Fujifilm SonoSite is collaborating with the Allen Institute of Artificial Intelligence (AI2) Incubator to use artificial intelligence to improve the accuracy of ultrasound images, the standard way to detect ovarian cancer.
A developer of bedside and point-of-care ultrasound, Fujifilm SonoSite paired with the AI2 Incubator to use deep learning models on portable ultrasound products. Deep learning is a subset of machine learning, where artificial neural networks, algorithms created from the human brain, learn from large amounts of data.
The two companies will work together to enhance image analysis to broaden the use of ultrasound in a wider range of scenarios.
“The AI2 Incubator was a perfect place to look for help in creating breakthrough technology,” Rich Fabian, Fujifilm SonoSite’s president and CEO, said in a press release. “They have the type of talent that is hard to recruit, combined with the ambition of a startup. We look forward to collaborating more.”
Added Diku Mandavia, MD, the company’s senior vice president and chief medical officer: “The combination of deep learning and medical imaging is very exciting for the future of detection —better care and catching anomalies earlier and faster is a core mission.”
Deep learning-based techniques in medical imaging have resulted in advances such as finding tuberculosis with X-rays and detecting metastatic breast cancer in pathology slides.
Ultrasounds are less expensive and more portable than X-rays, computerized tomography (CT) scans, and positron emission tomography (PET) imaging, and do not expose patients to ionizing radiation. While ultrasound quality — historically its relative disadvantage — has improved over the last 20 years, deep-learning algorithms may significantly increase the technology’s accuracy and rapid assessment capability.
Ovarian cancer is one of the chief causes of cancer-related death in women globally, largely because the disease has no specific signs and symptoms and is often caught late, when the chances of curing it are low. Abnormalities are usually detected by ultrasound.
“In tackling this challenge, we are pushing deep learning, computer vision, and medical imaging into uncharted territory,” said Vu Ha, technical director at AI2. “In building new AI-based capabilities in affordable ultrasound devices, we hope to bring them to underserved markets to improve healthcare around the world.”
The AI2 Incubator was launched by the Allen Institute for Artificial Intelligence, the world’s largest nonprofit AI organization.