Yatman Tsang1, Julie Duong2, Gareth Jackson3, Ramona DeSouza1, Kingsley Ko4, Prasana Nariyangadu3, Catharine Clark5
1Mount Vernon Cancer Centre, Radiotherapy, Northwood, United Kingdom; 2Mount Vernon Cancer Centre, Radiotherapy, Northwood , United Kingdom; 3Mount Vernon Cancer Centre, Radiotherapy Physics, Northwood, United Kingdom; 4University College London, Department of Medical Physics & Biomedical Engineering, London, United Kingdom; 5UCLH NHS Foundation Trust, Radiotherapy Physics, London, United Kingdom
Currently, no standard consensus in dose reporting in lung stereotactic body radiotherapy (SBRT) has been widely adopted yet and there is limited published literature on the correlation of Monte Carlo (MC) based lung SBRT dose distributions and patient’s physical lung density. Against this background, this study aims to investigate the relationships between the MC calculated dose parameters and the lung densities around target volumes of patients who received lung SBRT.
A retrospective review on lung function test results of all patients diagnosed with early-stage lung cancer (T1-2N0M0) who were treated in between January 2018 to December 2018 at our institution was conducted. All patients were treated with real-time motion tracking using a dedicated robotic stereotactic radiotherapy machine. All radiotherapy plans were optimised and calculated with MC calculation algorithms.
For each lung SABR plan, mean hounsfield units (HUmean) of the gross tumour volumes (GTV), planning target volumes (PTV), and PTV minus GTV were used as surrogate volumes representing the low-density components around target volumes in the lung SBRT plans. Spearman correlation tests were performed to establish the correlations between the densities of surrogate volumes and patient’s physical lung function (FEV1). The HUmean of the surrogate volume with the best correlation to the FEV1 were used to divide the cohort in to two groups: <=median and >median. Mann Whitney U tests were used to investigate if there were any statistically significant differences in the dosimetric parameters between the two groups.
This study included 42 lung SBRT plans for the analysis. The treatment schedules were either 54Gy in 3 fractions (83%) or 50Gy in 5 fractions (17%) with the range of prescription isodose from 71% to 77%. The median (range) of GTV and PTV volumes were 8.1cc (1.6 to 77.3) and 31.5cc (10.4 to 134.6) respectively.
The median (range) HUmean of GTV, PTV and PTV minus GTV were -294.5 (-29.7 to -658.6), -537.6 (-179.5 to -832.7) and -622.5 (-229.1 to -854.2). PTV minus GTV was suggested to be the surrogate volume with the highest correlation (rs=0.42, p<0.05) to FEV1 (Figure 1).
As illustrated in table 1, statistically significant differences were found in GTV D50, GTV D2, PTV D2, gradient index and prescription isodose between the two groups of HUmean PTV minus GTV (p<0.05). It is noted that plans were usually prescribed to lower isodose to improve the dose coverage in the group of <=median HUmean PTV minus GTV.
This study suggested that a moderate correlation exits between the density of surrogate volume PTV minus GTV and the patient’s physical lung function. This surrogate volume can be potentially used to proactively indicate planning difficulties due to low lung densities around the target volumes for patients with specific lung characteristics.