In precision oncology, the combination of the strengths of both histopathology and medical imaging provides a fertile ground for tumor characterization. Although histopathology offers a definitive cellular diagnosis, this approach is invasive and only provides a small-scale characterization of the tumor, while medical imaging modalities, such as X-ray, CT, MRI, ultrasound, and PET scans, provide a complete characterization of the tumor but, until recently, relied on the subjective ability of a human observer. The application of machine learning to radiomics aims at filling this gap, as images are mined to reveal patterns of disease not visible to the naked eye. In this perspective paper, the trajectory of machine learning in radiomics for oncology applications is critically discussed. By exploring studies using different imaging modalities, we seek to look beyond the achievements of innovative algorithms and identify the systemic weaknesses in the field, which are holding it back from translating to the clinic. In this regard, we identify two major challenges in the field: the significant effects of inter-modality and inter-scanner variability in model generalizability, and the ‘interpretability gaps’ in understanding the rationale for the decision-making process in ML algorithms. In this paper, we assert that these challenges are holding back even the best of algorithms and thus set the direction for the field in the future, advocating for the development of ML systems with emphasis on their performance in real-world settings as opposed to the lab.

The accurate detection and classification of brain tumors from magnetic resonance imaging (MRI) are critical for diagnosis and treatment planning. While deep learning has shown remarkable success in this domain, many state-of-the-art models rely on complex, end-to-end convolutional neural networks (CNNs) that require extensive computational resources and large, annotated datasets for training. This work proposes a novel and efficient methodology that, for the first time, leverages self-supervised DINO vision transformer backbones (DINO v1, DINOv2, and DINOv3) on a large corpus of natural images as powerful feature extractors for brain tumor analysis. We utilize the rich, general-purpose features from DINO-family backbones without fine-tuning the core model. These extracted features are then fed into a simpler, task-specific classifier (such as a support vector machine or a multi-layer perceptron) for the final detection and multi-class classification (e.g., glioma, meningioma, and pituitary tumor). Our methodology is evaluated on two benchmark medical imaging datasets with various classifying granularities. The results demonstrate that the proposed method achieves competitive and, in some cases, superior classification accuracy compared to representative fine-tuned convolutional neural networks and attention-based architectures, while significantly reducing the number of trainable parameters and training time. In particular, the best configuration achieves up to 98.17% accuracy and an F1-score of 98.18% on the 15-class dataset and 99.08% accuracy and an F1-score of 99.02% on the 4-class dataset. This study confirms the exceptional transfer learning capabilities of self-supervised vision transformers like DINO in the medical imaging domain, establishing it as a highly effective and efficient backbone for robust brain tumor detection and classification systems.