TY - JOUR T1 - <strong>Development of a dynamic lung phantom for use in lung ventilation studies</strong> JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 3306 LP - 3306 VL - 63 IS - supplement 2 AU - Glenn Woolley AU - Chinedum Ekpemiro AU - Nathaniel Brown AU - Louise France AU - Graham Wright AU - Steve Archibald Y1 - 2022/06/01 UR - http://jnm.snmjournals.org/content/63/supplement_2/3306.abstract N2 - 3306 Introduction: The development of novel lung tracers and delivery devices are complex in their nature requiring strict control over variables to investigate any hypothesis. One of the most difficult variables to control is a representable patient breath. The recent developments in 3D printing opens new possibilities to create anatomically and physiologically accurate phantoms for use in nuclear medicine. In this work, we have developed a hollow dynamic 3D lung phantom that simulates respiration based on anatomical volumes. Methods: To develop methods to create flexible 3D-printed structures we segmented the right horizontal lobe from a 1mm slice CT that showed no abnormalities. The software package ‘Aviso’ by Thermo Fisher Scientific was used to create a high-resolution triangular mesh model of the lung lobe. Basic air/tissue thresholds were used to generate the initial volume, before refining each fissure and bronchi outline on a slice-by-slice basis. A STL model was produced and imported into Solidworks (Dassault Systèmes) where the 3D volume was sampled into a set of 3mm planes. A 2mm hollow model was created with cleaning outlets and an internal support structure. A Stratasys J750 Digital printer with a blend of Agilus30™ Clear and VeroClear™ (Stratasys) material was used to print the model, before dissolving the support structure and jet cleaning. The cleaning holes were closed and the phantom was supported in a sealed Perspex cylindrical housing. A mass flow controller and vacuum plump was used to evacuate the air from the cylinder in a 2 second cycle with a flow rate of 38mL/s to mimic typical patient breathing. 1GBq of Tc-99mDTPA was nebulised for 90 seconds using a SmartVent (Diagnostic Imaging Ltd UK) mesh nebuliser. Planar dynamics and SPECT/CT were acquired on a GE Discovery 870DR throughout ventilation and at the end of ventilation respectively. A fully quantitative SPECT reconstruction was performed with standard parameters to compare output with typical patient scans. Results: The printed phantom maintained anatomical shape whilst being fully flexible and able to simulate a breathing pattern. Dynamic imaging showed a good ventilation profile throughout the breathing cycles. Quantitative SPECTCT showed that 0.7MBq of Tc-99m was delivered to the lung lobe. When scaled to represent total lung volume (right horizontal lobe ~10% of total lung volume) this represents 7MBq of delivered tracer ~40% of the activity typically used in our institution for ventilation studies. The dynamic images showed good ventilation with the aerosol filling the entire volume. The time activity curve revealed that a longer nebulisation period could have increased the delivered activity and that the tracer was retained. Most of the tracer settled in the bottom of the lung since there were no internal structures. Conclusions: Anatomically and physiologically accurate lung phantoms can be manufactured using flexible 3D printing materials. Phantoms are able to simulate inspiration and show the potential for evaluating new tracer and delivery technology outputs. Further work is required to match patient deposition through the development of flexible 3D printed internal structures. ER -