Early detection of breast cancer is diagnosed using mammography, the gold standard in breast screening. However, its increased use also provokes radiation-induced breast malignancy. Thus, monitoring and regulating the mean glandular dose (MGD) is essential. The purpose of this study was to determine MGD for full-field digital mammography (FFDM) and digital breast tomosynthesis (DBT) in the radiology department of a single centre. We also analysed the exposure factors as a function of breast thickness. A total of 436 patients underwent both FFDM and DBT. MGD was auto calculated by the mammographic machine for each projection. Patients’ data included compressed breast thickness (CBT), peak kilovoltage (kVp), milliampere-seconds (mAs) and MGD (mGy). Result analysis showed that there is a significant difference in MGD between the two systems, namely FFDM and DBT. However, the MGD values in our centre were comparable to other centres, as well as the European guideline (<2.5 mGy) for a standard breast. Although DBT improves the clinical outcome and quality of diagnosis, the risk of radiation-induced carcinogenesis should not be neglected. Regular quality control testing on mammography equipment must be performed for dose monitoring in women following a screening mammography in the future.

A set of national diagnostic reference levels (DRLs) was established in Malaysia for a range of breast thicknesses in 2013, but no updates for full-field digital mammography (FFDM) and digital breast tomosynthesis (DBT). Due to the increasing number of DBTs used and concern over radiation exposure, this study aimed to explore and establish local diagnostic reference levels for FFDM and DBT in Malaysia health facilities at different compressed breast thickness (CBT) ranges. The CBT, kilovoltage peak (kVp), Entrance surface dose (ESD), and average glandular dose (AGD) were retrospectively extracted from the mammography Digital Imaging and Communications in Medicine (DICOM) header. The 75th and 95th percentile values were obtained for the AGD distribution of each mammography projection for three sets of CBT range. The difference in AGD values between FFDM and DBT at three CBT ranges was determined. The DRLs for FFDM were 1.13 mGy, 1.52 mGy, and 2.87 mGy, while DBT were 1.18 mGy, 1.88 mGy, and 2.78 mGy at CBT ranges of 20–39 mm, 40–59 mm, and 60–99 mm, respectively. The AGD of DBT was significantly higher than FFDM for both mammographic views (p < 0.005). All three CBT groups showed a significant difference in AGD values for FFDM and DBT (p < 0.005). The local DRLs from this study were lower than the national DRLs, with the AGD of FFDM significantly lower than DBT.

As the total volume of mammograms in Dubai is increasing consistently, it is crucial to focus on the process of dose optimization by determining dose reference levels for such sensitive radiographic examinations as mammography. This work aimed to determine local diagnostic reference levels (DRLs) for mammography procedures in Dubai at different ranges of breast thickness. A total of 2599 anonymized mammograms were randomly retrieved from a central dose survey database. Mammographic cases for screening women aged from 40 to 69 years were included, while cases of breast implants and breast thickness outside the range of 20–100 mm were excluded. Mean, median, and 75 percentiles were obtained for the mean glandular dose (MGD) distribution of each mammography projection for all compressed breast thickness (CBT) ranges. The local DRLs for mammography in Dubai were found to be between 0.80 mGy and 0.82 mGy for the craniocaudal (CC) projection and between 0.89 mGy and 0.971.8 mGy for the mediolateral oblique (MLO) projection. Local DRLs were proposed according to different breast thicknesses, starting from 20 to 100 mm. All groups of CBT showed a slight difference in MGD values, with higher values in MLO views rather than CC views. The local DRLs in this study were lower than some other Middle Eastern countries and lower than the standard reference levels reported by the International Atomic Energy Agency (IAEA) at 3 mGy/view.

Background/Objective: This study evaluates the mean glandular dose (MGD) in mammography screening for women aged 40–69 in Dubai, based on a retrospective analysis of a dose survey involving 2599 participants. Methods: MGD was calculated using the Dance formula. Results: The average MGD was 0.96 ± 0.39 mGy for mediolateral oblique (MLO) views and 0.81 ± 0.33 mGy for craniocaudal (CC) views. Weak inverse correlations were found between age and organ dose (OD) for both views, while a direct relationship was observed between breast thickness and entrance skin dose (ESD). In adjusted models, ESD was strongly associated with MGD (β = 1.04, 95% CI: 0.97, 1.09), while OD showed a moderate association (β = 0.44, 95% CI: 0.40, 0.49). Significant variations in ESD, OD, and MGD were noted across age groups and breast thicknesses. Conclusions: Lower MGD indicates reduced radiation exposure risk, while higher MGD in MLO views suggests improved imaging quality. Monitoring and optimizing MGD are essential for enhancing patient safety and screening efficacy.

Background and Aims: Machine learning (ML) has shown promise in diagnosing tuberculosis (TB), but systematic evidence on its role in predicting and diagnosing drug‐resistant tuberculosis (DR‐TB) is lacking. This study integrates a systematic review and meta‐analysis to consolidate ML's performance in DR‐TB diagnosis and prediction to promote artificial intelligence in this field. Methods: Relevant studies were retrieved from PubMed, Cochrane, Embase, and Web of Science up to August 20, 2025, complemented by a manual search of Google Scholar. Risk of bias was evaluated with PROBAST. A bivariate mixed‐effects model pooled accuracy measures, with subgroup analyses stratified by ML tasks (diagnosis and prediction). Results: Twenty‐six studies, including 35,472 participants, were analysed. Diagnostic models outperformed prediction models, with a higher pooled AUC (0.94 vs. 0.87). Deep learning (DL)‐based diagnostic models consistently surpassed traditional ML across all key metrics, AUC (0.97 vs. 0.89). In the diagnostic model, internal validation showed superior performance to external validation AUC (0.95 vs. 0.85), and in the predictive model, the overall performance of the model in internal validation is slightly better than that in external validation AUC (0.88 vs. 0.85). Conclusion: ML models, particularly DL, demonstrate high diagnostic efficacy for DR‐TB, though performance declines in external data sets. Predictive models show moderate accuracy but remain useful for early risk stratification. Large multi‐center validations are needed to ensure robustness and clinical applicability.