analysis based on graph edit distances, thereby offering new insights into disease progression and treatment response. Ultimately, our method bridges the gap between flow cytometry data visualization and biological interpretation.

Copyright: © 2026 Shemonti et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

In plain English

This article described CARGO, a computer method for displaying and comparing flow cytometry data on blood cell markers. It did not study asbestos. Each sample became a network of cell types, with lines showing how similar their marker patterns were. A 'graph edit distance' measured change between samples. The method was tested on a dendritic cell-based immunotherapy trial of 14 malignant peritoneal mesothelioma patients, sampled before vaccination and after the first and third vaccines; one did not complete the trial. An acute myeloid leukemia dataset was also used. For one patient, graphs showed trends the authors called consistent with the earlier trial, including fewer naïve T cells after the third vaccine. Of the 13 trial completers, seven patients' distance scores kept rising and six declined after the third dose. In a descriptive comparison, the rising group showed longer progression-free survival. A baseline-to-third-vaccine test gave p less than or equal to 0.001; citing few meaningful permutations, the authors said it suggested immune shifts, not definitive clinical significance. The authors concluded the method may be useful for immune monitoring, but called this exploratory.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Computational methods paper: development of a graph-based visualization and comparison framework for flow cytometry data, demonstrated on two existing datasets
• Who was studied: 14 patients with malignant peritoneal mesothelioma in a dendritic cell-based immunotherapy trial at Erasmus MC Cancer Institute, Rotterdam (Group 1, N = 9; Group 2, N = 5). Blood cells were sampled at baseline, two weeks after the first vaccine and two weeks after the third vaccine. One of the 14 patients did not complete the trial. The method was also shown on a FlowCAP-II acute myeloid leukemia dataset of 359 individuals (316 healthy controls and 43 AML patients).
• Exposure: No asbestos or fiber exposure was studied. The mesothelioma data came from a trial of adjuvant dendritic cell-based immunotherapy given after cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (CRS-HIPEC).
• Main results:
  • In one mesothelioma patient (MCV005, co-inhibition panel), natural killer cells, effector memory T cells and terminally differentiated effector memory T cells increased in proportion over treatment. Naïve T cells decreased after the third vaccine. The authors described these as trends consistent with the earlier trial report by Dietz et al.
  • In the same patient, the link between NK cells and NKT cells grew stronger over vaccination. This indicated more similar marker-expression profiles between the two cell types.
  • Based on graph edit distance trajectories, six mesothelioma patients' values declined after the third dose and seven continued to increase. One of the 14 patients did not complete the trial. The authors described this as a descriptive grouping.
  • In a descriptive comparison with trial outcomes, mesothelioma patients with increasing graph edit distance trajectories showed longer progression-free survival than those with declining trajectories.
  • A pseudo-F test on the change in graph edit distance from baseline to the third vaccination across mesothelioma patients gave p less than or equal to 0.001. The authors described this test as exploratory and hypothesis-generating, and said it should be read as suggestive of immunological shifts because of the limited number of meaningful permutations.
  • In the leukemia dataset, one AML patient (Individual 5) showed reduced lymphocyte proportions and an expanded myeloblast population compared with a healthy donor (Individual 1).
• The authors' conclusion: The authors concluded that the correspondence they observed between graph edit distance patterns and progression-free survival suggests the method may be useful for immune monitoring in longitudinal studies. They stated that this finding remains exploratory.
• Limitations the authors state:
  • Substantial baseline heterogeneity and the small cohort size limit formal statistical inference about immunological shifts caused by the immunotherapy.
  • Because few meaningful permutations were possible, the p value should be read as suggestive rather than as definitive evidence of clinical significance.
  • The patient grouping was not pre-specified in the trial and is not a clinical classification.
  • The comparison with survival outcomes was illustrative and does not constitute clinical inference or validation.
  • The framework works on predefined cell populations and is not designed to discover new cell types.

Topics

Medical terms in this study, explained in our Lexicon: Clinical Trial, Immunotherapy, Peritoneal Mesothelioma, Disease Progression.

Related studies

• Robust Cell Detection and Segmentation for Image Cytometry Reveal Th17 Cell Heterogeneity — T. Tsujikawa, G. Thibault, et al. (2019)
• Diagnostic DNA-flow- vs. -image-cytometry in effusion cytology — H. Motherby, N. Pomjanski, et al. (2002)
• Optimal isolation of extracellular vesicles from pleural fluid and profiling of their microRNA cargo — T.M. Chee, H.E. O'Farrell, et al. (2023)
• A predictive assay for human tumor cellular response to hyperthermia using dansyl lysine staining and flow cytometry — S.Y. Woo, G.C. Rice, et al. (1988)
• Flow cytometry in the exploration of the physiopathology of occupational lung disease — A.D. Curran (1999)
• Comparison of imaging based single-cell resolution spatial transcriptomics profiling platforms using formalin-fixed paraffin-embedded tumor samples — N. Ozirmak Lermi, M. Molina Ayala, et al. (2025)

Browse return to medical research.