ESTRO 2023 - Abstract Book

S1659

Digital Posters

ESTRO 2023

Preliminary results obtained for surface guided ED prepositioning system show, in average, submillimeter accuracy for lateral and longitudinal axis for patients immobilized with 4PI stereotactic mask. However, less accuracy has been found for vertical axis, probably because of the external contour generation inaccuracy due to the thermoplastic mask. Consequently, it is necessary internal information from Cone Beam Computed Tomography or kV X-Ray images to achieve total submillimeter accuracy necessary for example in radiosurgery treatments.

PO-1910 Target-driven initial optimization planning for simulation-free, stereotactic abdominal adaptive RT

A. Price 1 , E. Laugeman 2 , J. Schiff 2 , P. Samson 3 , G. Hugo 4 , T. Zhao 5 , L. Henke 5

1 Washington University, Radiation Oncology, Saint Louis, USA; 2 Washington University in St. Louis, Radiation Oncology, Saint Louis, USA; 3 Washington University in St. Louis, Radiation oncology, Saint Louis, USA; 4 Washington University in St. Louis, RAdiation oncology, Saint Louis, USA; 5 Washington University in St. Louis, Radiation Oncology, St. Louis, USA Purpose or Objective We present an in silico analysis of an efficient adaptive (ART) treatment planning workflow to enable initiation of treatment delivery <1 hour after initial patient consult for CBCT-guided stereotactic abdominal ART using an existing commercial platform and tools, template-based ART planning, and visualization capabilities on a novel, 6-second-acquisition CBCT aboard a ring gantry linac. Here we evaluate a simulation-free workflow only requiring a physician to contour an abdominal target on the diagnostic CT scan for pre-plan creation, followed by a full comprehensive contouring session and plan finalization through a standard, online ART workflow at the time of first treatment. Materials and Methods Three abdominal patients (2 pancreas, 1 omental metastasis), imaged on a novel CBCT were included. Patients concurrently received standard, clinical ART, separate from this imaging and in silico workflow study. To decrease time from consult to treatment, in silico, the sim and OAR contouring processes were removed. An attending radiation oncologist contoured only the target (GTV/CTV) on each patient’s diagnostic image. Then, an abdominal-based planning template, extracted from a multi-institutional abdominal SBRT study, was preloaded for all patients in a commercial ART treatment planning system (TPS; Varian ETHOS). These templates include OARs, boolean structures, and planning structures to be contoured/generated during the ART process with the patient on the table. VMAT plans with two full-arcs were used. The simulated prescription was 50Gy/5fx. Figure 1 is a step-by-step breakdown of this workflow; OAR objectives are listed in Figure 2. A patient’s image acquired in 6s on a novel CBCT panel (Varian Hypersight) was injected into the ART emulator system. AI was used to define the liver, duodenum, and stomach. All other structures were manually contoured. A final, ART-based plan was developed using the target defined on the diagnostic scan and the template that had been defined at the time of pre-planning (Fig 1). After plan delivery, sim-free plans were compared to traditional plans made from planning CT using the same abdominal planning template. All sim-free steps were timed for analysis.