MEPhI: Intelligent Assistant Physician for Ultrasound Diagnostics of Thyroid Nodules

The main articles are:

• Neural networks (neural networks)
• Artificial Intelligence in Medicine
• Ultrasound (Ultrasound)
• Thyroid cancer
• Thyroid gland

2023: Creating a neural network that helps classify thyroid neoplasms based on ultrasound images

The development of neural networks, which, having trained on marked medical data, help doctors make diagnoses and even develop recommendations for treatment - one of the most advanced directions for using artificial intelligence technologies in medicine. A project of this kind was recently implemented by students of NRNU MEPhI, which announced this on July 6, 2023.

The project is called "Intellectual Assistant to the Doctor ultrasound diagnostics of Nodal Formations." thyroid gland As the name suggests, the intellectual system helps endocrinologists by classifying neoplasms in the human thyroid gland based on ultrasound images. "Assistant" classifies nodes formations based on TI-RADS (Thyroid Imaging Reporting and Data System) - an international standardized system for describing and processing data radiation studies of the thyroid gland

The system was developed by students of NRNU MEPhI with the participation of IIX professor Konstantin Zaitsev and VISH teacher Maxim Dunaev under the general guidance of IFIB director Alexander Garmash.

The partner of the project was the "National Medical Research Center of Endocrinology" of the Ministry of Health of Russia, whose specialists not only advised the developers, but, more importantly, provided marked-up medical data for training the neural network. In total, data from 137 patients were used, the diagnostic results of which were included in the training sample of 400 unique film loops (collections of several dozen images) and single images. Before the transfer to the developers, all data was anonymized.

From a technical point of view, the "Assistant" was based on the architectures of neural networks already in the public domain - the architecture of the Deeplabv3 + neural network was used to segment images and highlight nodes, and the neural network EfficientNetB6 was used to diagnose pathologies and classify neoplasms. However, these programs were taken only as a basis - the members of the student team had to work on the architecture of the final product.

In addition to the actual diagnostic functions, the "Assistant" performs others that are needed at the workplace of the doctor - this is the accounting of patients, and the maintenance of medical statistics, and the "expert mail" unit - a special interface that helps the doctor get a "second opinion" from another specialist.

It is also very important that the "Assistant" has a built-in feedback system that allows a doctor working with it to point out neural network errors, and thus increase its accuracy.

The system has been tested since the beginning of 2023 on the basis of the National Medical Research Center of Endocrinology; as of July 2023, three practitioners work with it. Work is also underway to register the "Assistant" as a medical device.

Meanwhile, as the project participant, student Ksenia Tsiguleva, said, the Assistant development team has big plans: in the future, the neural network should analyze data not only from ultrasound, but also from cytology, histological and genetic studies.

In addition, the development team plans to join another project being implemented at NRNU MEPhI - the development of a medical dermatological neural network.