3D-Printed Abdominal Phantom: What Does CT Reveal?

CT imaging research depends on models that make it possible to assess anatomy and image quality under controlled conditions. A useful phantom must represent relevant structures and produce X-ray attenuation similar to the tissues being studied. This paper reports an abdominal phantom made with 3D-printed components and readily available materials, and evaluates its CT appearance, including a small appendix simulation.

Key findings

• The abdominal phantom combined a 3D-printed mold and components with beeswax, epoxy resin and plaster.
• Measured CT values were −115.41 ± 20.29 HU for fat, 65.61 ± 18.06 HU for muscle and 510 ± 131.2 HU for bone.
• The fat simulation did not differ significantly from patient images (p = 0.428).
• Muscle and bone simulations had higher HU values than the corresponding patient-image values (p < 0.001).
• The bone simulation was statistically similar to that of a commercially available phantom (p = 0.063).
• Mean contrast for the appendix simulation was 818 HU at 80 kV and 881 HU at 120 kV.

Why develop an abdominal imaging phantom?

A phantom is a physical model used to represent selected properties of the human body for imaging tests, training or research. In computed tomography, image values are expressed as CT numbers, commonly measured in Hounsfield units (HU). These values help distinguish materials and tissues, but they also depend on factors such as the X-ray energy and scanning settings.

An anthropomorphic phantom has two related design challenges: its geometry should represent the anatomy of interest, and its materials should approximate the X-ray attenuation of the target tissues. Commercial phantoms provide established reference objects, but a custom model can be designed around a particular research question. Additive manufacturing, commonly called 3D printing, offers a way to make customized shapes and molds. However, a printed structure alone may not reproduce the attenuation of every tissue, which is why this study combined printed elements with casting materials.

The study set out to develop and characterize a composite beeswax–epoxy abdominal phantom, with 3D-printed parts and plaster used in its construction. One stated purpose was to analyze the appendix simulation and optimize scanning parameters. The work therefore connects product design and manufacturing choices with the practical evaluation of CT images.

Materials and fabrication approach

The abdominal mold was produced with a custom 3D printer using the selected filament. The paper says that printing parameters were optimized for dimensional accuracy, surface finish and structural integrity.

Commercial beeswax pellets were used to represent adipose tissue. The wax was melted, poured into designated cavities in the printed mold and allowed to cool and solidify. Epoxy resin was mixed with a selected muscle-mimicking filler and poured into the remaining cavities. Once that component had cured, a separate plaster mold was made within the assembled abdominal structure to form the bone simulation. A contrast agent with HU values closer to bone was incorporated into the final epoxy resin casting, which was poured into the plaster mold. The cured components were then assembled into the complete phantom.

This is a multi-material approach rather than a single-material printed model. The printed mold supports the desired geometry, while wax, resin and plaster are assigned to different regions. The paper’s evaluation focuses on CT attenuation and image presentation; it does not establish that the materials reproduce every mechanical or biological property of human tissue.

How the CT images were assessed

The appendix simulation was evaluated using three parameters: contrast, visible diameter and edge sharpness. Contrast is the difference between the mean CT number of the object and that of the background. The diameter describes the apparent width of the object in the image. Edge sharpness was represented by σ, estimated by fitting a curve to the transition from background to object. In practical terms, the analysis examined how distinctly the simulated structure appeared against its surroundings.

The study used 12 scan series. Tube voltage settings were 80, 100, 120 and 135 kV, with the listed tube currents of 300, 153, 87 and 80 mA across the respective series. Slice thickness was 0.5 or 1.0 mm. The CTDI values listed in the scan table were 3.9, 4.9, 6.0 and 7.6 mGy. The FC18 and FC08 reconstruction kernels were compared. Automatic tube current modulation was used, with a preset standard deviation of 12.5 and a dynamic anode current range of 50 mA to 300 mA.

CT results and what they show

The phantom’s fat, muscle and bone CT numbers were compared with values from patient images and a commercially available phantom. For fat, no statistically significant difference from patient images was found (p = 0.428). In contrast, the phantom’s muscle and bone simulations had higher HU values than the corresponding values in patient images (p < 0.001). The bone simulation was statistically similar to the commercial phantom’s bone simulation (p = 0.063). These findings show that performance differed by tissue type: the study does not support treating all three simulations as equally close to patient values.

The appendix analysis showed that tube voltage affected measured contrast. Mean contrast was 818 HU at 80 kV and 881 HU at 120 kV. In the individual results, contrast at 120 kV was 884 HU and 883 HU for the listed series with slice thicknesses of 0.5 and 1.0 mm. The study reported a positive correlation between tube voltage and contrast (p = 0.014). This is an important measurement detail: a HU result should be interpreted alongside the imaging conditions, rather than treated as an unchanging property of the material.

The comparison of FC18 and FC08 kernels found no significant difference in contrast (p = 0.594). There was also no significant difference in edge sharpness, described by σ, between these kernels (p = 1.000). The visible diameter did not differ significantly between kernels, and the correlation test between tube voltage and diameter gave p = 0.357. These results apply to the phantom and scan settings studied; they are not evidence that all CT scanners or protocols will behave identically.

What the work means for engineers and researchers

For mechanical engineers and product designers, the study provides an example of using 3D printing to create a custom mold and geometrically defined components, while casting materials form different phantom regions. For medical-imaging researchers, the phantom offers a test object for examining CT settings and the visibility of a small structure such as the appendix simulation. For students, it demonstrates how additive manufacturing, material selection, CT numbers and image analysis can come together in an applied research project.

A practical lesson is to evaluate each tissue substitute separately and document the scan settings. A close match for fat does not imply a similar match for muscle or bone. Likewise, the appendix contrast changed with tube voltage. For other uses, including education or product development, the phantom’s design and materials would need to suit the specific test objective.

Limitations and possible next steps

The measured muscle and bone values were higher than those in patient images, identifying areas where the material formulation could be further investigated. The experiments cover the stated CT settings and HU comparisons; they do not establish that the phantom reproduces every property of living tissue or every clinical imaging condition. Further work could examine how changes in material composition affect HU values and how consistently the measurements transfer across scanning protocols. These are potential follow-up questions, not results reported by the study.

Conclusion

This research describes a customizable abdominal phantom assembled from 3D-printed elements, beeswax, epoxy resin and plaster. Its fat simulation showed no significant difference from patient-image values, while the muscle and bone simulations had higher HU values; the bone result was statistically similar to that of a commercial phantom. The study offers a concrete example of how 3D printing and conventional casting materials can support abdominal CT research, while emphasizing the need to validate each material and imaging protocol against the intended application.

Frequently asked questions

Author: Adi Pandžić, PhD in Mechanical Engineering

This article summarises a published research paper; please refer to and cite the original paper. About the author

← All articles