Introduction: The technical objective of total knee arthroplasty (TKA) is to restore normal mechanical parameters to the knee. Patient-specific instrumentation (PSI) was developed to streamline the operative process and improve accuracy. PSI produces individualized cutting guides based on three-dimensional models of the patient's anatomy acquired from computed-tomography (CT) or magnetic-resonance imaging (MRI). However, the superiority of one modality over the other remains unclear. Therefore, we aimed to compare the accuracy of patient-specific cutting guides produced from MRI or CT imaging methods in TKA. Methods: Electronic databases were systematically searched using relevant keywords and MeSH terms for original-data English-language publications comparing the accuracy of CT and MRI-based PSI cutting guides in TKA. Data was extracted from the text, tables and figures of studies and meta-analysed. Results: MRI-based PSI cutting guides produced a lower proportion of coronal plane outliers (>3°) with regard to overall limb mechanical axis (OR 2.75, p = 0.01). There were no significant differences between the two in terms of sagittal femoral and tibial component placement, or coronal femoral and tibial placement, or femoral component axial rotation. Tibial rotation was not analysed in the literature. Conclusions: MRI-based patient-specific cutting guides produced a lower proportion of outliers in the overall coronal alignment of the limb compared to CT, with no significant difference between the two in terms of femoral or tibial component placement. Future studies should investigate the differences in resource usage and operative time between the two to inform surgeons' decision making when choosing an ideal imaging modality for PSI TKA. Study design: Meta-analysis. Level of evidence: III, systematic review of cohort and comparative studies.