Cancer is one of the most important global health problem. The Global Burden of Disease estimated that 9.56 million people died prematurely as a result of cancer in 2017. Medical professionals agree that improvements in the application of prevention and lifestyle changes (diet, smoking, obesity) and early diagnostic detection can reduce substantially incidence and mortality from most types of cancer, while increasing better treatment with less cost and better quality of life for cancer patients. In the last decade, many diagnostic tests and methods have been improved with higher accuracy. Tests are subdivided into Imaging (Radiology) tests, endoscopy procedures and biopsy and cytology tests.  These diagnostic tests inevitably produce vast amounts of data which need to be examined by experts and to differentiate between benign and malignant tumours. Now Artificial Intelligence (AI) applications can be used to examine large numbers of diagnostic imaging with improved accuracy amplifying the efficiency of the health system. New technological advancements in AI algorithms and improvements in computer hardware can be used to train artificial neural networks to acquire diagnostic experience from large number of scans. Already some deep learning and machine learning models can teach computers to compare and analyze enormous amount of scanning data for cancer. The diagnostic skills of highly sophisticated software have been tested and compared with the traditional diagnostic tools by cancer experts. Their accuracy is much improved and are considered very helpful in early diagnosis as well as prognosis of the extend of various cancers. Scientists turned into artificial intelligence (AI) systems that are capable of surpassing human experts in breast cancer prediction and much earlier diagnosis. Similarly, computer scientists devised AI algorithm and deep learning models that could predict which people would go on to develop lung cancer by analysing low-dose CT (computerized tomography) scans of the lungs. Recently, application of convolutional neural network (CNNs) was used in the diagnosis of the invasion depth of gastric cancer based on conventional endoscopy. Also, studies demonstrated that AI approaches combined with imaging can have considerable impact on early diagnosis of oral cancer outcomes, with algorithm-guided detection of oral lesion heterogeneity. This review collected some very interesting research papers on the subject