Quaisr technology is enabling engineers at Multiwave Metacrystal to automate complex physics simulations and integrate them with statistical and machine-learning algorithms.
Can we design cancer-detection technologies faster? At Quaisr, we've teamed up with partners at Multiwave Metacrystal, and thanks to €200k in funding from HUBCAP we will combine forces to deliver a 10x faster innovation cycle in the fight against cancer. The partnership will drastically lower the number of physics simulations required in the design of new cancer-detection technologies – reducing radiation exposure while increasing detector sensitivity.
Cancer is one of the leading causes of death worldwide: the World Health Organization estimates that cancer led to the death of nearly 10 million people in 2020. Devastatingly, the number is expected to rise to 16.4 million annually by 2040. Detecting cancer earlier and more accurately will reduce the number of people losing their lives.
Positron emission tomography (PET) scanners are crucial for early cancer diagnosis, which is the most important factor for increasing cancer survival rates. PET scanners work by detecting annihilation gamma rays, which result from the decay of radionuclides – a radioactive substance – injected into the patient. The emitted gamma rays are detected using scintillator-based systems that comprise the scanning device. Scan information is then analysed by computer codes, which reconstruct an image of the metabolic activity within the organ or tissue being examined. The power to visualise abnormal cell function makes PET a leading technology for the detection of not only cancer, but also serious neurodegenerative diseases such as Alzeheimers.
Particle-level simulations are critical for developing new PET-detector technologies – complex simulations that predict gamma ray interactions, photon production, transport and detection. Unfortunately, there are an infinite number of possibilities in which these processes could happen. A very slow and extremely computationally expensive process, Monte Carlo methods are proving to be the major bottleneck in the PET-detector innovation cycle.
Monte Carlo methods rely on repeated random sampling to obtain statistically significant solutions. To date, Monte Carlo simulations have been the best option when innovating with metamaterial technologies, allowing the Multiwave Metacrystal team to determine design characteristics including scintillator efficiency, together with energy, spatial, and timing resolutions.
Quaisr technology is enabling engineers at Multiwave Metacrystal to automate complex physics simulations and integrate them with statistical and machine-learning algorithms. Going forward, uncertainty-quantification (UQ) techniques will be used to reduce the number of simulations required, while supporting full determination of scintillator characteristics under alternative material and optical property assumptions. Next, a machine-learning surrogate-based optimiser (Bayesian based) with embedded UQ will search for optimal scanning time, resulting in higher image quality while using less radioactive tracer. By automating simulations and coupling them to state-of-the-art machine-learning techniques, we will be delivering a 10x speedup in design prototyping, faster time to market, and lower computational costs.
We have started on a journey to deliver faster innovation in the fight against cancer.
Multiwave Metacrystal exploits the emerging technology of metamaterials, which have transformed the field of wave physics. Metamaterials were pioneered by Sir John Pendry at Imperial College London at the turn of the 21st century. Resonance phenomena can be used to design materials with advanced properties, typically beyond those of any naturally occurring material. The potential of such materials to impact upon 5G networking, antenna theory, acoustic soundproofing and vibration control is now leading to novel devices, while the application to PET offers an opportunity to revolutionise cancer diagnosis. Multiwave Metacrystal’s technology is being used to design scintillator detectors at the heart of PET scanners, offering physicians potential gains of
HUBCAP is a €3M fund for SMEs to experiment and innovate with model-based designs for cyber-physical systems. More information can be found at hubcap.eu.
Resources
Explore more insights and resources related to this topic. Discover how AI and automation are transforming industries across the board.
Schedule a technical deep-dive to discuss your specific simulation stack, HPC security requirements, and data protocols. See exactly how Quaisr fits into your existing architecture.
Schedule a technical deep-dive to discuss your specific simulation stack, HPC security requirements, and data protocols. See exactly how Quaisr fits into your existing architecture.
