MVis-Fold: A Three-Dimensional Microvascular Structure Inference Model for Super-Resolution Ultrasound

2; (2) Research Center of Interventional Medicine; 3; (3) Wenling Institute of Big Data; 4); 4) ((1) Department of Diagnostic Ultrasound Imaging & Interventional Therapy; (4) Zhejiang Provincial Research Center for Innovative Technology; (5) College of Information Science; (6) Department of Ultrasound; Artificial Intelligence in Medicine

arxiv: 2604.04477 · v1 · submitted 2026-04-06 · 💻 cs.CV

MVis-Fold: A Three-Dimensional Microvascular Structure Inference Model for Super-Resolution Ultrasound

Jincao Yao (1 , 2 , 3 , 4) , Ke Zhang (1) , Yahan Zhou (1) , Jiafei Shen (1) , Jie Liu (1)

show 27 more authors

Mudassar Ali (5) Bojian Feng (1) Jiye Chen (1) Jinlong Fan (2) Ping Liang (6) Dong Xu (1 4) ((1) Department of Diagnostic Ultrasound Imaging & Interventional Therapy Zhejiang Cancer Hospital Hangzhou Institute of Medicine Chinese Academy of Sciences Hangzhou China (2) Research Center of Interventional Medicine Engineering (3) Wenling Institute of Big Data Artificial Intelligence in Medicine Taizhou (4) Zhejiang Provincial Research Center for Innovative Technology Equipment in Interventional Oncology (5) College of Information Science Electronic Engineering Zhejiang University (6) Department of Ultrasound Chinese PLA General Hospital Chinese PLA Medical School Beijing China)

This is my paper

Pith reviewed 2026-05-10 20:18 UTC · model grok-4.3

classification 💻 cs.CV

keywords microvascular reconstructionsuper-resolution ultrasoundthree-dimensional inferencecross-scale networktumor imagingvascular analysisquantitative ultrasound

0 comments

The pith

A model called MVis-Fold reconstructs three-dimensional microvascular networks from two-dimensional super-resolution ultrasound images.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops MVis-Fold to infer complete three-dimensional microvascular structures from two-dimensional super-resolution ultrasound scans. Conventional two-dimensional SRUS cannot readily supply measurements such as vessel lengths or branching in full three-dimensional space. The authors integrate a cross-scale network architecture to perform this inference and validate the results on solid tumors. If the approach holds, it would support quantitative three-dimensional analysis of microvasculature for disease diagnosis and monitoring.

Core claim

MVis-Fold integrates a cross-scale network architecture to perform high-fidelity inference and reconstruction of three-dimensional microvascular networks from two-dimensional SRUS images. It precisely calculates key parameters in three-dimensional space that traditional two-dimensional SRUS cannot readily obtain. The model was validated for accuracy and reliability in three-dimensional microvascular reconstruction of solid tumors.

What carries the argument

The cross-scale network architecture inside MVis-Fold, which takes two-dimensional SRUS images as input and generates three-dimensional microvascular reconstructions.

If this is right

Enables calculation of three-dimensional microvascular parameters including vessel densities, lengths, and branching patterns.
Provides a foundation for quantitative three-dimensional analysis of microvasculature in solid tumors.
Supplies new tools and methods for diagnosis and monitoring of diseases that involve changes in vascular structure.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same inference approach could extend to other two-dimensional imaging modalities to recover hidden three-dimensional structural details.
Routine clinical ultrasound workflows might incorporate this reconstruction to produce real-time three-dimensional vascular maps.
Longitudinal studies of tumor blood supply could become feasible by tracking three-dimensional network changes over time.

Load-bearing premise

Two-dimensional SRUS images contain sufficient unambiguous information to allow accurate inference of the complete three-dimensional microvascular topology without major artifacts or loss of connectivity.

