Prediction error and required internal margin provided for irregular respiratory movements: a phantom study.

Abstract

To estimate the prediction error and required internal margin provided for irregular respiratory movements. Twenty-eight patterns of irregular movement were prepared using a moving phantom. For irregular cycle movement, a cycle of 2.0-5.0 s was inserted into the baseline movement data (3.5-s cycle, 6-mm amplitude) every four cycles. In addition, a cycle of 2.5-4.5 s was further inserted into one of the irregular data points every four cycles to produce more complicated irregularity. For irregular amplitude movement, an amplitude of 3.0-9.0 mm was inserted into the baseline respiratory movement data. In addition, an amplitude of 4.0-8.0 mm was further inserted into one of the irregular data points every four cycles to produce more complicated irregularity. Images of the moving target were acquired using a real-time position management (RPM) and four-dimensional CT (4DCT) system. The displacement from the baseline image and required margin to compensate for irregular movement were calculated. Amplitude irregularity tended to lose stable placement and need larger margins to compensate for the reduction of image reproducibility than cycle irregularity. There was a large displacement and required margin when the target moved with more complicated irregular amplitude. The RPM and 4DCT system has a risk of prediction error, which may result from the complicated amplitude irregularity of respiratory movement.