16 January 2026 in European Journal of Surgical Oncology
Accurate estimation of tumour extent is crucial for breast cancer treatment and achieving tumour-free surgical margins. Breast MRI is a widely used imaging modality; however, the high cost and limited availability of MRI have increased interest in contrast-enhanced mammography (CEM) as a potential alternative. This meta-analysis compares the accuracy of tumour extent estimation between MRI and CEM, using histopathological size as the reference standard. Literature up to October 2024 was reviewed. Two independent reviewers extracted relevant data. For both modalities in comparison with pathology, mean differences in tumour size and pooled correlation coefficients were calculated. A direct comparison of tumour size estimates between both modalities was conducted using R2 and a two-sample t-test. The I2-statistic was used to assess between-study heterogeneity. Twenty-one studies involving 1060 patients were included. Both modalities showed a mean overestimation of tumour size by ∼3 mm, with no statistical significance (P = 0.345 and P = 0.308). Correlation coefficients with pathology were 0.78 for MRI and 0.79 for CEM. Direct comparison demonstrated strong agreement (R2 = 0.81) between the two modalities, and no significant difference (P = 0.65). Limited data suggested higher sensitivity but reduced specificity for MRI in the detection of multifocal/multicentric disease. MRI and CEM showed comparable performance in tumour size estimation, with no significant differences compared to pathology or each other. These findings suggest CEM could potentially be an alternative for MRI, especially in resource-limited settings or for patients contraindicated for MRI. However, small study sizes and limited subgroup data highlight the need for large-scale prospective research.
9 March 2026 in Seizure: European Journal of Epilepsy
Deep learning (DL) techniques may support localizing the epileptogenic zone (EZ) and improve surgical outcomes in drug-resistant epilepsy. This systematic review synthesizes current evidence on DL-assisted EZ localization from neuroimaging acquisitions, aiming to outline methodological trends, limitations, and future directions that bridge the gap between clinical translation and technological advances. We systematically searched PubMed, Scopus, and Embase (via Ovid) on April 15, 2025, for studies applying DL to localize the EZ using neuroimaging data. The bias and applicability of studies was assessed using the PROBAST+AI tool. We extracted methodological details, as well as key performance metrics. Thirty-six studies met the eligibility criteria, most focusing on segmenting epileptogenic lesions using structural MRI. Focal cortical dysplasia was the most commonly targeted pathology, with fully convolutional networks being the predominant DL architecture. Approximately two-thirds of the studies showed high risk of bias and clinical applicability concerns, limited by non-representative cohorts and suboptimal evaluation methods. Five studies reported promising EZ detection rate in MRI-negative cases using large multi-center cohorts, yet progress in fine-grained localization tasks, such as lesion segmentation, remained moderate. This review highlights methodological limitations hindering the clinical translation of current DL approaches for EZ localization and provides a comprehensive set of recommendations to address them. Future work should prioritize developing standardized, clinically informative evaluation frameworks and explore research avenues aligned with modern DL practices, spanning from uncertainty quantification to large-scale vision foundation models and synthetic data generation.