Dr Helen MulvanaJAMAAN MOHAMMED ALASMI2022-05-302022-05-30https://drepo.sdl.edu.sa/handle/20.500.14154/52409Cancer is an increasingly growing disease worldwide. It is considered to be a fatal and a heterogeneous disease having multiple subtypes. Early diagnosis and prognosis of cancer can help increase the survival rate of patients by medically treating them at early stages. Medical imaging modalities (Ultrasound, X-Ray, magnetic resonance imaging, etc.) play a significant role in the in-depth analysis of cancerous cells and tissues. However, the advent of medical imaging techniques gives rise to the need for observing the extracted images accurately and effectively. This responsibility lies completely on the expertise of medical practitioners which may lead to false-positive diagnosis sometimes. Also, inter- and intra-observer variability is an issue that needed to be addressed. Hence, Artificial Intelligence (AI) has gained the attention of researchers for the analysis, detection and prognosis of cancer. Various Machine Learning (ML) and Deep Learning (DL) techniques have been applied in recent literature, and the research field is still gaining popularity. Introduction of AI (ML and DL) makes learning and prediction of cancer accurate and efficient. Quantitative ultrasound (QUS) is a highly advanced type of ultrasound medical imaging. QUS can facilitate in cancer detection and bringing down the unnecessary biopsies and other non-invasive procedures used to detect types of tumors and lesions. AI backed up with QUS technologies, and Machine Learning (ML) algorithms have gained immense attention in the past decade owing to their ability to improve medical imaging and automatically interpreting them. Therefore, this research is a review of the role QUS technology plays individually and when combined with AI algorithms for cancer prognosis and detection. A review in terms of predictive performance variations in ML/ DL models according to the input features has been presented along with the identification of strengths and limitations of commonly used algorithms in cancer prediction.en