ESTRO 2024 - Abstract Book

S3769

Physics - Image acquisition and processing

ESTRO 2024

Material/Methods:

The Elekta XVI system can be calibrated for HU Scaling uses calibration measurements to calculate the HU mean values. The manufacturer specifies an accuracy of ±40 HU for small field of view (SFOV) imaging under specified conditions (default presets, CATPHAN 503). This level of accuracy may be suitable for dose calculations in certain scenarios, such as adaptive planning [1]. Elekta system advises against using VolumeView™ images for treatment planning. VolumeView™ images may not provide HU with sufficient accuracy for dose planning. The XVI system was fully calibrated and optimised to ensure the system provides reliable measurements and images for calibration [2, 3, 4]. For this study, two different phantoms were scanned to mimic the characteristics of different tissues in the human body for SFOV and middle field of view (MFOV): CIRS ELECTRON DENSITY PHANTOM MODEL 062M and RMI Tissue Characterization Phantom Model 467. A customized approach with narrow beam geometry cassettes of S2 (CTDI-S) and M2 (CTDI-M) was adopted to replicate CT-fanbeam acquisition. Optimised reconstruction presets were developed for this study. The scans were performed for 100kV and 120kV with filtration setting of F0 and F1 .The phantoms were scanned with and without scatter conditions. In the absence of scatter, the phantoms slabs with inserts characterizing different tissues' electron densities were scanned. In the presence of full scatter conditions, the additional attenuator material was placed on both sides of the slab. The reconstructions were performed with and without implemented HU correction. These phantoms were used to establish the relationship between the relative electron density (RED)/mass density (ρ/cc) of various tissues and their corresponding CT number in Hounsfield Units (HU)/pixel values. The outcomes of measured CT numbers/pixel values of tissue equivalent materials together with their relative electron densities were used to create the calibration curves for all scans taken for the CIRS (with and without scatter conditions) and RMI phantoms (without scatter conditions), under different exposure conditions (presets, kVp, filter, mA, ms), S2 and M2 collimators. It was found that the differences between scan conditions were: • insignificantly different for 100kV and 120kV, • insignificantly different for F0 and F1 filters, • insignificantly different for implemented HU correction to reconstruction vs no correction applied, • insignificantly different for scatter and not scatter conditions. Insignificancy was identified per SD of the pixel values (RMI SFOV SD=30.48-54.56, RMI MFOV SD=30.15-46.44, CIRS SFOV SD= 19.42-38.60, CIRS MFOV SD= 18.70-30.16) read in the established ROIs (0.71cm) in the inserts of reconstructed cross-sections of the phantoms, averaged for 3 slices. The results were averaged to receive the HU/pixel value = f (ED) calibration curves for RMI (CBCT SFOV), RMI(CBCT MFOV), RMI (CT), CIRS(CBCT SFOV), CIRS(CBCT MFOV), CIRS (CT). Figure 1 displays the obtained results/curves and the curve for Monaco TPS. Results: