Artificial intelligence (AI) integration in medical imaging has solved essential problems in oncology by enabling the early diagnosis and accurate segmentation of tumours. This chapter focuses on state-of-the-art AI-based tumour detection and segmentation techniques with special reference to ML and DL for modern diagnostic applications. New neural architectures have emerged and been substantiated, including convolutional neural networks (CNNs) and transformer-based models, allowing near-perfect accuracy in detecting tumors regardless of the imaging mode - MRI, CT or PET scans. This chapter focuses on the analysis techniques used to preprocess the data, train the model, and optimize detection to increase the sensitivity and specificity of the results. It also emphasizes the use of AI for visual analysis, generalization of tumor width, and general time saving for clinical practices. Recent AI frameworks, including U-Net and Mask R-CNN, are discussed from a systems perspective regarding their application in outlining the tumor borders with an accuracy of sub-millimeter. This chapter also addresses modality discrepancy in the data, interpretability in the case of AI, implementation of the AI in clinical settings, data augmentation, XAI and, lastly, legal compliance. Further developments of research on possible topics discussed in the light of the given work are possible on adding multimodal data fusion and real-time AI based decision support systems as topics. This chapter should be of interest to AI researchers, technologists and clinicians looking at the application of AI for diagnostic purposes and prognostic success of patients in cancer cases.