ESTRO 2023 - Abstract Book

S286

Sunday 14 May 2023

ESTRO 2023

( r ) was calculated between HU and the average/ standard deviation of the normalised dose metrics for each structure. The patient group was divided into two sub-groups for analysis based on tumour location relative to the kidneys (unilateral vs midline). Lastly, the Homogeneity Index (HI) was calculated for the CTV as D 5% /D 95% . Results Tab.1 shows the r values between simulated HU and the average/ standard deviation of all normalised dose metrics. The high correlations found indicate that an increase in GI air contents led to increased underdosage ( r <0) of targets and overdosage ( r >0) of OARs (kidneys, liver). Moreover, the standard variation across the patient group for each dose metric was overall larger in scenarios simulating increased gas contents ( r >0). Fig.1 shows the linear regression for CTV D 95% and kidneys D mean . The CTV D 95% was reduced in the P -500 scenario by an average of 1.3±1.9 (0.1–6.9)%; while the kidneys D mean increased by 5.5±16.1 (0.3-68.2)%. Dose distributions within the CTV became less homogeneous as the GI air contents increased with HI = {1.05±0.01, 1.06±0.01, 1.05±0.01, 1.05±0.01, 1.07±0.03, 1.08±0.03, 1.10±0.05} for HU = {nm, 50, 0, - 100, -500, -700, -1000}. Midline tumours experienced overall larger changes for most dose metrics than unilateral tumours, with differences between the groups becoming clear from HU = -500 (Fig.1a). One exception was liver V 19Gy which was more robust to bowel contents variation for midline tumours.