INTRODUCTION: Magnetic resonance imaging (MRI) has the potential to aid in determining the presence and extent of cracks/fractures in teeth because of better contrast without ionizing radiation. The objectives were to develop MRI criteria for root crack/fracture identification and to establish reliability and accuracy in their detection. METHODS: MRI-based criteria for crack/fracture appearance was developed by an MRI physicist and a panel of 6 dentists. Twenty-nine human adult teeth previously extracted after a clinical diagnosis of a root crack/fracture were frequency matched to 29 controls. Samples were scanned using an in vivo MRI protocol and the reference standard (ie, ex vivo limited field of view cone-beam computed tomographic [CBCT] imaging). A blinded, 4-member panel evaluated the images with a proportion randomly retested to establish intrarater reliability. Overall observer agreement, sensitivity, and specificity were computed for each imaging modality. RESULTS: Subjectively, MRI has increased crack/fracture contrast and is less prone to artifacts from radiodense materials relative to CBCT imaging. Intrarater reliability for MRI was fair to excellent (kappa = 0.38-1.00), and for CBCT imaging, it was moderate to excellent (kappa = 0.66-1.00). Sensitivity for MRI was 0.59 (95% confidence interval [CI], 0.39-0.76; P = .46), and for CBCT imaging, it was 0.59 (95% CI, 0.59-0.76; P = .46). Specificity for MRI was 0.83 (95% CI, 0.64-0.94; P < .01), and for CBCT imaging, it was 0.90 (95% CI, 0.73-0.98; P < .01). CONCLUSIONS: Despite advantages of increased contrast and the absence of artifacts from radiodense materials in MRI, comparable measures of sensitivity and specificity (to limited field of view CBCT imaging) suggest MRI quality improvements are needed, specifically in image acquisition and postprocessing parameters. Given the early stage of technology development, there may be a use for MRI in detecting cracks/fractures in teeth.