ESTRO 2024 - Abstract Book

S4256

Physics - Intra-fraction motion management and real-time adaptive radiotherapy

ESTRO 2024

treatment took place in DIBH guided by an external gating block placed on the thoracic cage. The target volume included mediastinal lymph nodes with implanted fiducial markers. Patient setup was guided by a CBCT, recorded during typically 3-4 breath-holds and matched on the primary lung tumor. CBCT projections were only recorded when the marker block was inside the 2mm wide gating window. A post-treatment CBCT was recorded at the first three fractions and then at every fifth fraction throughout the treatment course. Retrospectively, the implanted markers were segmented in all CBCT projections. The 3D marker motion in patient coordinates during CBCT acquisition was determined by a probability-based method and used as surrogate for the lymph node motion. For all BHs of at least 12s duration, the intra-BH drift motion was determined as the difference between the mean marker position in the last 3s and the first 3s of the BH (Figure 1, yellow). For all consecutive pairs of BHs of at least 12s duration, the inter BH drift motion was determined as the difference in mean marker position between the two BHs (Figure 1, blue). For all fractions with a post-CBCT, the pre-to-post CBCT drift motion was determined as the difference in mean marker position between the two CBCTs after accounting for the image-guided couch correction performed after the pre CBCT (Figure 1, green).

Results:

The analyzed dataset encompassed the 3D motion recorded at 11-15 Hz during 604 DIBH CBCTs. Intra-BH drift was determined for 1634 BHs, inter-BH drift for 1176 consecutive BH pairs, and pre-to-post CBCT drift for 136 fractions. Table 1 summarizes the population mean (±standard deviation) of the 3D drift motion on the three investigated time scales, while the distribution of shifts for each individual patient in the cranio-caudal (CC) direction is shown as box plots in Figure 2. Significant (p << 0.001) systematic drift motion occurred during BH in the cranial direction (mean: 1.2mm) and during the treatment session between pre-CBCT and post-CBCT in the cranial (mean: 1.1 mm) and posterior (mean: -0.7 mm) directions. This drift motion was highly patient-dependent (Figure 2) with mean values