3D Fabrication and Multiscanner CT Validation of a Novel Patient-Specific Anthropomorphic Pediatric Head Phantom.

RATIONALE AND OBJECTIVES: Anthropomorphic pediatric computed tomography (CT) phantoms are valuable tools for quantitative imaging validation and protocol optimization, however, available models are often generic, anatomically simplified, or difficult to reproduce, limiting their ability to represent patient-specific pediatric anatomy. This study aimed to develop a patient-specific, heterogeneous anthropomorphic pediatric head phantom using an indirect fabrication strategy and to validate its radiological performance across multiple CT scanners. MATERIALS AND METHODS: A retrospective head CT dataset from a 5-year-old child was segmented to generate models of the skull, mandible, teeth, brain, ventricles, eyes, cervical vertebrae, and spinal canal. Polylactic acid positive models and silicone negative molds were used to cast compartment-specific tissue-equivalent epoxy composites, while polyurethane foam preserved physiological air cavities. Material attenuation was assessed using Geant4 Application for Tomographic Emission (GATE) Monte Carlo simulations and XMuDat data from 10 to 150 keV. The completed phantom was scanned on four CT systems from three vendors. CT numbers, measurement reproducibility, and interscanner agreement were evaluated. RESULTS: GATE and XMuDat showed close agreement. At 10 keV, bone- and teeth-equivalent materials had the highest mass attenuation coefficients, at 13.99 and 18.01 cm² g⁻¹, respectively, while all materials converged to approximately 0.14-0.15 cm² g⁻¹ at 150 keV. Across scanners, osseous structures ranged from 512.0 to 1285.0 HU, soft-tissue regions from 19.1 to 65.3 HU, cerebrospinal fluid from 1.2 to 17.9 HU, and air cavities from -1023.9 to -903.7 HU. Overall interscanner agreement at 80 kVp was high (intraclass correlation coefficient = 0.992). CONCLUSION: The proposed workflow produced an anatomically realistic, radiologically heterogeneous pediatric head phantom with stable cross-platform CT behavior, supporting quantitative validation, scanner harmonization, and protocol-optimization studies.