[Threshold Segmentation of Pulmonary Subsolid Nodules on CT Images:
Detection and Quantification of the Solid...

Area of Science:

• Radiology
• Medical Imaging
• Pulmonary Medicine

Background:

• Accurate detection and quantification of solid components in pulmonary subsolid nodules (SSNs) are crucial for diagnosis, prognosis, and treatment planning.
• Currently, objective criteria for solid component analysis in SSNs are lacking.
• Computed tomography (CT) imaging is widely used for SSN assessment.

Purpose of the Study:

• To determine the optimal threshold values for detecting and quantifying solid components in SSNs using CT images.
• To evaluate the efficacy of threshold segmentation in analyzing SSNs.

Main Methods:

• Retrospective analysis of CT images from 102 SSNs.
• Manual measurements of solid component volume by observers to establish a reference standard.
• Application of threshold segmentation with varying settings to quantify solid volumes and calculate solid-to-total volume ratios.
• Comparison of segmentation results with reference standards using ROC curves and Wilcoxon tests.

Main Results:

• Thresholds of -250 HU and -300 HU demonstrated high diagnostic value for solid component detection (AUC 0.982 and 0.977, respectively).
• Cut-off values for solid-to-total volume ratio were 1.10% and 6.14% for the respective thresholds.
• Median solid component volumes measured by segmentation were not significantly different from the reference standard.

Conclusions:

• Threshold segmentation is a valuable method for detecting and quantifying solid components in SSNs on chest CT images.
• Recommended thresholds of -250 HU and -300 HU provide reliable results for SSN analysis.
• This technique aids in objective assessment and management of SSNs.