Related Experiment Video
Updated: Oct 1, 2025
Quantification of Cerebral Vascular Architecture using Two-photon Microscopy in a Mouse Model of HIV-induced Neuroinflammation
Published on: January 12, 2016
Microvessel quantification by fully convolutional neural networks associated with type 2 inflammation in chronic
Wendong Liu1, Xing Liu2, Nan Zhang3
1Department of Otolaryngology, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong, PR China; Otorhinolaryngology Institute, Sun Yat-sen University, Guangzhou, Guangdong, PR China; The First Affiliated Hospital, Sun Yat-sen University, International Airway Research Center, Guangzhou, Guangdong, PR China.
Chronic rhinosinusitis (CRS) involves unclear pathogenesis, but this study used a fully convolutional network (FCN) to quantify microvessels. Angiogenesis is closely linked to CRS endotypes, suggesting anti-angiogenesis therapies for refractory cases.
Area of Science:
- Otorhinolaryngology
- Pathology
- Medical Imaging
Background:
- The pathogenesis of chronic rhinosinusitis (CRS) remains poorly understood, with limited knowledge regarding the role of angiogenesis.
- Current understanding of angiogenesis in CRS is insufficient for targeted therapeutic development.
Purpose of the Study:
- To quantify microvessels in CRS tissues using a fully convolutional network (FCN).
- To investigate the association between microvessel quantification and inflammation in CRS endotypes.
Main Methods:
- A fully convolutional network (FCN) was developed and validated using 552 images from 27 CRS tissue samples.
- Microvessel quantification was performed on 79 CRS patients and 17 controls using the optimized FCN.
- Correlation analyses were conducted between microvessel density and various clinical and endotyping parameters, including cytokine levels.
Main Results:
- Significant differences in microvessel quantification were observed between type 2 and non-type 2 CRS, with higher levels in type 2 CRS.
- A strong negative correlation was found between microvessel area ratio and tissue levels of tumor necrosis factor alpha and transforming growth factor beta.
- A mild positive correlation was noted between microvessel area ratio and tissue concentrations of IL-5 and eosinophilic cationic protein.
Conclusions:
- Fully convolutional networks (FCNs) effectively facilitate microvessel analysis in airway tissues.
- This study highlights a strong association between angiogenesis and CRS endotyping.
- Targeting angiogenesis may offer a therapeutic strategy for recurrent and refractory chronic rhinosinusitis.
