A breast scan for detecting cancer takes less than a minute using an experimental system that combines photoacoustic and ultrasound imaging, according to a study in IEEE Transactions on Medical Imaging.
The system does not require painful compression like mammography. Instead, patients stand and gently press their breast against an imaging window.
In tests involving four healthy individuals and 61 breast cancer patients, it produced clear, artificial intelligence-powered 3D images of common breast cancer subtypes such as Luminal A, Luminal B and Triple-Negative Breast Cancer.
Our system, which is called OneTouch-PAT, combines advanced imaging, automation and artificial intelligence –all while enhancing patient comfort,” says the study’s corresponding author Jun Xia, PhD, professor in the University at Buffalo’s Department of Biomedical Engineering.
Mammography is widely available but is less accurate among women with dense breast tissue, involves radiation and is painful. Ultrasound, which is often used in conjunction with mammography, is better with dense breast tissue, but it can produce false positives and its quality is reliant upon the skill of the sonographer.
Other tools such as MRI are effective but expensive, time-consuming and not widely available.
Xia and colleagues have been studying photoacoustic imaging, which works by emitting laser pulses that cause light-absorbing molecules to heat up and expand. This in turn creates ultrasound waves that allow medical professionals to detect blood vessels that often grow more in cancerous tissues.
Typically, these systems require a sonographer to manually scan the breast, or they rely on separate devices for photoacoustic imaging and ultrasound imaging.
OneTouch-PAT combines both scans automatically – in other words, there is no potential for operator error – with the patient in the same standing position. The device performs a photoacoustic scan first, followed by an ultrasound scan, then repeats this pattern in an interleaved way until the entire breast is covered.
The system then processes the data using a deep learning network to improve image clarity. Depending on the computing power in this step, this may take only a few minutes. Ultimately, the research team found that OneTouch-PAT provides a more in-depth and clearer view of breast tumors compared to photoacoustic and ultrasound imaging systems that are operator-dependent.
image: University at Buffalo