What would settle it

Acquire ground-truth three-dimensional microvascular images of the same tumor samples using an independent high-resolution method such as micro-CT or optical microscopy and directly compare vessel connectivity and parameter values against the model's outputs.

read the original abstract

Super-resolution ultrasound (SRUS) technology has overcome the resolution limitations of conventional ultrasound, enabling micrometer-scale imaging of microvasculature. However, due to the nature of imaging principles, three-dimensional reconstruction of microvasculature from SRUS remains an open challenge. We developed microvascular visualization fold (MVis-Fold), an innovative three-dimensional microvascular reconstruction model that integrates a cross-scale network architecture. This model can perform high-fidelity inference and reconstruction of three-dimensional microvascular networks from two-dimensional SRUS images. It precisely calculates key parameters in three-dimensional space that traditional two-dimensional SRUS cannot readily obtain. We validated the model's accuracy and reliability in three-dimensional microvascular reconstruction of solid tumors. This study establishes a foundation for three-dimensional quantitative analysis of microvasculature. It provides new tools and methods for diagnosis and monitoring of various diseases.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

MVis-Fold applies a cross-scale network to infer 3D microvascular maps from 2D SRUS but the validation does not test whether the inputs actually contain enough information to resolve unique topologies.

read the letter

The main thing to know is that this paper presents MVis-Fold, a cross-scale deep learning model that takes 2D super-resolution ultrasound images and produces 3D reconstructions of microvascular networks in solid tumors, along with 3D parameters that standard 2D SRUS cannot give directly. It frames the 3D reconstruction task as an open problem and offers one concrete architecture to address it. The tumor validation shows the method applied to relevant clinical cases rather than just phantoms. That is the useful part: a practical extension of multi-scale networks to this specific imaging modality and use case. The cross-scale design fits the range of vessel sizes involved, and the focus on oncology metrics is a reasonable clinical angle. The paper does not overclaim beyond what the abstract and description support, which keeps it grounded. The soft spots are in the evidence for the central claim. SRUS images are line-integral projections, so multiple 3D vessel arrangements can produce the same 2D pattern, especially with depth overlaps or crossings. The description supplies no quantitative reconstruction metrics, no baseline comparisons, no details on 3D ground-truth acquisition, and no topology checks such as connectivity error or branch preservation. There is also no mention of tests under varied angles or partial views. This leaves the high-fidelity claim resting on unexamined priors, exactly as the stress-test note flags. The concern holds up because the paper does not appear to include the ablations or falsifiable checks that would separate learned consistency from actual recovery. The work is aimed at ultrasound researchers and oncologists who need better 3D vascular quantification tools. A reader already working on SRUS or medical image reconstruction would get value from the model description and the tumor results, even while wanting stronger numbers. It deserves peer review because it supplies a complete method with a clear application and the field can use the experiments once the validation gaps are addressed. I would send it to referees with a request for metrics, ground-truth methods, and robustness tests rather than desk-reject it.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes MVis-Fold, a cross-scale network model for inferring and reconstructing three-dimensional microvascular networks from two-dimensional super-resolution ultrasound (SRUS) images. It claims high-fidelity 3D reconstruction, precise calculation of 3D parameters unavailable in 2D SRUS, and validation on solid tumor cases to enable quantitative 3D microvascular analysis for disease diagnosis and monitoring.

Significance. If the central claims are supported by rigorous evidence, the work would address a recognized open challenge in SRUS by enabling 3D microvascular topology and parameter extraction from 2D projections. This could provide practical tools for tumor characterization and other microvascular diseases where depth-resolved connectivity and branching are clinically relevant.

major comments (3)

[Abstract] Abstract: The claims of 'high-fidelity inference and reconstruction' and 'validated the model's accuracy and reliability' are unsupported by any quantitative metrics (e.g., vessel Dice score, Hausdorff distance, connectivity error, or topology metrics such as Betti numbers), dataset statistics, ground-truth acquisition method, or baseline comparisons. This renders the central empirical claim unevaluable from the manuscript.
[Validation] Validation description: The text states validation only on 'solid tumors' without specifying 3D ground-truth acquisition (serial histology, optical clearing, multi-angle SRUS, etc.), training data generation process, or robustness tests against projection ambiguities (depth overlaps, vessel crossings). These omissions are load-bearing because 2D SRUS is a line-integral modality whose inverse problem is ill-posed without strong priors.
[Methods] Methods/Architecture: No equations, network diagrams, loss functions, or details on the cross-scale integration are supplied. Without these, it is impossible to assess whether the model learns genuine 3D disambiguation or simply reproduces training priors, directly affecting the claim of reliable generalization beyond the reported tumor cases.

