Characterization of free breathing patterns with 5D lung motion model.

PURPOSE: To determine the quiet respiration breathing motion model parameters for lung cancer and nonlung cancer patients.
METHODS: 49 free breathing patient 4DCT image datasets (25 scans, cine mode) were collected with simultaneous quantitative spirometry. A cross-correlation registration technique was employed to track the lung tissue motion between scans. The registration results were applied to a lung motion model: X(-->) = X(-->)0 + alpha(-->)v + beta(-->)f, where X(-->) is the position of a piece of tissue located at reference position X(-->)0 during a reference breathing phase (zero tidal volume v, zero airflow f). alpha(-->) is a parameter that characterizes the motion due to air filling (motion as a function of tidal volume v) and beta(-->) is the parameter that accounts for the motion due to the imbalance of dynamical stress distributions during inspiration and exhalation that causes lung motion hysteresis (motion as a function of airflow f). The parameters alpha(-->) and beta(-->) together provide a quantitative characterization of breathing motion that inherently includes the complex hysteresis interplay. The alpha(-->) and beta(-->) distributions were examined for each patient to determine overall general patterns and interpatient pattern variations.
RESULTS: For 44 patients, the greatest values of /alpha(-->)/ were observed in the inferior and posterior lungs. For the rest of the patients, /alpha(-->)/ reached its maximum in the anterior lung in three patients and the lateral lung in two patients. The hysteresis motion beta(-->) had greater variability, but for the majority of patients, /beta(-->)/ was largest in the lateral lungs.
CONCLUSIONS: This is the first report of the three-dimensional breathing motion model parameters for a large cohort of patients. The model has the potential for noninvasively predicting lung motion. The majority of patients exhibited similar /alpha(-->)/ maps and the /beta(-->)/ maps showed greater interpatient variability. The motion parameter interpatient variability will inform our need for custom radiation therapy motion models. The utility of this model depends on the parameter stability over time, which is still under investigation.