Session Item

CT images are commonly used for modelling of the patient for the absorbed dose calculation in external beam radiotherapy. The dose calculation is sensitive to variations in density and effective atomic number (EAN) of the tissue. The mass density (MD) and relative electron density (RED) images are usually established from CT-numbers (Hounsfield units) using pre-defined CT-calibration curves. New techniques are now available where information about material properties can be provided directly from the CT software. This study compares MD and RED obtained with single-energy CT (SECT) as well as RED and EAN obtained with dual-energy CT (DECT) to theoretical reference values for materials with known MDs and elemental compositions.

The head sections of two electron density phantoms (Model 1467, Sun Nuclear, SN) and (062M, CIRS) were placed inline in the holder belonging to the first phantom (Figure 1). The SN phantom was setup with its tissue surrogate rods, and a cylindrical container with 5 mg/cm³ iodine solution. The CIRS phantom was setup with its tissue surrogates, as well as some non-tissue equivalent materials (Table 1). The phantom setup was scanned with a radiotherapy specific CT scanner (SOMATOM go.Open Pro, Siemens Healthineers) with a SECT (120 kV) and a dual-spiral DECT (80 kV/Sn140 kV) protocol. The SECT images were reconstructed with the Sd40 and Sm40 (DirectDensity™ (DD), Siemens Healthineers) kernels which provide RED and MD images, respectively. The RED was also estimated from conventional SECT images (Qr40 kernel) in a more traditional way using a pre-defined CT-calibration curve created based on the CIRS phantom. The DECT images were converted to RED and EAN images in the Rho/Z application in (syngo.via.VB50A, Siemens Healthineers). The values obtained with the different methods were compared to the reference MD and theoretical reference RED and EAN of the phantom inserts.

Figure 1. The head section of the CIRS 062M phantom is shown in A) and an opposite view of the setup that shows the Sun Nuclear 1467 phantom is shown in B).

Table 1. The reference values of the MD, RED and EAN, as well as the relative differences of the studied methods compared to the reference values. Deviations >±3% are written in red. Non-tissue equivalent materials are shaded in grey. 

Both SECT methods underestimated the RED and MD for PMMA, graphite and PTFE (Table 1). The RED and MDs of the iodine solution were overestimated for the SECT methods, although less using the traditional calibration compared to DD. However, the traditional CT-calibration resulted in an underestimation of the RED of the three most compact bone surrogates in the SN phantom with 4%. DECT provided RED values within 1.4% for all materials in the phantom setup. Larger deviations were observed for the EAN.

Generally, the DECT method resulted in RED values closer to the references values compared to both SECT methods studied and especially for non-tissue-equivalent materials.