Malignant or benign? Quick and accurate diagnosis
Stiffness levels and distributions of various biological materials reflect disease-related information from cells to tissues. For example, malignant breast tumors are generally stiffer and have a more irregular shape than benign breast tumors. Ultrasound elastography can noninvasively determine the degree and shape of tissue stiffness and is used to diagnose breast cancer due to its low cost. However, experienced expert opinion is essential for interpreting ultrasound elastography images, but different experts differ in their accuracy.
Korea Institute of Science and Technology (KIST) President Dr. Seok-Jin Yoon announced that the team of Dr. Hyunjung Yi at the Spin Convergence Research Center and Suyoun Lee, the Director of the Spin Convergence Center neuromorphic engineering, had developed a simple yet highly accurate disease diagnostic technology by combining touch neural devices with artificial neural network learning methods. Unlike previously reported artificial tactile neural devices, this tactile neural device can determine the stiffness of objects. Neuromorphic technology is a field of research that aims to imitate the information processing method of the human brain, which is capable of high-level functions while consuming a small amount of power using electronic circuits. Neuromorphic technology is gaining attention as a new data processing technology suitable for AI, IoT and autonomous driving, requiring real-time processing of complex and vast information. Sensory neurons receive external stimuli through sensory receptors and convert them into spike electrical signals. Here, the generated spike pattern varies based on external stimulus information. For example, a higher stimulus intensity results in a higher generated peak frequency. The research team developed an artificial touch neuron device with a simple structure that combines a pressure sensor and an ovonic threshold switching device to produce such sensory neuron features. Applying pressure to the pressure sensor causes the resistance of the sensor to decrease and the peak frequency of the connected ovonic switching element to change. The artificial touch neuron device developed is a high-response, high-sensitivity device that allows pressure force to generate faster electrical spikes while improving pressure sensitivity, which focuses on making stiffer materials lead to faster pressure detection when pressed.
The electrical spike duration (or 1/frequency) generated by the developed device is less than 0.00001 s, which is more than 100,000 times faster than the few seconds it usually takes to press an object. In addition, while the existing devices could detect low pressure (about 20 kPa, similar to light pressure force) with a peak frequency change of 20-40 Hz, the developed device can detect low pressure with changes peak frequency of 1.2 MHz. . This allows real-time conversion of pressing force changes into spikes.
To deploy the developed device to the actual diagnosis of the disease, the research team used elastography images of malignant and benign breast tumors and used a neural network learning method. Each pixel in the color-coded ultrasound elastography image that correlates to the stiffness of the material being imaged was converted into a peak frequency change value and used for AI training. As a result, it was possible to determine the malignancy of a breast tumor with an accuracy of up to 95.8%.
The KIST research team stated that “the developed technology of artificial touch neurons is able to detect and learn mechanical properties with a simple structure and method.” The team added: “Through follow-up research, it will be possible to solve the problem of noise reflection, which is a disadvantage of ultrasound elastography if artificial touch neurons can collect the elastography image from an object obtainable using ultrasound elastography. The team also expects the device to be useful in low-power, high-accuracy disease diagnosis and in applications such as robotic surgery where a surgical site must be quickly determined in an environment that humans cannot. cannot contact directly.
KIST was established in 1966 as the first government-funded research institute in Korea to establish national development strategy based on science and technology and disseminate various industrial technologies to promote the development of major industries. KIST now elevates the status of Korean science and technology through the pursuit of world-class innovative research and development. For more information, please visit the KIST website at https://eng.kist.re.kr/
This research, with the support of the Ministry of Science and ICT (Minister Jong-Ho Lee), was carried out under the institutional project KIST, the individual research project in basic sciences and engineering of the National Research Foundation, the Next-Generation Intelligent Semiconductor Technology Development Project, the Future Semiconductor New Device Resource Technology Development Project, and the Nano•Material Technology Development Project. The research results are published as an article inside the back cover in the latest issue of Advanced materialsthe international journal in the field of materials.
The title of the article
An artificial touch neuron enabling peak representation of stiffness and disease diagnosis
Publication date of articles
Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of press releases posted on EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.