Study objectives: Endobronchial ultrasonography (EBUS) is currently the sole clinical method available for delineating the bronchial wall structure; however, the image resolution is inadequate. Thus, an improved image analysis system is needed for both a more accurate and more readily interpretable endobronchial ultrasonogram.
Materials and methods: A total of 10 patients underwent pulmonary resection for lung cancer. EBUS was performed on the bronchi of the resected lungs, which had been immersed in physiologic saline solution. The same bronchial lesion in each specimen was imaged with two probes: 20 MHz and 30 MHz. The images were analyzed using the plot profile derived from freeware image analysis software: NIH Image (National Institutes of Health; Bethesda, MD). The measured echo intensity of the bronchial wall was statistically analyzed.
Results: A normal bronchial wall image consists of five layers, and the plot profile shows a W-shaped curve. The mean value of the echo intensity of each peak or trough of the W-shaped curve was calculated and compared for both probes. The differences in the mean echo intensity between both the third and fourth layer and the second and fourth layer were found to be significantly greater with the 30-MHz probe than with the 20-MHz probe. The echo intensity curve of a central-type lung cancer was not W shaped, indicating that the bronchial wall was not composed of the normal five layers.
Conclusion: We employed image analysis software and drew a plot to obtain a W-shaped curve from the EBUS image data. This enabled us to make an objective assessment of the laminar structure of the bronchial wall. In order to clearly recognize the laminar structure of the bronchial wall, the 30-MHz probe was found to be more useful than the 20-MHz probe.