Effects of respiration on target and critical structure positions during treatment assessed with movie-loop electronic portal imaging

Purpose: To determine the extent of organ and target motion due to patient respiration during chest radiotherapy using electronic portal imaging, to examine these effects on treatment volumes and to show that simulation and treatment port films do not reflect this range of motion. Materials and Methods: Twenty four patients consisting of 17 tangential breast and 7 AP-PA lung field arrangements were imaged during daily radiation treatment. Eight to 10 sequential movie-loop images were acquired during each field of each fraction with a liquid ion chamber electronic portal imaging device (EPID). Motion relative to the reference image was assessed orthogonally to the central axis of the beam. In tangential breast images, cranial, caudad and lateral lung-chest wall landmarks were used; for AP-PA lung, visible tumor, mediastinum and bronchus. Inter and intra-fractional landmark displacements were determined through off-line analysis. Intra-fractional displacements, determined from multiple images within one fraction, indicate motion of the landmark during treatment. Inter-fractional data represents motion between treatment fractions as seen in routine portal film imaging. The effects on treatment volumes were assessed for the largest displacements using the EPID data together with CT reconstruction. Results: The mean, maximum and standard deviation (σ) for observed respiration induced displacements in the cranio-caudad (CC) and lateral directions relative to the beam are summarized both within (intra) and between (inter) fractions: These data indicate that while the mean displacements are small, the standard deviations are significant and the maximum motion observed during a fraction due to respiration may exceed 3 cm in certain cases. In addition, the intra-fractional displacements significantly exceed the inter-fractional displacements, which suggests that anatomical motion is not fully quantified in routine portal imaging. In lung treatments where the largest excursions occurred, some portion (exceeding 10%) of the GTV was noted to move beyond the treatment field during the treatment. During tangential breast treatments, increases of up to 20% in treated lung volumes were observed. Conclusions: Diagnostic, simulation and verification images used in radiation therapy are snapshots of dynamic patient anatomy. The magnitude of physiologic motion due to respiration may significantly impact on target volume coverage during the treatment fraction. Movie-loop EPID data recorded during treatments of breast and lung treatments show that while the mean displacements are small, large >3cm displacements and broad standard deviations in target or normal anatomy position exist. The implied effects on decreasing target volume treated or increased normal tissue irradiated will compromise our efforts to deliver dose-escalated, three-dimensional and dynamic radiation therapy