Computer software, to run partly as a Cloud computing service and partly on local devices, using new, patented forms of machine learning and artificial intelligence to achieve personalised organ-scale medicine in a way that has been previously impossible.
In other words, we're using advanced mathematics to achieve personalised medicine for complex diseases, partly by data-mining medical histories in a completely new way. This is achieved without waiting for genetics-based therapies and in a way that cannot be replaced by genetics-based therapies.
Already validated in silico (i.e. using computer simulations) and by data-mining actual medical histories, this software is most relevant for enhancing medication in chronic or degenerative diseases: our immediate application is enhanced insulin dosing for type-1 diabetes. The objective is to improve both quality of life and therapeutic outcomes for patients, particularly those suffering from highly unstable forms of type-1 diabetes.
Our flagship application is therefore machine-intelligent artificial pancreas software, designed to interrogate medical data to construct personalised diabetes models for people with unstable type-1 diabetes, and then generate suggested insulin infusion control laws to improve control of blood glucose, steering it to desired levels and keeping it there. Other forms of artificial pancreas algorithms do not cope well with unstable glucose-insulin dynamics, so this is an under-served cohort desperately in need of better technology.
In 2011 Dr Jenny Gunton, then Head of the Diabetes and Transcription Factors Group at the Garvan Institute of Medical Research and Dr Nigel Greenwood, of NeuroTech Research Pty Ltd, were awarded an Innovative grant by the JDRF in New York, to demonstrate the prototype Neuromathix Artificial Pancreas. With additional funding provided by the Queensland Government and by the directors of NeuroTech Research Pty Ltd, the team used a mix of simulated and actual human diabetic data to demonstrate the software, with spectacular results (see News).
One of the purposes of this project was to demonstrate that existing medical hardware used for diabetes therapies (insulin pumps and continuous glucose monitors) can have its performance transformed simply by using sophisticated mathematics and high-performance computing.
A dedicated start-up company, Diabetes Neuromathix Pty Ltd (DNx), was then formed to complete the clinical study and trialling of this software and engage in worldwide commercialisation of this intellectual property, to improve therapies for type-1 and the insulin-using subsets of type-2 and type-3c diabetes.
DNx has demonstrated successful generation of personalised predictive models of diabetes for people with highly unstable type-1, by data-mining medical histories and (literally) evolving personalised computational models. Consequently, DNx was a semi-finalist in the 2018 Diabetes Innovation Challenge in Boston, USA. Furthermore, its associated team, Team MachineGenes, was listed in December 2018 as one of the top-10 teams in the world in the IBM Watson AI XPRIZE. The AI XPRIZE imposed a rigorous auditing process across 2019, including having a "Red Judge" embedded in the team over five months, doing rigorous due-diligence on our technology.
Also in 2019, DNx in Team MachineGenes was one of the AI teams whose work was showcased in the United Nation's 'AI for Good' Global Summit in Geneva.
In January 2020 Team MachineGenes was announced as one of the ten semi-finalist teams worldwide in the AI XPRIZE. We were the only team from outside North America and Europe/Israel.
Given that the US FDA has a dedicated track for artificial pancreas candidates (due to their urgent need) and the extremely stable, safe design of the Neuromathix Artificial Pancreas, it is anticipated that the first of our tools ('KnowsMe') will reach the market by 2023, with versions of the Neuromathix Artificial Pancreas software reaching the market in late 2024 ('ShowsMe') and 2026 ('StabilizesMe'), provided sufficient funding is achieved to enable a suitable clinical study and clinical trials.