Verification of Breath-Hold Technique in Stereotactic Body Radiation Therapy for Lung Cancer: What Are the Optimal Margins?

Abstract

To reduce the irradiated volume in stereotactic body radiotherapy (SBRT) for lung cancer, breath-hold (BH) techniques may be more effective than tracking or gating techniques. However, the optimal margin in the technique has not been established yet. In this prospective study (UMIN 000026080), we evaluated the optimal margin in our BH technique. The patient eligibility criteria were: 1) 20 years old or older; 2) primary stage I lung cancer or 3 or less solitary metastatic lung cancer; 3) BH capability for 15 seconds or longer following our sign; and 4) target respiratory motion of 10 mm or more. This study was performed from 2015 to 2017 at a single institution. Computed tomography (CT) simulation scans for treatment planning were taken 1-2 weeks after placement of fiducial gold markers near the target. Patients underwent respiratory training to hold the breath at the expiratory end phase. CT images for treatment planning were obtained with breath holding during the expiratory phase. The clinical target volume (CTV) was defined as the visible gross tumor volume. As the BH margin, 4, 3, and 3 mm were assumed for the superior-inferior (SI), left-right (LR) and anterior-posterior (AP) directions. In addition to the BH margins, 3-mm set-up margins were added to all directions. The planning target volume was defined as the CTV plus these margins. All treatment was delivered using a linear accelerator with an on-board imager. To confirm the reproducibility, on-board imaging was performed 10 times before each treatment. When the errors were larger than the assumed BH margins (4 mm in the SI direction, and 3 mm in the LR and AP directions), the treatment was ceased, and planning was made all over again. After each treatment, the dry run process was repeated to confirm the intra-fractional error. The prescribed doses were 50-52 Gy in 4 fractions or 60 Gy in 8 fractions. Thus, BH errors were measured using 160 or more images per patient. In total, 1260 images of 7 patients were analyzed. Six patients were male and 1 was female. Five targets were primary lung cancer and 2 were metastatic lung cancers. The median patient age was 77 years (range, 56-86). One target was located in the upper lobe and 6 in the lower lobe. The median diameter of the targets was 42 mm (range, 13-54 mm). In total, median errors (5th percentile, 95th percentile) were 0.60 mm (-0.24, 2.8 mm) in the SI direction, 0.30 mm (-0.90, 1.7 mm) in the LR direction, and -0.30 mm (-1.6, 1.9 mm) in the AP direction. The errors after treatment tended to be larger than those before treatment (p< 0.001, 0.16 and 0.004 in the SI, LR, and AP directions, respectively). According to van Herk’s formula, the calculated BH margin was 2.8 mm in the SI direction, 1.5 mm in the LR direction, and 2.5 mm in the AP direction. In SBRT using the BH technique for lung cancer, the optimal margin may be 4 mm in the SI direction, 2 mm in the LR direction, and 3 mm in the AP direction. or more.