Dosimetric assessment of patient‐specific breath‐hold reproducibility on liver motion for SBRT planning

PURPOSE: To investigate the impact of breath‐hold reproducibility on liver motion using a respiratory motion management device. METHODS: Forty‐four patients with hepatic tumors, treated with SBRT with breath‐hold, were randomly selected for this study. All patients underwent three consecutive computed tomography (CT) scans using active breath‐hold coordinator (ABC) with three repeated single breath‐hold during simulation. The three CT scans were labeled as ABC1‐CT, ABC2‐CT, and ABC3‐CT. Displacements of centroids of the entire livers among the three ABC‐CTs were measured as a surrogate for intrafractional motion. For each patient, two different treatment plans were prepared: (a) a clinical plan using a 5‐mm expansion of an ITV that encompassed all three GTVs from each of the three ABC‐CTs, and (b) a research plan using a 5‐mm expansion of the GTV from only ABC1‐CT to create PTV. The clinical plan acceptance criteria were that 95% of the PTV and 99% of the GTV received 100% of the prescription dose. Dosimetric endpoints were analyzed and compared for the two plans. RESULTS: All shifts in the medial‐lateral direction (range: −3.9 to 2.0 mm) were within 5 mm while 7% of shifts in the anterior–posterior direction (range: −10.5 to 16.7 mm) and 11% of shifts in the superior–inferior direction (range: −17.0 to 8.7 mm) exceeded 5 mm. Six patients (14%) had an intrafraction motion greater than 5 mm in any direction. For these six patients, if a plan was created based on a PTV from a single CT (ex. ABC1‐CT), 5 of 12 GTVs captured from other ABC‐CTs would fail to meet the clinical acceptance criteria due to poor breath‐hold reproducibility. CONCLUSIONS: Non‐negligible intrafractional motion occurs in patients with poor breath‐hold reproducibility. To identify this subgroup of patients, acquiring three CTs with active breath‐hold during simulation is a feasible practical method.