Applications of AI in Cancer Diagnosis and Categorization

Abstract

Cancer is a highly individualized disease, as tumors both take on traits from their host and develop their own mutations. Thus, it is critical to better understand the characteristics and behavior of tumors to treat patients more effectively. There are various artificial intelligences (AI) that can and are being used to improve the development of oncological technologies. AI is designed to replicate human intelligence, but not all models are capable of the same tasks. In diagnostics, AI models can be used to analyze multimodal data, to improve the accuracy of cancer diagnostics. AI can also identify cancer stage and grade. A tumors stage depends on tumor size, while its grade depends on how much the tumor has spread. By subtyping tumors, researchers are able to develop more targeted drugs, as similar tumors most likely arose through similar pathways. However, data from studies is unverified and different, which makes it hard to find formatted data to train AI models. Although the application of AI clinically is promising, AI still needs to go through more trials before its applied clinically.