Automated 3D Quantitative Analysis of Intrapulmonary Vessel Volume on Noncontrast CT in Healthy Individuals

Ming Ying, Yu Zhang, Ran Xiao, Ruijie Zhao, Jiaru Wang, Sirong Piao, Lan Song, Y. T. Xu, Xin Sui, Wei SongCurrent 2025 · DOI

TL;DR

This paper introduces an automated method for quantifying intrapulmonary vessel volume (IPVV) from non-contrast CT scans, revealing that lung windows provide a more comprehensive measure than mediastinal windows. This technique offers a feasible and objective tool for assessing pulmonary vasculature, potentially aiding in early disease detection.

What Problem Does This Paper Address?

Understanding the intricate network of pulmonary blood vessels is crucial for diagnosing and managing various lung diseases. Traditionally, assessing these vessels relied on invasive methods or subjective interpretations of autopsy findings. While CT scans offer a non-invasive window into pulmonary anatomy, accurately and quantitatively analyzing the vast and complex network of intrapulmonary vessels, especially distinguishing between arteries and veins, has been challenging. Existing methods often struggle with manual segmentation, which is time-consuming and prone to variability. Furthermore, the choice of CT window settings (e.g., lung vs. mediastinal) can influence measurements, and there's a lack of consensus on which window is optimal for comprehensive IPVV analysis in healthy individuals. This study aims to bridge this gap by developing an automated, quantitative approach to IPVV analysis and comparing measurements obtained from different CT window settings.

Methods

The researchers developed and employed an automated algorithm for the quantitative analysis of intrapulmonary vessel volume (IPVV) from non-contrast chest CT scans. The core of their method involves a post-processing algorithm that first separates pulmonary arteries and veins. This is followed by a U-Net-based model for inference, utilizing a sliding window approach to handle varying volume sizes. The algorithm preprocesses CT images by cropping, normalizing, and resampling them to a specific resolution. Post-processing steps include outlier removal and adjustments for vessel branching.

A key aspect of their methodology is the comparison of IPVV measurements derived from two distinct CT window settings: the "lung window" and the "mediastinal window." The algorithm was applied to extract IPVV data from both window types for each participant. To ensure accuracy, two experienced radiologists visually reviewed the segmented vessels, manually correcting any mislabeling of arteries and veins. The study also calculated Body Surface Area (BSA) for each participant to adjust vascular parameters, acknowledging that vessel size can be influenced by body size. Statistical analysis was performed using standard software to compare IPVV values between the two window settings and to investigate sex-based differences.

Datasets

The study utilized a dataset of 258 healthy individuals who underwent routine physical examinations and non-contrast chest CT scans between January and November 2023. These participants were retrospectively enrolled. The inclusion criteria ensured image quality suitable for segmentation, negative CT scans for pulmonary disease, and no history of dust exposure. Exclusion criteria addressed factors that could confound the analysis, such as thoracic surgery, poor image quality, or significant pulmonary lesions. The paper does not mention the introduction of a new benchmark dataset; rather, it uses existing CT scans from a clinical setting.

Key Insights and Ideas

  • Lung Window Superiority: The most significant finding is that the lung window consistently yields higher IPVV measurements compared to the mediastinal window. This suggests that the lung window is more effective at capturing the full extent of intrapulmonary vasculature, particularly smaller vessels.
  • Automated Analysis Feasibility: The study demonstrates that automated 3D quantitative analysis of IPVV is not only feasible but also effective in identifying morphological characteristics of intrapulmonary vessels. This opens doors for more objective and reproducible assessments.
  • Sex-Based Differences Emerge: While overall IPVV is higher in males, the study identified nuanced sex-based differences in specific vessel diameter ranges. Females showed higher IPVV in pulmonary vessels between 1.6 and 2.4 mm, while males generally had larger IPVV in other categories, potentially linked to differences in body size and cardiac output.
  • Age-Related Decline: IPVV in both males and females shows a progressive decrease with advancing age, highlighting the impact of aging on the pulmonary vascular system.
  • Potential for Early Detection: The ability to quantitatively assess IPVV in healthy individuals provides a baseline. Deviations from these normal values could serve as early indicators of pulmonary vascular abnormalities, aiding in the timely detection of diseases.

Key Results

The study found that IPVV for the entire lung, intrapulmonary arteries, intrapulmonary veins, and small pulmonary vessels were significantly higher when extracted from the lung window compared to the mediastinal window (p<0.01). For instance, Total Intrapulmonary Vessel Volume (TIPVV) was 120.42 mL/m² with the lung window versus 102.33 mL/m² with the mediastinal window. No significant sex-based differences were observed for pulmonary vessels with diameters between 0.8 and 1.6 mm, and for pulmonary veins between 2.4 and 3.2 mm. However, females had significantly higher IPVV than males for pulmonary arteries and veins with diameters between 1.6 and 2.4 mm. In all other vessel diameter categories, males exhibited higher IPVV than females. The algorithm achieved an extraction success rate of approximately 100%.

Limitations and Future Work

The authors acknowledge several limitations:

  • Single-Center Retrospective Study: The findings are based on data from a single center and a retrospective design, which may limit generalizability to broader populations.
  • Potential Selection Bias: The selection of participants might introduce bias.
  • CT Pixel Size Limitation: The resolution of current CT scanners (~0.8 mm pixel size) limits the ability to capture the smallest pulmonary microvasculature, potentially affecting the accuracy of measurements for very small vessels.

Despite these limitations, the study suggests promising future directions:

  • Broader Population Studies: Expanding the study to include a larger, more diverse population is crucial for validating the findings and establishing more robust normative data.
  • Clinical Application: The developed automated IPVV analysis method holds significant potential for clinical application, particularly in the early detection and monitoring of pulmonary vascular diseases like Idiopathic Pulmonary Fibrosis (IPF) and Chronic Obstructive Pulmonary Disease (COPD).
  • AI Integration: Further integration of Artificial Intelligence (AI) could enhance the accuracy and efficiency of these analyses.
  • Future CT Technology: With advancements in CT technology (e.g., 1024 matrix machines), pixel value limitations will decrease, potentially allowing for even more precise measurements of smaller vessels.

Summary updated

Abstract

This method of automatic IPVV extraction and quantitative assessment has been proven to be feasible. Automated IPVV expression effectively identified morphological characteristics of intrapulmonary vessels. The study has concluded IPVVs extracted from the lung window to be generally larger than those extracted from the mediastinal window.
BibTeX
@inproceedings{ying2025automated,
  author = {Ming Ying and Yu Zhang and Ran Xiao and Ruijie Zhao and Jiaru Wang and Sirong Piao and Lan Song and Y. T. Xu and Xin Sui and Wei Song},
  title = {{Automated 3D Quantitative Analysis of Intrapulmonary Vessel Volume on Noncontrast CT in Healthy Individuals}},
  booktitle = {Current},
  year = {2025},
  doi = {10.2174/0115734056354924241115102310},
}
Figure