BSC initiates clinical validation of radiation-free breast cancer diagnostic method at Vall d'Hebron Hospital

QUSTom Project: Advancing Breast Cancer Diagnosis with 3D Ultrasound and Supercomputing, Led by Josep de la Puente from Barcelona Supercomputing Center

The QUSTom project, backed by the European Union and managed under the guidance of the Barcelona Supercomputing Center (BSC-CNS), is breaking new ground by exploring a novel approach to medical imaging. Utilizing 3D tomographic ultrasound and advanced supercomputing, the project aims to enhance early breast cancer detection. In collaboration with the Hospital Universitari Vall d'Hebron, the project is now entering the clinical validation stage. This cutting-edge technique holds the potential to transform breast cancer diagnostics, offering a non-invasive method that provides a more comprehensive and functional view of breast tissue compared to traditional imaging modalities like X-rays.

Over the coming weeks, women participating in the hospital’s early breast cancer screening program will be invited to join the clinical trials for this pioneering technology. Unlike conventional ultrasound, which provides real-time images, the focus here is on high-resolution diagnostics to improve accuracy. The aim is to complement, and possibly replace, X-ray-based mammography with this more sophisticated technique, which promises not only better quality imaging but also heightened patient comfort.

The project uses a revolutionary 3D Ultrasound Computed Tomography Scanner (3D USCT III), designed and built by the Karlsruher Institut für Technologie (KIT) in Germany—a key partner in QUSTom. This scanner is one-of-a-kind globally, equipped with 2304 transducers arranged in a hemispherical formation. These transducers function as both transmitters and receivers to detect pathological changes in breast tissue. KIT has developed further prototypes, but the device currently being tested on patients in Barcelona is the first undergoing validation using full-wave inversion algorithms.

Before it reached this stage, the 3D USCT device was extensively tested for electrical safety and ultrasound performance in a certified lab in Germany, passing all tests successfully. The raw data collected during the imaging sessions will be processed using advanced 3D full-wave inversion algorithms to generate detailed images. Much of this heavy computational lifting will be done on BSC's MareNostrum 5 supercomputer, employing UBIware software by FrontWave Imaging, a spin-off of BSC and Imperial College London. This innovative approach integrates multimodal imaging and delivers precise real 3D images, establishing a new frontier in breast cancer screening through ultrasound technology.

Generating these high-resolution images in 3D is no small feat—it requires a staggering 50,000 ultrasound wave simulations per image. While 2D simulations can be performed on standard Graphics Processing Units (GPUs) in commercial cloud environments, handling the data complexity in 3D is a mammoth task. “We are charting new territory by running these 3D image reconstructions on MareNostrum 5,” says Josep de la Puente, principal investigator of QUSTom at BSC. “What would take a conventional computer years to process, we can achieve in just a few days with the power of BSC’s infrastructure.”

At the heart of the project lies the creation of a digital twin—an exact digital replica of the breast tissue and the ultrasound device. This digital model mirrors the physical device used by radiologists, simulating its ultrasound emissions. “This allows us to produce not just an image, but a full 3D map displaying the tissue’s properties down to the pixel level,” de la Puente explains.

“This new diagnostic tool will provide a more comprehensive and functional imaging approach and eliminate the need for harmful ionizing radiation while also improving women’s comfort during routine breast exams,” says Ana María Rodríguez Arana, Head of the Women’s Radiology Service at Vall d'Hebron Hospital. She also holds a principal investigator position in the Molecular Medical Imaging group at the Vall d’Hebron Research Institute (VHIR).

A key differentiator of QUSTom’s technology is its non-reliance on radiation, unlike mammography which uses X-rays. Additionally, this new ultrasound-based device offers superior potential for accurate imaging and tumor monitoring, an aspect particularly beneficial for women with denser breast tissue—an issue affecting around 40% of the female population globally according to the Spanish Society of Senology and Breast Pathology (SESPM).

BSC’s expertise in solving mechanical wave problems has been instrumental in developing the algorithms required for this specialized medical imaging. These algorithms have their roots in areas of research that are worlds apart, such as geophysical applications studying the Earth’s subsurface.

Alongside BSC, Vall d’Hebron Research Institute (VHIR), and FrontWave Imaging, the project partners include Karlsruher Institut für Technologie, Arctur, and Imperial College London, contributing additional expertise.

The Spanish Society of Medical Oncology (SEOM) reports that breast cancer is one of the most diagnosed cancers worldwide, with around 2.3 million cases in 2020, resulting in approximately 700,000 deaths that year. In Spain, projections for 2024 indicate around 36,395 new cases, a slight uptick from previous years, according to REDECAN.

Early detection remains vital for improving survival rates. While mammography remains a critical tool in the fight against breast cancer, the need for alternative, non-invasive methods that forgo ionizing radiation is becoming ever more apparent.

Launched in 2022, QUSTom was selected as part of the first call of the Pathfinder Open program funded by the European Innovation Council. Under the Horizon Europe Framework, the initiative has received 2,744,300 euros to support its European operations. Out of 868 projects evaluated by the European Commission, QUSTom was among the 56 chosen, with 11 coming from Spain.

3459 likes 123 282 views
No comments
To leave a comment, you must .
reload, if the code cannot be seen