Recent advances in quantum computing have broadened the available techniques towards addressing existing computing problems. One area of interest is that of the emerging field of machine learning. The intersection of these fields, quantum machine learning, has the ability to perform high impact work such as that in the health industry. Use cases seen in previous research include that of the detection of illnesses in medical imaging through image classification. In this work, we explore the utilization of a hybrid quantum-classical approach for the classification of brain Magnetic Resonance Imaging (MRI) images for brain tumor detection utilizing public Kaggle datasets. More specifically, we aim to assess the performance and utility of a hybrid model, comprised of a classical pretrained portion and a quantum variational circuit. We will compare these results to purely classical approaches, one utilizing transfer learning and one without, for the stated datasets. While more research should be done for proving generalized quantum advantage, our work shows potential quantum advantages in validation accuracy and sensitivity for the specified task, particularly when training with limited data availability in a minimally skewed dataset under specific conditions. Utilizing the IBM’s Qiskit Runtime Estimator with built in error mitigation, our experiments on a physical quantum system confirmed some results generated through simulations.