Effectiveness of multi-agent technology in three-day campaign for early diagnosis of cutaneous melanoma

BACKGROUND: Melanoma — is a tumor that in most cases affects the skin and is characterized by an extremely aggressive course and a steadily increasing morbidity in the world. However, diagnosed at an early stage, cutaneous melanoma (CM) has a good prognosis with correct treatment. The results of the diagnosis of CM can be improved by joint the efforts of dermatologists and artificial intelligence (AI). AIM: Cutaneous melanoma is a tumor with an unpredictable course. The article discusses ways to solve the problem of early diagnosis using multi-agent technology and an expert system based on AI. MATERIALS AND METHODS: In a large industrial city with more than three million population, a three-day campaign for the early diagnostics of cutaneous melanoma was carried out, which revealed 4 cases of CM (4.35%) at pT1a stage in 96 patients registered in 2019. A total of 800 people were examined. RESULTS: As a result of diagnostics, the following data were obtained: specificity of self-diagnostics of the region was 6.78% by the inhabitants, 78.89% by dermatologists, and 95.24% by expert oncologists. In prospective quality control of diagnostics within 6 months, such parameters as the sensitivity of diagnosing cutaneous melanoma by oncologists and dermatologists were both 100%. As a result of the study, it was found that the multi-agent technology is necessary to improve the results of CM diagnostics, and also for a more complete assessment of the onco-epidemiological situation, and for forecasting of the necessary resources in the region. CONCLUSIONS: The multi-agent technology can improve diagnostic results, but for a more complete assessment of the onco-epidemiological situation, a large-scale population screening in the region is required.