[Method for Spectral Analysis of the Respiratory Cycle Based on Dynamic Pulmonary Computed Tomography]

Dynamic pulmonary computed tomography (CT) enables morphological analyses of tumor adhesion and infiltration and functional analyses of the lungs based on four-dimensional data. However, the functional analysis requires visualization of the respiratory cycle. The aim of the present study was to investigate the utility of spectral analysis as part of the functional analysis of the lungs based on dynamic pulmonary CT. In this study, the reference curves for the respiratory cycle were obtained using measurements of all phases of respiration based on the movement of the diaphragm. The reference curves and fields of the unaffected lung were divided into three sections: upper, middle, and lower. The central position within each lung field in the axial section was used as the fixed location, and the lung field concentrations (CT values within each lung field) were measured. Using the maximum entropy (ME) method, the spectral analysis was performed for the lung field concentration curves obtained in this manner. The investigated items were the peak frequency in the power spectrum based on the ME analysis of the reference curve and the time difference from this peak frequency in the upper, middle, and lower lung field concentration curves. The time differences (median±standard deviation) from the reference values were 0.18±0.20, 0.34±0.33, and 0.34±0.35 s in the upper, middle, and lower lung fields, respectively, indicating the smallest time difference in the upper lung field. Performing spectral analysis using the ME method on lung field concentration curves enables assessment of the respiratory cycle based on dynamic pulmonary CT, and this approach is consistent with the visual assessment of the respiratory cycle.