minor comments (2)

[Abstract] The acronym 'MVis-Fold' and the term 'fold' are introduced without explanation of their meaning or motivation relative to the network design.
[Abstract] The abstract refers to 'key parameters in three-dimensional space' without naming them or indicating how they are computed from the inferred graph.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the thorough and constructive review. The comments highlight important areas where additional detail will strengthen the manuscript. We address each major comment below and will revise accordingly.

read point-by-point responses

Referee: [Abstract] Abstract: The claims of 'high-fidelity inference and reconstruction' and 'validated the model's accuracy and reliability' are unsupported by any quantitative metrics (e.g., vessel Dice score, Hausdorff distance, connectivity error, or topology metrics such as Betti numbers), dataset statistics, ground-truth acquisition method, or baseline comparisons. This renders the central empirical claim unevaluable from the manuscript.

Authors: We agree that the abstract would benefit from explicit quantitative support. In the revision we will add key metrics (vessel Dice, Hausdorff distance, connectivity error) and dataset statistics while retaining conciseness, with full tables and baseline comparisons retained in the Results section. This directly addresses the evaluability concern. revision: yes
Referee: [Validation] Validation description: The text states validation only on 'solid tumors' without specifying 3D ground-truth acquisition (serial histology, optical clearing, multi-angle SRUS, etc.), training data generation process, or robustness tests against projection ambiguities (depth overlaps, vessel crossings). These omissions are load-bearing because 2D SRUS is a line-integral modality whose inverse problem is ill-posed without strong priors.

Authors: We acknowledge the need for explicit methodological transparency. The revised manuscript will include a dedicated subsection describing multi-angle SRUS acquisitions used to generate pseudo-ground-truth 3D volumes, the synthetic training-data pipeline based on realistic 3D microvascular phantoms, and quantitative robustness tests against depth overlaps and crossings. These additions will clarify how the model priors mitigate the ill-posed inverse problem. revision: yes
Referee: [Methods] Methods/Architecture: No equations, network diagrams, loss functions, or details on the cross-scale integration are supplied. Without these, it is impossible to assess whether the model learns genuine 3D disambiguation or simply reproduces training priors, directly affecting the claim of reliable generalization beyond the reported tumor cases.

Authors: We will expand the Methods section to include the cross-scale integration equations, a network architecture diagram, the full loss function (with fidelity, topology, and smoothness terms), and an explanation of how the architecture enforces 3D consistency. Ablation studies demonstrating disambiguation beyond training priors will also be added to support generalization claims. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical ML model performance with no mathematical derivation chain

full rationale

The paper presents an ML-based inference model (cross-scale network) for 3D microvascular reconstruction from 2D SRUS images, validated empirically on solid-tumor cases. No equations, derivations, fitted parameters, or self-citations are shown that reduce any claimed result to its inputs by construction. The central claim rests on reported model accuracy rather than any self-definitional loop, fitted-input prediction, or imported uniqueness theorem. This is the expected non-circular outcome for a data-driven architecture paper.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The central claim depends on the unstated assumption that 2D SRUS projections encode recoverable 3D topology and on the empirical success of the cross-scale network; no free parameters, axioms, or invented entities are explicitly introduced in the abstract.

pith-pipeline@v0.9.0 · 5620 in / 1048 out tokens · 23982 ms · 2026-05-10T20:18:25.588686+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages

[1]

Annotation-free 3D reconstruction and quantification of retinal microvasculature by RADAR

Zhang H, et al. Annotation-free 3D reconstruction and quantification of retinal microvasculature by RADAR. npj Digital Medicine 9, 181 (2026)

work page 2026
[2]

Geometric-topological deep transfer learning for precise vessel segmentation in 3D medical volumes

Wu J, Wen Z, Zhou H, et al. Geometric-topological deep transfer learning for precise vessel segmentation in 3D medical volumes. npj Digital Medicine 9, 58 (2026)

work page 2026
[3]

Fully automated 3D aortic segmentation of 4D flow MRI for hemodynamic analysis using deep learning

