Different approaches to Imaging Mass Cytometry data analysis.
Part of the results reproduced; minor but material deviations remained.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
- ✓Any deviation was negligible
- 🔴Could not use the authors’ exact input data
- 🔴Reported values were only indirectly comparable
- 🟡A deviation arose in the data or preprocessing
- 🟡A deviation was attributed to the published material
- 🟡Reported values were not (fully) derivable from the shared data
- 🟡The central claim did not (fully) hold under reproduction
- 🟡Overall, the reproduction showed a material discrepancy
▸Reproduction agent’s raw note
DROP (non_pipeline), independently re-verified 2026-07-15 via NCBI efetch XML. PMID 37092034 'Different approaches to Imaging Mass Cytometry data analysis' (Milosevic V., single author, Univ. Bergen; Bioinformatics Advances 3(1):vbad046, 2023) is a REVIEW article: PubMed PublicationTypeList = 'Journal Article' + 'Review' (UI D016454); abstract states its aim is 'to give a systematic synopsis of all the available classical image analysis tools and pipelines useful ... for IMC data analysis.' It has NO Methods section, NO Data/Code Availability statement, and NO <DataBank> element in the PubMed record (DataBankList count = 0) -> no accession belongs to this paper. It defines no author-run pipeline and reports no author-computed quantitative result. The few numbers in prose (IMC-Denoise F1, Cellpose 'outperformed all', CellSighter '80-100% accuracy') are other groups' published results re-quoted from the cited tools' own papers, not reproducible FROM this paper. The two auto-extracted links are text-mining FALSE POSITIVES confirmed against the reference list: code github.com/angelolab/MAUI is reference #9 (Baranski A. et al. 2021, MAUI, PLoS Comput Biol, PMC8084329) and data zenodo:10.5281/zenodo.3841961 is MAUI's own example dataset -- neither belongs to this review (same trap as prior review-drop RU pmid-36920617). The datasets[] entry is the misattributed Zenodo link recorded for traceability only (grade D = not this paper's data). No «our HPC» compute spent: control-plane decision independent of VPN/cluster state. NOT attempted: any pipeline run, since the paper defines no reproducible pipeline-derived result of its own.
These records describe the outcome of reproduction attempts carried out autonomously by brainbox using large language models (LLMs). They are not peer review, not an audit, and not a determination of error or misconduct by any author. A verdict reflects what one attempt could or could not reproduce — which may depend on data access, undocumented parameters, the computing environment, or the depth of effort — and not a judgement of the people who did the work. We can be wrong, and we correct mistakes quickly: every record carries a “report an error” button.
Assessment versions
Every reproduction run is kept as an immutable version — anchored to the data as it stood, with a tamper-evident chain hash. A rerun (e.g. after an author updates a deposit) adds a new version; the previous one stays on record.
-
v1 current initial assessmentassessed: 2026-06-19 ⛓ 3b8c00cfff04
✎ I am an author of this paper
Updated or fixed a deposit, or is there an erratum? Ask us to re-run the metrics. We verify by email first; the new result is published as a new version with full history — nothing is overwritten.
Provenance — full disclosure
When this reproduction was carried out, which methodology version was used, and by whom — so the record can be audited and checked independently.
- Reproduced
- 2026-07-15
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-19no human curator yet
- Last updated
- 2026-08-05
Provisional, curator- or AI-assessed, and independently checkable. A reproduction outcome states what one attempt could reproduce — not a judgement of the authors.
