Optimization of low-dose, low cost mobile 3D X-ray Imaging

Digital Tomosynthesis (DT) is a 3D mode of X-ray imaging. This imaging mode extracts both the typical
2D information you would expect from an x-ray image, as well as depth information on the body being
imaged. This is conventionally achieved by altering the angle of an x-ray tube over a set angular range,
where the image created at each unique angle corresponds to a certain layer of depth of the imaged body
[1]. These images can be reconstructed using reconstruction algorithms [2] to create a singular image
with enhanced contrast between the object of interest and background than is seen in 2D x-ray images.
However, conventional DT imaging machines for medical purposes require high power to be operated, and
are too large to be portable and hence have no application in bedside imaging.
A portable DT system created by Adaptix Ltd [3] has changed this. This system has been designed and
constructed which uses cold-cathode, flat panel x-ray source (FPS) arrays. These arrays are constructed of
25 individual emitters that emit sequentially, creating the multiple projections conventionally created by
the rotation of the x-ray tube [4]. This device has been successful for small animal and human orthopaedic
imaging. Hence, proof-of-concept for the scaling up of this for chest imaging purposes has been shown [5]
and active research is now underway for the development of this potential low-cost and portable method
of chest DT imaging. This technology sees additional application in non-destructive material evaluation
(NDE) and dentistry contexts.
This PhD project aims to utilise Geant4 [6], a Monte-Carlo simulation software, to address the engineering
and science questions that arise within the further development of these devices. Initial work has been
completed to investigate the effect of different x-ray target constructions on the x-ray spectra produced
and hence the dosage to the patient and the quality of the image outputted. Ongoing work endeavours to
develop this simulation framework further by implementing a realistic detector to allow closer comparison
of the images created from this work to those outputted from the physical device itself. This should
additionally inform a discussion on the optimal detector construction for this DT approach and identify
how best to parameterise detector quality in this context. Future work is planned to follow on from
this by applying this benchmarked simulation environment to look at modern x-ray approaches, such
as dual-energy imaging, to identify how best to adapt the device for these novel capabilities. This will
create simulation data that can then be experimentally verified with the device itself. Hence, the project
narrative will reflect the development of an accurate simulation environment within the Geant4 software
which can be used to address practical questions raised when constructing this technology in a more
cost- and labour-efficient method than can be achieved experimentally. Comparison with existing data
will ensure accurate benchmarking to prove the reliability of the simulation and hence allow it to inform
future developments in combination with experimental work.
References
[1] Shinn Huey Chou, Greg A. Kicska, Sudhakar N. Pipavath, and Gautham P. Reddy. Digital tomosynthesis
of the chest: Current and emerging applications. Radiographics, 34:359-372, 3 2014.
[2] Vadim Y Soloviev, Kate L Renforth, Conrad J Dirckx, and Stephen G Wells. Physics in medicine
biology physics in medicine biology meshless reconstruction technique for digital tomosynthesis. Phys.
Med. Biol, 65:85010, 2020.
[3] Adaptix Ltd, https://adaptix.com/.
[4] Thomas Primidis. Design and optimisation of ultra-compact, high-resolution 3d x-ray imaging systems.
2022.
1
[5] Thomas G Primidis, Stephen G Wells, Vadim Y Soloviev, and Carsten P Welsch. 3d chest tomosynthesis
using a stationary flatpanel source array and a stationary detector: a monte carlo proof of
concept. 2021