Developing Lower Cost Radiotherapy Solutions for Low and Middle Income Countries
Date
2023-08-30
Authors
O'Connell, Jericho
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Radiotherapy research often focuses on state-of-the-art methods to provide small improvements in the treatment of patients in high-income countries, while less focus is put on providing low-cost treatments suitable for the majority of people in the world who have little or no access to radiotherapy. In an effort to remedy this paradigm we optimise, design, and benchmark lower cost radiotherapy treatment modalities with the overarching goal of increasing treatment accessibility in low- and middle-income countries (LMICs).
This body of work has focused on the simulation of cost reducing modification to linacs, in an effort to reduce linac production cost while maintaining treatment efficacy. One method is to simplify current linacs through the elimination of the kilovoltage on-board CT imaging system (kV-OBI). The removal of the kV-OBI would greatly reduce production costs of treatment linacs. However, the removal of the makes many image guided radiotherapy (IGRT) treatments infeasible, treatments which form the bulk of modern radiotherapy. To allow IGRT without the kV-OBI, novel megavoltage cone beam CT (MV-CBCT) methods are simulated to find low-cost setups that provide image quality similar to kV-OBIs. Additionally, cost reduction can be achieved through redesigning current radiotherapy machines by replacing expensive linear accelerator-based treatment heads with low-cost x-ray tubes. To validate this methodology, arc treatments were simulated on a simplified isocentric kilovoltage arc (SITKA) treatment machine with a novel treatment planning system. Additionally, machine learning solutions are used to ameliorate the systems OBI image quality such that a separate planning CT machine is not needed.
The Fastcat tool was developed to rapidly simulate CBCT through a combination of pre-calculated Monte Carlo (MC) data and GPU raytracing. To demonstrate this tool improvements, addressing the low contrast to noise ratio in MV-CBCT were studied and proposed. The rapid prototyping of CBCT setups available through Fastcat greatly improve the development of CBCT systems by providing a fast alternative to time-consuming MC simulations in key development situations: Allowing researchers to efficiently optimize a CBCT detector design based on quick feedback in terms of image quality in phantoms for a given dose. This was seen through utilization of Fastcat for the purpose of virtual clinical trials and detector design optimization of novel perovskite and cadmium tungstate (CWO) detectors. Design optimization results demonstrate the benefit of replacing existing cesium iodide (CsI) and gadolinium oxysulfide (GOS) detectors with next generation Perovskite direct conversion detectors. Perovskite detectors showed higher contrast to noise ratio (CNR) and spatial resolution, resulting in better image quality for clinical tasks such as patient positioning and micro-calcification detection, all at a very low manufacturing cost. Likewise, the novel, low-cost system in combination with adaptive machine learning methods and novel planning is demonstrated to provide clinically practical lung treatments that meet the urgent and increasing demand for radiotherapy treatment in low-income countries as well as rural and remote areas.
Description
Keywords
Radiotherapy, Fastcat, MV-CBCT, Hybrid MC simulation, MC Simulation, kV Treatment, SITKA