On the Feasibility of Machine Learning Augmented Magnetic Resonance for Point-of-Care Identification of Disease