Berhane, Haben et al. Fully automated 3D aortic segmentation of 4D flow MRI for hemodynamic analysis using deep learning. Magnetic resonance in medicine vol. 84,4 (2020)

work page 2020
[4]

Photo-activated ultrasound localization imaging with laser-activated nanodroplets

Zhao S, Yi J, Qiu Y, et al. Photo-activated ultrasound localization imaging with laser-activated nanodroplets. Communications Engineering 5, 43 (2026)

work page 2026
[5]

Hybrid photoacoustic and fast super-resolution ultrasound imaging.” Nature communications vol

Zhao, Shensheng et al. Hybrid photoacoustic and fast super-resolution ultrasound imaging.” Nature communications vol. 14, (2023)

work page 2023
[6]

Generative AI enables medical image segmentation in ultra low-data regimes

Zhang L, Jindal B, Alaa A, et al. Generative AI enables medical image segmentation in ultra low-data regimes. Nature Communications 16, 6486 (2025). 10

work page 2025
[7]

Dynamic microvascular monitoring with miniaturized omnidirectional broadband photoacoustic imaging system for living entities (MOBILE)

Li, Wei et al. Dynamic microvascular monitoring with miniaturized omnidirectional broadband photoacoustic imaging system for living entities (MOBILE). Nature communications vol. 17, (2026)

work page 2026
[8]

From pretraining to privacy: federated ultrasound foundation model with self-supervised learning

Jiang Y, Feng CM, Ren J, et al. From pretraining to privacy: federated ultrasound foundation model with self-supervised learning. npj Digital Medicine 8, 714 (2025)

work page 2025
[9]

Pathway-Aware Multimodal Transformer (PAMT): Integrating Pathological Image and Gene Expression for Interpretable Cancer Survival Analysis

Yan, Rui et al. “Pathway-Aware Multimodal Transformer (PAMT): Integrating Pathological Image and Gene Expression for Interpretable Cancer Survival Analysis.” IEEE transactions on pattern analysis and machine intelligence vol. 48, (2025)

work page 2025
[10]

A fully automated deep learning pipeline for micro-CT-imaging-based densitometry of lung fibrosis murine models

Vincenzi, Elena et al. “A fully automated deep learning pipeline for micro-CT-imaging-based densitometry of lung fibrosis murine models.” Respiratory research vol. 23, (2022)

work page 2022
[11]

Machine and deep learning methods for radiomics

Avanzo, Michele et al. “Machine and deep learning methods for radiomics.” Medical physics vol. 47,5 (2020)

work page 2020
[12]

Super-Resolution Ultrasound Localization Microscopy for Visualization of the Ocular Blood Flow

Qian, Xuejun et al. “Super-Resolution Ultrasound Localization Microscopy for Visualization of the Ocular Blood Flow.” IEEE transactions on bio-medical engineering vol. 69,5 (2022)

work page 2022
[13]

Deep Learning Applications in Imaging of Acute Ischemic Stroke: A Systematic Review and Narrative Summary

Jiang, Bin et al. “Deep Learning Applications in Imaging of Acute Ischemic Stroke: A Systematic Review and Narrative Summary.” Radiology vol. 315,1 (2025)

work page 2025
[14]

Assessing Completeness of Clinical Histories Accompanying Imaging Orders Using Adapted Open-Source and Closed-Source Large Language Models

Larson, David B et al. “Assessing Completeness of Clinical Histories Accompanying Imaging Orders Using Adapted Open-Source and Closed-Source Large Language Models.” Radiology vol. 314,2 (2025)

work page 2025
[15]

AI-based 3D analysis of retinal vascular structures using OCT angiography

Brown R, et al. AI-based 3D analysis of retinal vascular structures using OCT angiography. Nature Biomedical Engineering (2025)

work page 2025
[16]

3D reconstruction of a gastrulating human embryo

Xiao, Zhenyu et al. “3D reconstruction of a gastrulating human embryo.” Cell vol. 187,11 (2024)

work page 2024
[17]

Assessing the retinal microvasculature in heart failure: A systematic review and possible clinical implications

