Sechenov.AI_nephro Medical Decision Support System (SPMS)

The main articles are:

• Medical Decision Support Systems (CDS)
• Kidneys

2025

Accreditation of Roszdravnadzor

An AI-based medical device developed at Sechenov University has been accredited by Roszdravnadzor. The university announced this on June 16, 2025.

A service to support medical decision-making and navigation of operations in patients with surgical kidney diseases Sechenov.AI_nephro developed at the Center for Neural Network Technologies of the Institute of Urology and Human Reproductive Health of Sechenov University. Thanks to the platform, the surgeon in a few minutes receives a single three-dimensional image of the zone of interest: a kidney tumor with the vessels that feed it, elements of the cup-pelvis system, veins, neighboring structures. Preoperative planning allows the doctor to assess the situation in detail and perform virtual resection of the kidney with different planes, the patient - to preserve the organ.

Kidney parenchyma cancer is one of the most common diseases faced by cancer doctors. About 20 years ago, a patient with this disease had no choice - he was threatened with kidneys. But gradually, kidney resection began to enter clinical practice - surgery to remove the tumor with the preservation of the organ.

To preserve the kidney and reduce the risk of postoperative complications, the surgeon before the operation must know the exact location of the tumor and what anatomical structures it borders, that is, to obtain a 3D image of the area of ​ ​ interest.

Since 2008, the Institute of Urology of Sechenov University began to use 3D modeling of the pathological process in patients with neoplasms of kidney parenchyma. In manual mode, when a team of three specialists works: a urologist, a radiation diagnostic doctor and an IT specialist, the construction process took up to a day. Today, thanks to the automation of the Sechenov.AI_nephro system, image acquisition is a matter of several minutes.

To do this, you need to perform only a few actions: create a patient's card, upload his CT study to the interface and highlight the area of ​ ​ interest - tumors in the kidney, - said Ivan Chernenky, developer of the web resource, leading software engineer at the Center for Neural Network Technologies at Sechenov University, - thanks to the 3D model, you can assess the depth of tumor germination into the kidney, perform virtual resection on different planes and obtain other information important for the doctor.

Thanks to the application of the preoperative planning program at the Institute of Urology, the structure of the operation has changed dramatically. If in the early 2000s kidney resection was no more than 20-30% of operations, then ten years later the number of such operations increased to 80-90%. Thus, thanks to the preoperative planning program, most patients are left with two functioning organs.

Sechenov.AI nephro is an AI-based project developed at Sechenov University. The web resource has been registered by Roszdravnadzor, which means that our product meets all safety and quality requirements. As of June 2025, the following options are available on the web platform: 3D modeling of the pathological process in patients with neoplasm of kidney parenchyma, the system can also build observation by a patient with kidney cysts and hydronephrosis, "explained Evgeny Sirota, head of the Center for Neural Network Technologies at the Institute of Urology and Human Reproductive Health, Sechenov University.

Passing of technical tests before registration

The Center for Neural Network Technologies of the Institute of Urology and Human Reproductive Health of Sechenov University has developed a system for supporting medical decision-making Sechenov.AI_nephro in the treatment of patients with neoplasms of kidney parenchyma. This was announced on February 27, 2025 by representatives of Sechenov University.

𝓲

Фото: Freepik

According to the company, the web platform automatically combines four phases of the patient's CT examination into a single three-dimensional image in just a few minutes. Work on the creation of the product was carried out in 2022.

A kidney tumor is one of the most common diseases faced by cancer doctors. About 20 years ago, a patient with this disease had no choice - only surgery to remove the kidney. But gradually, kidney resection began to enter clinical practice - surgery to remove the tumor with the preservation of the organ.

To plan for such surgery to reduce the risk of postoperative complications, the surgeon must know the exact location of the tumor and what anatomical structures it borders. It became clear that a 3D image is required, which will not only have a kidney tumor, but also the vessels that feed it, elements of the cup-pelvis system, veins and neighboring structures.

Since 2008, the Institute of Urology of Sechenov University began to use 3D modeling the pathological process in patients with neoplasms of kidney parenchyma. In manual mode, when a team of three specialists works: a urologist, a doctor radiation diagnostics and IT a specialist, the construction process took up to a day.

For February 2025, thanks to the automation of the system, image acquisition is a matter of several minutes. On the Sechenov.AI_nephro web resource, you need to create a patient card, download his CT examination in the interface and highlight the area of ​ ​ interest - tumors in the kidney area.

Before going to the server, this zone is cut out and sent for processing and already on the server side, the neural network distinguishes the necessary anatomical structures: arteries, veins, ureters, kidney parenchyma, tumor, cysts and creates an additional file with these "masks." After that, it is returned to the user and 3D construction is already carried out, with which you can work. Thanks to the 3D model, it is possible to assess the depth of the tumor immersion in the kidney, perform virtual resection on different planes and obtain other information important to the doctor. told Ivan Chernenky, leading software engineer at the Center for Neural Network Technologies, Sechenov University

Thanks to the application of the preoperative planning program at the Institute of Urology and Human Reproductive Health, the structure of the operation has changed dramatically. If in the early 2000s kidney resection was no more than 20-30% of operations, then ten years later the number of such operations increased to 80-90%. Thus, thanks to the preoperative planning program, most patients are left with two functioning organs.

In the near future, the medical decision support system will Sechenov.AI_nephro be registered. For February 2025, the following options are available on the web platform: 3D modeling of the pathological process in patients with kidney parenchyma neoplasm, and the system can also build observation by a patient with kidney cysts and hydronephrosis.

The plans include training the web platform for digital biopsy, that is, it will be able to determine whether a patient has a benign or malignant tumor, and a digital puncture to recognize the type of tumor. By the end of 2025, we hope to implement this in the form of an alpha version of the program. told representatives of the university

Also, the web platform can become useful for transplant doctors who are engaged in kidney transplantation. An agreement on cooperation in this area has already been signed with the Brest Clinical Hospital (Belarus).

Colleagues from Brest have accumulated significant experience in performing kidney transplantation. The proportion of disciplines that we plan to add to the platform is very important for transplant doctors, both at the planning stage in terms of functional state and at the postoperative stage to assess the functional state of the transplant. explained by Evgeny Sirota

One of the directions of development of the innovative scientific school "Consortium of neural network systems of 3D modeling for preoperative planning" is the use of AI in the diagnosis of bladder cancer. After endoscopic reduction of the bladder tumor, patients need to be monitored every three months and submit the material for biopsy. But in the case of bladder cancer, the situation is complicated by the fact that it is very difficult to recognize, since bladder cancer is a disease of the bladder mucosa. And experts of the Center neural network technologies have hypothesized to segment the video sequence of endoscopic observations for subsequent training of AI to determine this type of cancer with high accuracy. Segmentation of the video sequence has already begun as of February 2025.

INSh "Consortium of Neural Network 3D Modeling Systems for Preoperative Planning" is implemented within the framework of the program strategic academic leadership "Priority-2030" and the strategic project "Network for the Development of Best Practices in Medicine, Science and Education."