Part-solid pulmonary nodule phantoms with realistic morphology and densities by stereolithography-based 3D-printing:...

Oversimplified pulmonary nodule phantoms limit the clinical translation of computed tomography (CT) research. Therefore, we manufactured and preliminarily validated patient-realistic part-solid nodule models with heterogeneous radiodensities using a stereolithography apparatus (SLA) benchtop three-dimensional (3D) printing.

Patient-derived CT images were utilised upon Ethics Committee approval to determine part-solid nodule HU ranges and morphologies. To lower the density of the applied resin (Formlabs Clear V4), we designed variable 3D-beams (340, 510, or 680 µm) into lattice structures with variable gap thicknesses (from 680 to 2,040 µm). These lattice structures were merged with patient-derived nodule segmentations. The 3D-prints were incorporated in a Lungman phantom and evaluated using low-dose CT imaging. A multicentre, single-blinded reader study, involving seventeen radiologists, assessed whether 3D-printed nodules were distinguishable from real part-solid nodules using χ² goodness-of-fit testing.

Through varying combinations of material thicknesses and void sizes, we reproduced multiple target radiodensities of clinical nodules and successfully manufactured pulmonary nodule phantoms consisting of three different ground-glass components around a solidly printed core. There was significant evidence (χ² = 136.13; p = 1.864e-31; 5% confidence level) against readers reliably distinguishing patient nodules from our 3D-printed models. Average accuracy across all radiologists was 53.5%. Moreover, 47.5% of the 3D-printed nodules were incorrectly classified as real nodules.

Our SLA 3D-printing workflow produces patient-realistic part-solid pulmonary nodules that are more cost-effective than commercially available counterparts. This methodology could provide customisable ground truth phantom models for CT imaging studies, including software validation, acquisition and reconstruction parameter optimisation and/or image quality evaluation.

This study marks the first successful application of SLA 3D-printing to manufacture part-solid pulmonary nodule phantoms, incorporating multiple radiodensities and mimicking patient-realistic morphologies. Our developed methodology offers potential to 3D-printed phantoms with higher degrees of customisation and adaptation to research-specific objectives in CT imaging compared to commercially available standardised phantoms.

Part-solid pulmonary nodule phantoms with patient-realistic morphologies and multiple radiodensities were manufactured using benchtop SLA 3D-printing. Clinical relevance of our 3D-printed nodules is demonstrated and statistically substantiated in a multicentre, single-blinded reader study including seventeen reading radiologists. Our methodology renders pulmonary nodule models that overcome limitations of generic, standardised, commercially available phantoms, often lacking complexity and realism. Manufactured nodule phantoms can provide an absolute ground truth for software training and validation, CT protocol optimisation and (image) quality assurance. Our modified 3D-printing method is readily available to other groups and can be customised to specific research applications.