Use of Three-Dimensional Printing in Modelling an Anatomical Structure with a High Computed Tomography Attenuation Value: A Feasibility Study

Abstract

Introduction

Three-dimensional (3D) printing provides an opportunity to develop anthropomorphic computed tomography (CT) phantoms with anatomical and radiological features mimicking a range of patients’ conditions, thus allowing development of individualised, low dose scanning protocols. However, previous studies of 3D printing in CT phantom development could only create anatomical structures using potassium iodide with attenuation values up to 1200 HU which is insufficient to mimic the radiological features of some high attenuation structures such as cortical bone. This study aimed at investigating the feasibility of using 3D printing in modelling cortical bone with a non-iodinated material.

Methods

This study had 2 stages. Stage 1 involved a vat photopolymerisation 3D printer to directly print cube phantoms with different percentage compositions of calcium phosphate (CP) and resin (approach 1), and approach 2 using a material extrusion 3D printer to develop a cube mould for infilling of the CP with hardener as the phantom. The approach able to create the cube phantom with the CT attenuation value close to that of a tibial mid-diaphysis cortex of a real patient, 1475 ± 205 HU was employed to develop a tibial mid-diaphysis phantom. The mean CT numbers of the cube and tibia phantoms were measured and compared with that of the original CT dataset through unpaired t-test.

Results

All phantoms were scanned by CT using a lower extremity scanning protocol. The moulding approach was selected to develop the tibia mid-diaphysis phantom with CT attenuation value, 1434 ± 184 HU which was not statistically significantly different from the one of the original dataset (p=0.721).

Conclusion

This study demonstrates the feasibility to use the material extrusion 3D printer to create a tibial mid-diaphysis mould for infilling of the CP as an anthropomorphic CT phantom and the attenuation value of its cortex matches the real patient’s one.