Vanreusel, Inne et al. “Assessing the retinal microvasculature in heart failure: A systematic review and possible clinical implications.” Microvascular research, 104949. 29 (2026)

work page 2026
[18]

Development and deployment of a functional 3D-bioprinted blood vessel

Dell, Annika C et al. “Development and deployment of a functional 3D-bioprinted blood vessel.” Scientific reports vol. 15, (2025)

work page 2025
[19]

Functional ultrasound localization microscopy reveals brain-wide neurovascular activity on a microscopic scale

Renaudin, Noémi et al. “Functional ultrasound localization microscopy reveals brain-wide neurovascular activity on a microscopic scale.” Nature methods vol. 19,8 (2022)

work page 2022
[20]

Photoacoustic tomography detects response and resistance to Bevacizumab in breast cancer mouse models

Quiros-Gonzalez I, et al. Photoacoustic tomography detects response and resistance to Bevacizumab in breast cancer mouse models. Cancer Res. (2022)

work page 2022
[21]

Myocardial Perfusion PET for the Detection and Reporting of Coronary Microvascular Dysfunction: A JACC: Cardiovascular Imaging Expert Panel Statement

Schindler, Thomas H et al. “Myocardial Perfusion PET for the Detection and Reporting of Coronary Microvascular Dysfunction: A JACC: Cardiovascular Imaging Expert Panel Statement.” JACC. Cardiovascular imaging vol. 16,4 (2023). 11

work page 2023
[22]

Microscopy based methods for characterization, drug delivery, and understanding the dynamics of nanoparticles

Gupta, Priyamvada et al. “Microscopy based methods for characterization, drug delivery, and understanding the dynamics of nanoparticles.” Medicinal research reviews vol. 44,1 (2024)

work page 2024
[23]

Deep learning for cellular image analysis

Moen E, Bannon D, Kudo T, et al. Deep learning for cellular image analysis. Nature Methods 16, 1233–1246 (2019)

work page 2019
[24]

Recent Advances in Deep Learning and Medical Imaging for Head and Neck Cancer Treatment: MRI, CT, and PET Scans

Illimoottil, Mathew, and Daniel Ginat. “Recent Advances in Deep Learning and Medical Imaging for Head and Neck Cancer Treatment: MRI, CT, and PET Scans.” Cancers vol. 15, (2023)

work page 2023
[25]

CLOOME: contrastive learning unlocks bioimaging databases for queries with chemical structures

Sanchez-Fernandez, Ana et al. “CLOOME: contrastive learning unlocks bioimaging databases for queries with chemical structures.” Nature communications vol. 14, (2023)

work page 2023
[26]

Incorporating the image formation process into deep learning improves network performance: Richardson-Lucy network for three-dimensional fluorescence microscopy deconvolution

Li Y, et al. Incorporating the image formation process into deep learning improves network performance: Richardson-Lucy network for three-dimensional fluorescence microscopy deconvolution. Nature Methods 19, 1427-1437 (2022)

work page 2022
[27]

A practical guide to bioimaging research data management in core facilities

Schmidt, Christian et al. “A practical guide to bioimaging research data management in core facilities.” Journal of microscopy vol. 294,3 (2024)

work page 2024
[28]

SegR-Net: A deep learning framework with multi-scale feature fusion for robust retinal vessel segmentation

Ryu, Jihyoung et al. “SegR-Net: A deep learning framework with multi-scale feature fusion for robust retinal vessel segmentation.” Computers in biology and medicine vol. 163 (2023)

work page 2023
[29]

Deep learning for fast super-resolution ultrasound microvessel imaging

Luan, Shunyao et al. “Deep learning for fast super-resolution ultrasound microvessel imaging.” Physics in medicine and biology vol. 68, (2023)

work page 2023

[1] [1]

Annotation-free 3D reconstruction and quantification of retinal microvasculature by RADAR

Zhang H, et al. Annotation-free 3D reconstruction and quantification of retinal microvasculature by RADAR. npj Digital Medicine 9, 181 (2026)

work page 2026

[2] [2]

Geometric-topological deep transfer learning for precise vessel segmentation in 3D medical volumes

