Room: AAPM ePoster Library
Purpose: Simulate patients’ respiratory trends during breathing-gated carbon radiotherapy and produce a respiratory curve for breathing control.
Methods: Ten patients with thoracic cancer (lung:8; trachea:1; esophagus:1) were involved in this retrospective study. All patients underwent respiratory-gated therapy in our center. During treatment, the patient’s breathing amplitude signal was digitized by an Anazi-733V pressure sensor (Anazi Medical, Japan) located on the contralateral abdomen. The signal was recorded 40 times per second. For each patient, breathing records were acquired during the planning CT and in 14 following fractions. Each record was chronologically divided into five equal periods. For each period, a fast Fourier transformation (FFT) was performed on the breathing data to analyze the signal frequency distribution. The first peak of the spectrum for each period was recognized as the patient’s breathing frequency (BF). A sine function was also fit to the planning CT breathing data to draw a simulated respiratory curve and estimate the beam-on time (BOT).
Results: For the first few fractions, the patient breathing was generally erratic. For subsequent fractions, the BF stabilized. The FFT analysis showed the mean BF value to be 0.327±0.056Hz during the first five fractions but lower by 13.1%(1.5%-64.5%) during the last five fractions. The average BF value was 0.300±0.053Hz during planning CTs and 0.309±0.067Hz during treatments. A different fitting function was applied directly on the planning CT respiratory signals using the following formula: breathing amplitude =a*sin(2*p*BF*breathing time+c). According to this fitting, the mean value of BF was equal to 0.29 Hz. The estimated average BOT using the fitted data was 0.71s when a gating window of ex20% to in20% was applied.
Conclusion: Patients’ BF become more stable after several treatments. By analyzing the respiratory data taken during the planning CT, a virtual respiratory curve can be generated for personalized breathing training.