Deep full-text extraction
Model: sonnet- ★ Imaging Mass Cytometry (IMC) is a high-multiplexing imaging platform capable of simultaneously detecting and visualizing up to 40 different protein targets in tissue sections. finding
- ★ Due to the characteristics of IMC-derived data, the image analysis approach can diverge from classical image analysis pipelines used for standard micrographic images. finding
- ★ This review provides a systematic synopsis of classical and IMC-specific image analysis tools and pipelines to help researchers select the most suitable methodology for a given analysis. resource
- ★ IMC combines suspension-based mass cytometry (CyTOF) technology with UV laser ablation (Hyperion Tissue Imager) performed on stained tissue sections to add spatial resolution. method
- Antibodies in CyTOF/IMC are coupled to stable metal isotopes via metal-chelating polymers instead of fluorophores, with detection/quantification via time-of-flight mass spectrometry. mechanism
- ★ IMC is largely unaffected by autofluorescence and signal spillover common in immunofluorescence, but can show weak signal, low signal-to-noise ratio, channel crosstalk, hot pixels and speckle artifacts. finding
- ★ ilastik-based pixel classification for background noise removal also normalizes signal across samples, thereby removing batch effects. finding
- CATALYST generates a spillover matrix from adjacent-channel signals to correct channel crosstalk between metal isotope channels. method
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| Imaging Mass Cytometry (protein marker imaging) | FFPE and frozen tissue sections | none | spatial protein marker expression (up to 35-40 markers) | Hyperion Tissue Imager / Helios mass cytometer (Standard BioTools) |
| Raw data visualization/conversion | IMC raw MCD data files | none | converted multichannel/single-channel .tiff images | MCD Viewer; napari with napari-imc plugin; readimc python package |
| Channel crosstalk / spillover correction | IMC images, agarose glass slide antibody conjugates | none | corrected signal via spillover compensation matrix | CATALYST R/Bioconductor package |
| Image preprocessing (speckle/aggregate removal, crosstalk correction, background reduction) | IMC and MIBI images | none | denoised/binarized pixel intensity images | MAUI (MATLAB-based) |
| Background noise pixel classification | IMC images | none | binary expression maps (signal vs background pixels) | ilastik random forest pixel classifier |
- – IMC enables simultaneous detection of 35-40 different markers with minimal crosstalk between channels.
- – The Helios mass cytometer can discriminate between isotope masses differing by only 1 Da. 1 Da
- – ilastik pixel binarization removes background noise and also eliminates batch effect issues across samples.
- – readimc/MCD viewer are the only tools discussed that are used exclusively for IMC data; all other tools accept .tiff input and are also usable with other multiplexed imaging technologies (e.g. CODEX).
- count up to 40 different protein targets (maximum multiplexing capacity of IMC)
- count 35 to 40 different markers (practical marker detection range with minimal crosstalk)
- other 1 µm² ablation spot size (UV laser ablation resolution per cycle)
- other λ = 193 nm (wavelength of the stationary UV ablation laser)
- other mass resolution of 1 Da (precision of Helios mass cytometer in discriminating isotope masses)
- count 29 tools/pipelines listed (Table 1 overview of IMC data analysis tools by processing stage)
Statistical methods review
Model: sonnetA neutral, descriptive read of the statistical approach — what was done, and (for shared learning, not as criticism) what could also have been done.
This is a narrative review article surveying software tools and pipelines for analyzing Imaging Mass Cytometry (IMC) data (visualization, preprocessing, segmentation, cell phenotyping, and spatial/pixel analysis). It does not present original experimental data, group comparisons, or inferential statistical analysis; it instead describes and tabulates the capabilities, platforms, and advantages/disadvantages of existing computational tools.
-
The review summarizes and compares IMC analysis tools primarily through narrative description and qualitative feature tables (e.g. presence/absence of capabilities, advantages/disadvantages).↳ Could also: A structured benchmarking study with quantitative performance metrics (e.g. segmentation accuracy, F1 score, runtime, reproducibility across datasets) could also be used — Quantitative benchmarking would allow objective, statistically comparable evaluation of tool performance across shared datasets, complementing the qualitative synthesis presented here.
-
Background noise correction (e.g. via ilastik pixel classification) is described as converting continuous signal intensities into binary signal/background maps.↳ Could also: Continuous intensity normalization methods (e.g. quantile normalization, arcsinh transformation as used in CyTOF/IMC workflows) could also be used instead of binarization — Retaining continuous intensity values (rather than binarizing) can preserve graded expression information, which may be useful for downstream analyses that benefit from continuous rather than binary marker readouts.
-
Channel crosstalk/spillover correction is described using the CATALYST compensation matrix approach based on single-antibody control slides.↳ Could also: Computational post-acquisition compensation methods not requiring dedicated control slides (e.g. the approach described by Wang et al. 2019, mentioned in the text) could also be used — Post-acquisition computational compensation can reduce the additional wet-lab preparation burden associated with generating dedicated single-stain control slides.
-
The review presents tool comparisons in a single summary table without a formal weighting or scoring framework for decision-making.↳ Could also: A formal multi-criteria decision analysis (MCDA) or scoring rubric weighting factors like accuracy, scalability, and usability could also be used — A structured scoring framework could help readers more systematically translate the qualitative comparisons into a tool-selection decision tailored to their specific data and resources.