Wu J, Wen Z, Zhou H, et al. Geometric-topological deep transfer learning for precise vessel segmentation in 3D medical volumes. npj Digital Medicine 9, 58 (2026)

work page 2026

[3] [3]

Fully automated 3D aortic segmentation of 4D flow MRI for hemodynamic analysis using deep learning

Berhane, Haben et al. Fully automated 3D aortic segmentation of 4D flow MRI for hemodynamic analysis using deep learning. Magnetic resonance in medicine vol. 84,4 (2020)

work page 2020

[4] [4]

Photo-activated ultrasound localization imaging with laser-activated nanodroplets

Zhao S, Yi J, Qiu Y, et al. Photo-activated ultrasound localization imaging with laser-activated nanodroplets. Communications Engineering 5, 43 (2026)

work page 2026

[5] [5]

Hybrid photoacoustic and fast super-resolution ultrasound imaging.” Nature communications vol

Zhao, Shensheng et al. Hybrid photoacoustic and fast super-resolution ultrasound imaging.” Nature communications vol. 14, (2023)

work page 2023

[6] [6]

Generative AI enables medical image segmentation in ultra low-data regimes

Zhang L, Jindal B, Alaa A, et al. Generative AI enables medical image segmentation in ultra low-data regimes. Nature Communications 16, 6486 (2025). 10

work page 2025

[7] [7]

Dynamic microvascular monitoring with miniaturized omnidirectional broadband photoacoustic imaging system for living entities (MOBILE)

Li, Wei et al. Dynamic microvascular monitoring with miniaturized omnidirectional broadband photoacoustic imaging system for living entities (MOBILE). Nature communications vol. 17, (2026)

work page 2026

[8] [8]

From pretraining to privacy: federated ultrasound foundation model with self-supervised learning

Jiang Y, Feng CM, Ren J, et al. From pretraining to privacy: federated ultrasound foundation model with self-supervised learning. npj Digital Medicine 8, 714 (2025)

work page 2025

[9] [9]

Pathway-Aware Multimodal Transformer (PAMT): Integrating Pathological Image and Gene Expression for Interpretable Cancer Survival Analysis

Yan, Rui et al. “Pathway-Aware Multimodal Transformer (PAMT): Integrating Pathological Image and Gene Expression for Interpretable Cancer Survival Analysis.” IEEE transactions on pattern analysis and machine intelligence vol. 48, (2025)

work page 2025

[10] [10]

A fully automated deep learning pipeline for micro-CT-imaging-based densitometry of lung fibrosis murine models

Vincenzi, Elena et al. “A fully automated deep learning pipeline for micro-CT-imaging-based densitometry of lung fibrosis murine models.” Respiratory research vol. 23, (2022)

work page 2022

[11] [11]

Machine and deep learning methods for radiomics

Avanzo, Michele et al. “Machine and deep learning methods for radiomics.” Medical physics vol. 47,5 (2020)

work page 2020

[12] [12]

Super-Resolution Ultrasound Localization Microscopy for Visualization of the Ocular Blood Flow

Qian, Xuejun et al. “Super-Resolution Ultrasound Localization Microscopy for Visualization of the Ocular Blood Flow.” IEEE transactions on bio-medical engineering vol. 69,5 (2022)

work page 2022

[13] [13]

Deep Learning Applications in Imaging of Acute Ischemic Stroke: A Systematic Review and Narrative Summary

Jiang, Bin et al. “Deep Learning Applications in Imaging of Acute Ischemic Stroke: A Systematic Review and Narrative Summary.” Radiology vol. 315,1 (2025)

work page 2025

[14] [14]

Assessing Completeness of Clinical Histories Accompanying Imaging Orders Using Adapted Open-Source and Closed-Source Large Language Models

Larson, David B et al. “Assessing Completeness of Clinical Histories Accompanying Imaging Orders Using Adapted Open-Source and Closed-Source Large Language Models.” Radiology vol. 314,2 (2025)

work page 2025

[15] [15]

AI-based 3D analysis of retinal vascular structures using OCT angiography

Brown R, et al. AI-based 3D analysis of retinal vascular structures using OCT angiography. Nature Biomedical Engineering (2025)

work page 2025

[16] [16]

