EUVIMEDEUVIMED entdecken
Open Access

A Data‐Efficient Computational Framework for Reconstruction and Simulation of Microcirculatory Perfusion

Peilun Li iD, Xiang Geng, Yanfei Shen, Molei Yan, Qing Pan iD, Gangmin Ning iD

Microcirculation · 2026-10 · 33 (7)

DOI: 10.1111/micc.70084 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Microcirculation, online erschienen am 31.08.2026. Verfasst von Peilun Li und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Biomedical Engineering, Zhejiang University, Hangzhou, China. Hauptthemen laut MEDLINE: Microcirculation, Models, Cardiovascular, Computer Simulation. Die Arbeit nennt 66 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: National Key Research and Development Program of China, National Natural Science Foundation of China, Natural Science Foundation of Zhejiang Province.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose

Microcirculation is essential for maintaining tissue viability and organ function by delivering oxygen and nutrients. Impaired microvascular perfusion may result in tissue dysfunction despite apparently normal macrocirculatory parameters. However, comprehensive experimental characterization of entire microvascular networks remains technically challenging, limiting quantitative assessment of network‐level perfusion. This study aimed to develop a data‐efficient computational framework for reconstructing microcirculatory network structure and estimating blood perfusion from sparse local vascular measurements.

Methods

Rat mesenteric microvascular networks, which share selected quasi‐two‐dimensional structural characteristics with human sublingual microcirculation, were used for methodological validation. Morphological and flow information was extracted from individual vessel segments, and three datasets, each containing 10 randomly selected venous bifurcations, were used to emulate limited data availability. A conditional generative adversarial network (cGAN) was pretrained on a larger source corpus and fine‐tuned using each sparse target dataset. Venous trees were subsequently generated and integrated with derived arterial trees and capillary connections to reconstruct complete microvascular networks. Initial flow rates were assigned from local measurements, followed by network optimization using a structural adaptation model.

Results

Five microcirculatory networks were generated for each sparse bifurcation dataset, enabling quantification of network‐level perfusion descriptors, including blood‐flow fractal dimension, multifractal spectrum, and capillary‐flow heterogeneity. Under controlled distribution shift at the bifurcation‐generation level, fine‐tuning successfully shifted the distributions of generated bifurcation exponent and asymmetry ratio toward the corresponding sparse target data. The reconstructed networks exhibited broadly similar diameter‐conditioned velocity and pressure trends to those in the reference network. Structural adaptation further differentiated the diameter–flow relationships in arterial and venous, with the fitted slopes approaching the corresponding reference values.

Conclusion

The proposed framework provides a data‐efficient framework for physiologically plausible reconstruction of microcirculatory networks and estimation of network‐level perfusion from sparse local bifurcation data. Validation in rat mesenteric microvascular networks demonstrates the methodological feasibility of the framework and highlights its potential as a quantitative tool for perfusion analysis and advancing translational studies of microcirculatory modeling.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

MicrocirculationModels, CardiovascularComputer SimulationAnimalsRatsGenerative Adversarial NetworksHumansMale

Stichwörter der Autor:innen

data‐efficient reconstructionmicrocirculationmicrovascular modelingperfusion characterizationstructural adaptationtransfer learning

Autor:innen und Institutionen

  • Department of Biomedical Engineering, Zhejiang University, Hangzhou, China.
  • Department of Orthopedics, Huashan Hospital Affiliated to Fudan University, Shanghai, China.
  • Department of Intensive Care Unit, Zhejiang Hospital, Hangzhou, China.
  • College of Information Engineering, Zhejiang University of Technology, Hangzhou, China.

ORCID-Profile: Peilun Li, Qing Pan, Gangmin Ning

Förderung

  • National Key Research and Development Program of China
  • National Natural Science Foundation of China
  • Natural Science Foundation of Zhejiang Province

Zitieren

APA

Li, P., Geng, X., Shen, Y., Yan, M., Pan, Q., & Ning, G. (2026). A Data‐Efficient Computational Framework for Reconstruction and Simulation of Microcirculatory Perfusion. Microcirculation, 33(7). https://doi.org/10.1111/micc.70084

Vancouver

Li P, Geng X, Shen Y, Yan M, Pan Q, Ning G. A Data‐Efficient Computational Framework for Reconstruction and Simulation of Microcirculatory Perfusion. Microcirculation. 2026;33(7). doi:10.1111/micc.70084

BibTeX
@article{Li2026,
  author = {Li, Peilun and Geng, Xiang and Shen, Yanfei and Yan, Molei and Pan, Qing and Ning, Gangmin},
  title = {A Data‐Efficient Computational Framework for Reconstruction and Simulation of Microcirculatory Perfusion},
  journal = {Microcirculation},
  year = {2026},
  volume = {33},
  number = {7},
  doi = {10.1111/micc.70084}
}
Kostenlos starten →