Program Information
Improving External-Internal Correlation by Correcting Respiratory Phase Shift Based On Simultaneous Internal and External Motion Waveforms
A Milewski*, D Olek , A Rimner , G Li , Memorial Sloan Kettering Cancer Center, New York, NY
Presentations
TH-AB-205-5 (Thursday, August 3, 2017) 7:30 AM - 9:30 AM Room: 205
Purpose: To study the reproducibility of patient-specific respiratory phase shift and to determine whether phase shift correction improves the reliability of an external motion surrogate.
Methods: Internal navigator and external bellows waveforms were acquired simultaneously during respiratory-correlated four-dimensional magnetic resonance imaging (RC-4DMRI) scans lasting 5-15 minutes. Ten volunteers participated in this IRB-approved protocol study. The navigator was placed at the lung-diaphragm interface while the bellows was placed 5cm inferior to the sternum. Segments of the respiratory waveforms were analyzed at the beginning, middle, and end of each scan, with 3-10 minutes intervening between the first and last segments. Five fitting techniques were used to quantify the phase shifts, including principal component analysis (PCA), phase space ellipsoidal fitting, waveform cross-correlation, fast Fourier transform, and Hilbert transform. The calculated phase shifts were compared among all three waveform segments to determine whether an individual’s respiratory phase shift is reproducible across time. Upon correcting the estimated phase shift, the correlation coefficient between the internal and external waveforms increased and the ellipsoidal phase space curve flattened.
Results: The external waveform leads the internal waveform in all subjects, and the PCA method yields an average phase difference of 63°±19° among our volunteers. The PCA, phase space ellipsoidal fitting, and cross-correlation methods provide the most consistent results. Seven out of ten volunteers exhibit fairly reproducible phase shifts. The internal-external correlation coefficient, however, is enhanced in all subjects after phase shift correction and increases on average from 0.45±0.21 to 0.85±0.15 (p-value<0.001) when the PCA method is used or to 0.83±0.16 (p-value<0.001) when the results are combined from the three best methods.
Conclusion: This exploratory study provides a new look at the existing external-internal motion correlation and reveals that phase shift correction can enhance the correlation between these motions. This is potentially useful for improved clinical respiratory-gated radiotherapy.
Contact Email: