Cone Beam CT (CBCT) Guided Tumor Positioning for Laryngeal Cancer Patients With Large Interfraction Time Trends in Tumor Setup

Abstract

Listen

Translate article

Comparing CBCT guided set-up correction protocols for laryngeal cancer patients with large interfraction time trends in primary tumor set-up relative to the treatment isocenter. For 30 laryngeal cancer patients (2 T1, 7 T2, 16 T3, 5 T4), cone beam CT measured displacements in the fractionated treatments were used to establish residual set-up errors for 1) the No Action Level (NAL) off-line protocol: CBCT only acquired in the first 3 fractions, mean displacement vector in these fractions used for corrections in all later fractions, 2) extended NAL (eNAL): NAL with weekly follow-up CBCT for weekly adjustment of the correction vector, and 3) daily CBCT acquisition with on-line analysis and patient re-positioning prior to dose delivery. Without set-up corrections, large Superior-Inferior (SI) interfraction time trends in primary tumor set-up would have occurred (46% of patients ≥ 8 mm/7 weeks). Because of the trends, application of the often used NAL would have led to large residual errors (required SI margin around the primary tumor: 6.7 mm). Due to the weekly correction vector adjustments, the trends could be largely compensated with eNAL (SI margin 3.5 mm). Correlation between movements of the primary and nodal CTVs in SI direction was poor (r2 = 0.15). For this reason, even with perfect on-line set-up corrections of the primary CTV (no residual errors), the required SI margin for the nodal CTV would be as large as 7 mm. In Left-Right (LR) and Anterior-Posterior (AP) directions, movements of the primary and nodal CTVs were largely correlated (r2 = 071/0.64). Even with the two off-line protocols, required PTV margins in LR and AP directions were limited to 2-3 mm for both CTVs. SI motion correlation between the cervical vertebrae and the primary CTV was poor ((r2 = 0.19). When using the vertebrae as reference structure in image guidance, the required SI margin around the primary CTV would be 7 mm, even with daily on-line repositioning. Without set-up corrections, laryngeal cancer patients showed large SI interfraction time trends in set-up of the primary tumor that could partially or almost fully be compensated with eNAL and daily on-line, respectively. Due to poor SI motion correlation of the primary tumor and the nodes, even perfect daily repositioning of the primary CTV required a 7 mm PTV margin around the nodes. Small nodal PTV margins would require an adaptive approach to correct for the non-rigid motion. Even with daily on-line corrections, the required SI PTV margin around the primary CTV would be as large as 7 mm, when using the vertebrae as reference structure in image guidance. Generally applied 5 mm PTV margins in head and neck cancer were often inadequate for the studied laryngeal cancer patients.