News

The Head of the Laboratory Presented at the Conference "Diagnosis and Treatment of Oncological and Hematological Diseases in HIV Infection"

Alexander Biryukov Presented at the Conference "Diagnosis and Treatment of Oncological and Hematological Diseases in HIV Infection"
The Head of the Laboratory Presented at the Conference "Diagnosis and Treatment of Oncological and Hematological Diseases in HIV Infection"

At the conference "Diagnosis and Treatment of Oncological and Hematological Diseases in HIV Infection," the Head of the Clinical Decision Support Laboratory based on AI Technologies at Sechenov University, Alexander Alekseevich Biryukov, spoke in the SYMPOSIUM section "ARTIFICIAL INTELLIGENCE — PHYSICS FOR MEDICINE" with a report titled "CDSS for Pathologist in Colorectal Cancer."

Alexander Alekseevich presented the results of an experiment on the use of the Clinical Decision Support System (CDSS) by pathologists on the basis of the Institute of Clinical Morphology and Digital Pathology at Sechenov University. He emphasized that the experiment revealed that AI positively affects the sensitivity of doctors in identifying both metastases and tumor deposits in colorectal cancer. Moreover, the AI model demonstrated sensitivity close to the results of experienced pathologists.

Special attention was given to the specificity of identifying healthy lymph nodes. The AI model showed the highest specificity in this task, which is a crucial factor for accurate diagnosis and subsequent treatment.

Regarding the time spent processing each sample, it was found that experiment participants work significantly faster with analogous specimens. Comparing the work with specimen scans (WSIs) before and after AI processing revealed that using AI significantly accelerates the work of the pathologist with WSIs.

One interesting observation presented in the report was that young specialists, using AI, achieved greater sensitivity in identifying metastases in colorectal cancer than experienced pathologists. This is explained by the fact that young specialists trust AI more, which tends toward overdiagnosis. At the same time, experienced pathologists have greater sensitivity in identifying tumor deposits, which are sometimes easily confused with metastases, a mistake more common among young specialists.

Alexander Biryukov's report demonstrated the significant potential of using AI in pathology, especially in the diagnosis of colorectal cancer. We continue to work on improving our technologies and their implementation in practice to enhance the quality of medical services.