Development of a Phantom for Quality Assurance of Deviceless 4DCT
Abstract
Purpose
Deviceless 4DCT (D4D) reconstruction relies on changes to patient anatomy, such as lung volume or external contour to generate a pseudo respiratory trace which is used to bin 4DCTs for use in radiotherapy planning. This poses a challenge due to the lack of anthropomorphic QA phantoms that undergo physiological motions. Here we present a solution in the form of an in-house constructed phantom.
Methods
A truncated cone was designed in SolidWorks and 3D printed on a Prusa XL 2T with PLA filament. Once in motion, the phantom’s shape provided a longitudinally changing contour needed for D4D. The phantom’s centre accommodated the insertion of 3 Catphan CTP504 Modules which allows standard CT tests to be performed as part of our 4DCT QA. The phantom assembly was placed on an in-house motion platform programmed to move 1 cm in the cranio-caudal direction with a 3.5s period, meant to mimic respiratory motion. An RPM marker block was placed on the stage of the motion platform to capture the motion with an external surrogate. A clinical thorax 4DCT was captured and reconstructed to generate 10 respiratory bins and a static CT was captured for baseline.
Results
The in-house phantom provided the necessary geometry for D4D to succeed. Catphan analysis showed that all bins, maximum intensity projection (MIP) and average CT were equivalent using D4DCT vs Respiratory Gating for Scanners (RGSC)-based 4DCT. This included HU accuracy, low and high contrast spatial resolution, and position integrity of moving targets. Visually, D4D exhibited more motion blur and streak artifacts in the phase bins, the clinical impact of which is difficult to quantify.
Conclusion
A simple yet elegant solution is proposed to complement our QA program that facilitates the use of already established tools for use of D4D QA.