Researchers at the University Hospital of Brest and the University of Western Brittany in France have developed a 3D printed lung phantom that can simulate a pulmonary embolism (PE) for comparing CT and nuclear medicine imaging.

In a study, highlighted in 3D Printing in Medicine, a radiologist and two nuclear medicine physicians judged the phantom’s images highly realistic across dual-energy CT (DECT) iodine mapping, planar scintigraphy and positron emission tomography (PET).

PE is usually diagnosed with CT angiography, while follow-up often relies on nuclear medicine. The pulmonary vascular obstruction index (PVOI), the share of lung receiving no blood flow, has been identified as an independent predictor of recurrence in unprovoked PE. 

The authors said no independent gold standard exists for comparing how each modality measures it, and that commercial phantoms have generic shapes unsuited to modeling PE.

A sealed compartment inside a printed lung

The team, including first author Pauline Tripet of the hospital’s Department of Emergency Medicine, began with a CT scan of the lung insert from a Kyoto Kagaku PET/SPECT thorax phantom. Anonymized lung segmentation data was mapped onto it through rigid and deformable registration. 

The superior segment of the left lower lobe, roughly 5% of lung volume and a common PVOI threshold, became the embolism. Two screw plugs let the healthy cavity and PE compartment be filled separately with contrast agent or radiotracer.

Because the phantom had to hold radioactive solutions, the authors chose selective laser sintering to print the nested cavity as a single watertight block. Formlabs produced the 14.3 x 24.1 x 10.4 cm part in Nylon 12 on a Fuse 1+ system at 0.11 mm layers, and it passed a leak test.

Filled with polystyrene beads to mimic lung density, the phantom was imaged with technetium-99m for scintigraphy and gallium-68 for PET, while a motorized platform simulated breathing. Reading blind, both physicians estimated the PVOI at one segment, or about 5%.

The authors also reported limits. The printed wall was 72% larger by volume than the original insert’s, the PE compartment measured 7.6% of the prototype’s volume against a 5% design, and nylon walls visible on CT meant PVOI was not quantified from CT.

The study put the cost at €770, including labor, printing, material and shipping (Formlabs supplied the print free of charge, so the figure reflected estimated market rates).