2965 - Motion Amplitude Variability during Pancreatic Cancer Treatment and its Association with Breathing Irregularity

Purpose/Objective(s): Tumor motion amplitude can be estimated from 4D CT for treatment planning; however, its variability over the course of treatment remains unclear. The recently developed Elekta Comprehensive Motion Management (CMM) system for MR-Linac provides real-time, high-resolution tumor motion monitoring from 2D cine images, enabling quantitative assessment of motion amplitude variability and informing margin selection.

Materials/Methods: Twenty pancreatic cancer patients treated on an MR-linac with CMM (301 total fractions) were retrospectively analyzed. Tumor motion on 4DCT was 8.8 ± 3.8 mm. Tumor positions during treatment were extracted from CMM log files. For each fraction, the oscillatory respiratory component was isolated by removing baseline drift using a low-pass filter. Peak (exhale) and valley (inhale) positions were identified for each breathing cycle, and motion amplitude was calculated from paired inhale–exhale positions. Uncertainties in amplitude, inhale, exhale, and other respiratory phases were computed. A Fast Fourier Transform (FFT) determined the fundamental breathing frequency and power spectrum. Breathing regularity was quantified using spectral concentration near the frequency peak, spectral entropy, and peak width (full width at half maximum). Pearson correlation evaluated relationships between breathing irregularity metrics and motion amplitude uncertainties.

Results: Motion amplitude ranged from 5.1 to 16.9 mm (mean 9.1 ± 2.9 mm). Tumor position uncertainty (SD) was 1.1 ± 0.5 mm at exhale and 1.9 ± 0.8 mm at inhale; inhale uncertainty was significantly larger (p < 0.0001) and strongly correlated with exhale uncertainty (r = 0.82, p < 0.001). Phase analysis showed the largest uncertainty at inhale, gradually decreasing across phases toward exhale. Inhale uncertainty corresponded to 21.7% ± 7.2% of motion amplitude (range 12.7%–36.4%). Amplitude uncertainty showed a moderately strong correlation with average motion amplitude (r = 0.66, p = 0.001), indicating greater variability with larger motion. FFT peak width correlated with relative uncertainty (uncertainty/amplitude) for both inter- and intra-fraction measures (r = 0.573, p = 0.008; r = 0.70, p < 0.001). Spectral concentration correlated inversely with intrafraction exhale and inhale uncertainties (r = -0.574, p = 0.008; r = -0.72, p < 0.001). These findings indicate that more regular and stable breathing patterns are associated with smaller motion uncertainties.

Conclusion: Inter- and intra-fraction variations in respiratory motion amplitude were quantified in real time using the CMM system, providing clinically relevant information for margin selection, particularly when gating or tracking is unavailable. Breathing irregularity, as quantified by FFT peak width and spectral concentration, is strongly associated with increased motion uncertainty and may serve as a practical guide for margin adaptation for individual patients.