Ultrasound Microvessel Imaging with AI to Improve Cancer Detection

Ultrasound is widely recognized for its role in monitoring fetal development during pregnancy. However, it's also a valuable tool for inspecting abnormal tissue masses and nodules that could be cancerous. Traditional ultrasound machines often fall short when it comes to revealing the tiny blood vessels, or microvessels, that make up these tumors. To address this gap, researchers have created a software tool that enhances the resolution of ultrasound imaging. This high-resolution ultrasound imaging software is compatible with a range of ultrasound machines and promises to significantly enhance the clarity and detail of images.

The investigational software named quantitative high-definition microvessel imaging (q-HDMI) was developed by researchers at Mayo Clinic (Rochester, MN, USA). It has the capability to produce high-resolution 2D and 3D images of microvessels as small as 150 microns, about double the thickness of a human hair. The researchers also developed an algorithm that incorporates specific biomarkers related to the small vessels' characteristics, such as shape and pattern. This algorithm categorizes the imaged masses as either benign or malignant. In a clinical trial, the q-HDMI software, coupled with artificial intelligence (AI), successfully identified a malignant breast cancer mass measuring just 3 millimeters across in a 40-year-old woman. Early detection of such tiny cancerous masses can be crucial for successful treatment.

The researchers extended their study by applying this new technology to 521 patients with suspicious breast masses who had previously undergone traditional ultrasound imaging. The results were remarkable: the technology achieved nearly a 100% accuracy rate in distinguishing between benign and malignant tumors, irrespective of their size. The team then focused on thyroid nodules, testing 92 patients. Thyroid nodules are prevalent, and it's often difficult to discern between cancerous and non-cancerous ones using standard imaging techniques. The team identified 12 new biomarkers to differentiate between the two and integrated them into their AI algorithm, which achieved an 84% accuracy rate. This is a significant improvement over the 35-75% accuracy rate of traditional ultrasound techniques. The researchers believe their q-HDMI tool could be especially beneficial in regions with limited medical expertise and resources, such as rural and developing areas. They are also partnering with cancer specialists to adapt the q-HDMI tool for ongoing cancer treatment monitoring, enabling adjustments to individualized therapies in real time.

"If we can visualize and capture the microvessel in the earliest stages of cancer, we can better diagnose and treat it earlier, which improves the outcome for the patient," said physician-scientist Azra Alizad, M.D., who specializes in ultrasound technology for cancer imaging.

"This technology provides a quantitative value that shows the probability of malignancy," added biomedical engineering scientist Mostafa Fatemi, Ph.D. "It's a tool to extract information in a way that can be useful to clinicians."

Related Links:
Mayo Clinic