Scientists at Sechenov University have developed a neural network to detect lymphovascular invasion — the penetration of cancer cells into blood vessels — in lung tissue samples in adenocarcinoma.
The development is designed to more accurately and quickly determine the risks of metastasis, which will allow timely adjustment of the patient's therapy regimen.
Lymphovascular invasion is an important prognostic factor for the course of the disease, as cancer can spread through blood or lymphatic vessels, forming metastases. This requires the appointment of adjuvant therapy, but accurate detection of invasion is difficult due to possible differences in the assessment of pathologists.
If a tumor grows into blood vessels, it can enter the bloodstream, which leads to the formation of metastases. Therefore, it is important to detect the invasion before the tumor cells enter the bloodstream.
The new system will help to more accurately predict the development of the disease and adjust treatment in time.
For the development of the system, 162 histoscans were used, containing 8212 marked vessels, which were marked by pathologists as invasion sites. The neural network was trained on this data and showed an accuracy of more than 95% in identifying both blood vessels and tumor invasion into them.
It is expected that this automated system will help pathologists obtain more objective and accurate results, minimizing the risk of errors in diagnosis.
Read materials on the topic:
A nanodevice for the diagnosis of cancer and Alzheimer's has been developed in Moscow
AI on guard of health: a project to analyze tomography for cancer diagnosis has been launched at NSU