Deep Learning-Based Cancer Detection Using Machine Learning

This area under reference is concerned with such a part of artificial intelligence that receives artificial neural

networks as machine learning techniques to recognize patterns or predict from very large data sets. The enthusiastic adoption of

deep learning in the healthcare sector, and accordingly, the availability of deeply well-characterized datasets for the cancer

prognosis have made the field much more electrifying in applying deep learning tools to unravel cancer complex biology. Early

findings are promising, but it remains an ever-changing world where information on cancer biology and deep learning keeps

pouring in.

This research offers an insight into the latest techniques in deep learning and their applications in oncology. Focus will mainly

be placed on the applications of deep learning towards omics data types-on genomic, methylation, transcriptomic data,

histopathology-inductive genomic inference, and provide commentaries on perspectives toward integrating these data types into

decision support tools. Finally, examples for applying deep learning to cancer diagnosis, treatment, and management will be

illustrated. In addition, we will highlight challenges and limitations to applying deep learning in precision oncology: the scarce

availability of data, especially of phenotype-rich type, and the need for more interpretable deep learning models. At last, scope

shall give a view of how these challenges can be fought with and sorted out in prospect for the future clinical application of deep

learning.