Radiopaque Three-dimensional Printing: A Method to Create Realistic CT Phantoms

Purpose To develop a method to create anthropomorphic phantoms of individual patients with high precision of anatomic details and radiation attenuation properties. Materials and Methods Inkjet cartridges were filled with potassium iodide solutions (600 mg/mL) and prints were realized on plain paper (80 g/m²). Stacks of 100 prints resulted in three-dimensional phantoms of 1 cm thickness. In a first step, reproduction of patient anatomy was tested by printing computed tomographic (CT) images of a real patient abdomen scan. In a second step, gray scales, iodine deposition, and Hounsfield units were investigated by printing geometric phantoms with gray scales ranging from 0% (white) to 100% (black). On the basis of these results, a gray-scale-correction procedure was developed to achieve realistic Hounsfield units in the patient phantom. In a third step, reproduction of the real patient's Hounsfield units was verified by printing the initial patient CT scan again after application of the gray-scale-correction procedure. Data were analyzed by using Pearson correlation, linear regression, and nonlinear regression. Results The first abdomen phantom showed a detailed reproduction of the patient anatomy and demonstrated feasibility of the concept. However, individual-organ Hounsfield units deviated from the real patient CT scan. Analysis of the geometric phantoms revealed an exponential correlation between template gray scales and printer deposition. Application of a correction procedure to the template gray scales allowed for a linear correlation (r = 0.9946; 95% confidence interval: 0.9916, 0.9966). After the same correction procedure was applied to the abdomen phantom, linear correlation of phantom and patient Hounsfield units was confirmed (r = 0.9925; 95% confidence interval: 0.9635, 0.9985). Conclusion The method presented in this work can realize realistic and customizable phantoms for diagnostic and therapeutic radiology, including the reproduction of individual patients. ^© RSNA, 2016.