Production

To analyse the sample quality of the CM and MM, a set of ten samples was produced for each method. We used a commercially available PVA (substance: Kuraray Poval^® 15-99 PVA, Kuraray Europe GmbH, Hattersheim am Main, Germany). The PVA (10 wt.%) was dissolved in distilled water. Samples were produced according to different methodologies specified by the following parameters:

• CM: utilizing a water bath with a beaker on a hot plate (operating at a temperature of 95 °C, heating power of 630 W) and
• MM: employing a microwave with a polypropylen container (operating at a power of 300 W).

For both methodologies, evaporated water was replaced after dissolution of PVA. The samples were prepared using polypropylene containers as moulds. Each sample had a volume of 40 ml. Processing finalized by two freeze-thaw cycles (Fig. 1).

Fig 1: Examples of PVA-C samples produced by the microwave (left) and by the...

Scanning

CT imaging was performed using a standard CT (Fig. 2) used for diagnostic imaging and fluoroscopic guided interventions (Somatom X.Cite, Siemens Healthineers, Erlangen, Germany).

Fig 2: Test measurement setup in the Somatom X.Cite CT (different PVA-C samples in 40...

The CT-protocol represents the manufacturer standard for abdominal imaging with:

• pitch p = 0.8,
• rotation time t = 0.5 s,
• collimation of 64 x 0.6 mm,
• X-ray tube voltage of 120 kV (automated tube voltage function CARE kV was switched off) and
• automatic exposure control (CARE Dose4D).

Images were reconstructed using an iterative CT algorithm (ADMIRE, Level 3) with and a filter kernel optimized for soft tissue (Br40d).

Evaluation

Image analysis of the reconstructed slices (Fig. 3) was performed by using MATLAB (version: 2022b, The MathWorks, Inc., Natick, Massachusetts, USA).

Fig 3: Reconstructed PVA-C slice from microwave (top) and hot plate method (bottom),...

Each sample was segmented by a semi-automatic algorithm employing intensity-based thresholding. This segmentation aimed to distinguish the PVA-C material from the surrounding material (e.g. air and the sample container). To assess the HU values, a cubic volume of interest (VOI, volume = 1 cm³) was positioned in the centre of each sample (n = 1458 values inside the VOI). The segmentation and positioning of the VOI were verified by visual inspection. Data were tested for normal distribution by Kolmogorov-Smirnov test. Wilcoxon rank sum test was used to identify significant differences regarding processing methodology (CM vs. MM).