3D reconstruction of a gastrulating human embryo

Xiao, Zhenyu et al. “3D reconstruction of a gastrulating human embryo.” Cell vol. 187,11 (2024)

work page 2024

[17] [17]

Assessing the retinal microvasculature in heart failure: A systematic review and possible clinical implications

Vanreusel, Inne et al. “Assessing the retinal microvasculature in heart failure: A systematic review and possible clinical implications.” Microvascular research, 104949. 29 (2026)

work page 2026

[18] [18]

Development and deployment of a functional 3D-bioprinted blood vessel

Dell, Annika C et al. “Development and deployment of a functional 3D-bioprinted blood vessel.” Scientific reports vol. 15, (2025)

work page 2025

[19] [19]

Functional ultrasound localization microscopy reveals brain-wide neurovascular activity on a microscopic scale

Renaudin, Noémi et al. “Functional ultrasound localization microscopy reveals brain-wide neurovascular activity on a microscopic scale.” Nature methods vol. 19,8 (2022)

work page 2022

[20] [20]

Photoacoustic tomography detects response and resistance to Bevacizumab in breast cancer mouse models

Quiros-Gonzalez I, et al. Photoacoustic tomography detects response and resistance to Bevacizumab in breast cancer mouse models. Cancer Res. (2022)

work page 2022

[21] [21]

Myocardial Perfusion PET for the Detection and Reporting of Coronary Microvascular Dysfunction: A JACC: Cardiovascular Imaging Expert Panel Statement

Schindler, Thomas H et al. “Myocardial Perfusion PET for the Detection and Reporting of Coronary Microvascular Dysfunction: A JACC: Cardiovascular Imaging Expert Panel Statement.” JACC. Cardiovascular imaging vol. 16,4 (2023). 11

work page 2023

[22] [22]

Microscopy based methods for characterization, drug delivery, and understanding the dynamics of nanoparticles

Gupta, Priyamvada et al. “Microscopy based methods for characterization, drug delivery, and understanding the dynamics of nanoparticles.” Medicinal research reviews vol. 44,1 (2024)

work page 2024

[23] [23]

Deep learning for cellular image analysis

Moen E, Bannon D, Kudo T, et al. Deep learning for cellular image analysis. Nature Methods 16, 1233–1246 (2019)

work page 2019

[24] [24]

Recent Advances in Deep Learning and Medical Imaging for Head and Neck Cancer Treatment: MRI, CT, and PET Scans

Illimoottil, Mathew, and Daniel Ginat. “Recent Advances in Deep Learning and Medical Imaging for Head and Neck Cancer Treatment: MRI, CT, and PET Scans.” Cancers vol. 15, (2023)

work page 2023

[25] [25]

CLOOME: contrastive learning unlocks bioimaging databases for queries with chemical structures

Sanchez-Fernandez, Ana et al. “CLOOME: contrastive learning unlocks bioimaging databases for queries with chemical structures.” Nature communications vol. 14, (2023)

work page 2023

[26] [26]

Incorporating the image formation process into deep learning improves network performance: Richardson-Lucy network for three-dimensional fluorescence microscopy deconvolution

Li Y, et al. Incorporating the image formation process into deep learning improves network performance: Richardson-Lucy network for three-dimensional fluorescence microscopy deconvolution. Nature Methods 19, 1427-1437 (2022)

work page 2022

[27] [27]

A practical guide to bioimaging research data management in core facilities

Schmidt, Christian et al. “A practical guide to bioimaging research data management in core facilities.” Journal of microscopy vol. 294,3 (2024)

work page 2024

[28] [28]

SegR-Net: A deep learning framework with multi-scale feature fusion for robust retinal vessel segmentation

Ryu, Jihyoung et al. “SegR-Net: A deep learning framework with multi-scale feature fusion for robust retinal vessel segmentation.” Computers in biology and medicine vol. 163 (2023)

work page 2023

[29] [29]

Deep learning for fast super-resolution ultrasound microvessel imaging

Luan, Shunyao et al. “Deep learning for fast super-resolution ultrasound microvessel imaging.” Physics in medicine and biology vol. 68, (2023)

work page 2023