Find in Library

Oncolytic viral therapy is a promising approach to cancer treatment. The first oncolytic virus in the world was the genetically unmodified ECHO-7 strain enterovirus RIGVIR, which was approved in Latvia in 2004 for the treatment of skin melanoma. Our goal is to understand – how could Machine Learning help in the RIGVIR treatment administration. Machine learning for skin disease may well be transferred to practice. Digital image analysis and melanoma genomics studies are promising approaches, but they are currently in their infancy. Machine learning saves time for doctors, but unfortunately does not increase the amount of knowledge about melanoma, since the computer-generated features are difficult to interpret. The analysis shows that the possibilities of classical Linear Discriminant Analysis (LDA) have been successful for cancer diagnostics, but til now it has not been applied to RIGVIR studies. The promising future work is to develop decisive rules in the form of LDA for diagnosing the stages of melanoma during treatment with RIGVIR on the basis of measurements of CD4+, CD8+, and CD38+ lymphocytes and/or their ratios.