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — PMID 37092034
- Title: Different approaches to Imaging Mass Cytometry data analysis.
- Author: Milosevic V. (single author), Univ. Bergen / CCBIO, Norway.
- Journal: Bioinformatics Advances 3(1):vbad046, 2023. DOI 10.1093/bioadv/vbad046.
- PMID 37092034 · PMCID PMC10115470.
Article type — confirmed REVIEW (no original computation)
PubMed PublicationTypeList = Journal Article + Review (verified via NCBI
efetch XML, 2026-06-19). Abstract verbatim:
"This review has for an aim to give a systematic synopsis of all the available classical image analysis tools and pipelines useful to be employed for IMC data analysis and give an overview of tools intentionally developed solely for this purpose..."
PMC full text confirms it is a narrative/systematic synopsis organised in four phases (raw-data visualisation/conversion, preprocessing, cell segmentation, downstream phenotyping/spatial analysis). It surveys ~40 third-party tools (readimc, CATALYST, MAUI, IMC-Denoise, Steinbock, nf-core/imcyto, MCMICRO, Cellpose, DeepCell/Mesmer, histoCAT, Astir, CELESTA, CellSighter, ...).
In-scope pipeline-derived results: NONE
- No Methods section. The author runs no pipeline of his own.
- No Data Availability / Code Availability statement. NCBI
DataBankListcount = 0 → no deposited accession belongs to this paper. - No original dataset. No organism/tissue/ROI/marker counts are reported as the author's own data; all numbers in prose are quoted from the cited tools' own papers (e.g. "IMC-Denoise ... evaluating ... F1 score", "Cellpose ... outperform all other compared algorithms", "CellSighter ... accuracy between 80% and 100%"). These are other groups' published results re-stated in the review, not values the review author computed — none are reproducible from this paper, and reproducing the cited tools' own papers is a different RU.
The two auto-extracted links are text-mining FALSE POSITIVES
code_url = github.com/angelolab/MAUI— MAUI is reference #9 (Baranski et al., PLoS Comput Biol 2021, PMC8084329), one of dozens of tools merely cited. It is not the review's code (the review has no code).data = zenodo:10.5281/zenodo.3841961— this is MAUI's own example dataset (Angelo lab), not data used or deposited by this review.
Associating either with PMID 37092034 would be incorrect (same trap documented for the prior review-drop RU pmid-36920617).
Decision
DROP — drop_reason = non_pipeline (text-mining false positive: a Review
article with no Methods, no original data, no author-run pipeline, and no
pinnable reported value derivable from this paper). No «our HPC» compute spent;
this is a control-plane decision, independent of VPN state.
No individual results have been recorded for this entry yet.
Assessments & scoring basis
Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.
An automated assessment. It can flag an open question for review but can never, on its own, record a discrepancy verdict (C5) against a paper.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
This is a Review article (Milosevic V., Bioinform Adv 2023) with no Methods, no original dataset, no Data/Code Availability statement, and no author-computed result — correctly dropped as non_pipeline. The few numbers in prose are other groups' published metrics re-quoted from cited tool papers, not derivable from this paper, so there is nothing to put against a reproduced value. The extracted code (angelolab/MAUI) and data (zenodo:3841961) links are text-mining false positives belonging to the cited MAUI tool, not this review. No deviation or fabrication is implicated — the case is simply uncheckable, hence yellow overall rather than red.
Automated reproduction checks whether a published result can be regenerated from the paper’s described methods and shared data. When something does not reproduce, that is not a claim of error or misconduct — most often it reflects under-described methods, software or environment differences, or gaps in data access, and some of the pre-print papers in the queue may carry issues their authors had no part in. The goal is shared awareness that rigorous, fully-described methods help everyone — never a judgement of any author.
Are you an author? We would genuinely like to hear from you — to clarify the record, add data or code, re-run the pipeline after an accession update, and publish your response right next to the assessment. Everything here is open and auditable.
🚩 Report an error in this record
Spotted something wrong — a verdict you’d contest, a data or value error, or a private detail that slipped through? Tell us, with a short justification. Authors and readers are equally welcome to write in; we review every report.
Prefer email, or the form below not working? Contact us at support@doesitreproduce.com.
Reproduction footprint
claude-opus-4-8Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.