Alternative scheme of organizing the automated workstation of a radiologist

AIM: To compare the two aforementioned approaches in terms of their functionality and financial burden on the medical institution. MATERIALS AND METHODS: Data on AI-based software were obtained from participants of the Moscow Computer Vision Experiment, and information on vendor-specific solutions was partly collected from open sources, partly provided by vendors. To determine the relevance of the software functions, we conducted a survey among 40 center specialists. RESULTS: Today, AI-based software is able to analyze more than 20 areas of radiology, and 22 areas are under development; ready-made vendor - dependent solutions are presented for 18 areas. The cost of one ready-made solution, together with additional modules for the analysis of various modalities and anatomical regions, averages to 6 million rubles (based on 5 pathologies), while the main cost of AI-based software consists of the price of each study, actually evaluated by the software. With the processing cost per one study evaluated by AI-based software being at 50 rubles (estimated cost calculated based on the tariff agreement for 2023), and considering 24 CT/MRI studies per day and 22 working days per month, the cost of using AI-based software equals 316 800 rubles per year. CONCLUSION: Thus, the relevance of choosing one or another way of organizing the workspace of a radiologist currently depends on the patient flow of the institution and the actual use of decision support systems by the doctor. In general, AI-based software is more cost-effective due to reduced use of medical equipment and independence from specific company solutions, and has broader functionality than off-the-shelf vendor-specific solutions in the near future.