Radiologists across nine Emirates Health Services facilities in the United Arab Emirates now have their first locally derived radiation dose benchmarks for mammography, the result of a multi-center analysis of 1,184 examinations performed between February 2023 and June 2025. The study, published in the Egyptian Journal of Radiology and Nuclear Medicine, established Local Diagnostic Reference Levels, or LDRLs, for both two-dimensional full-field digital mammography and three-dimensional digital breast tomosynthesis, giving the health network a quantitative yardstick against which every mammography unit can be judged.
Diagnostic reference levels are a radiation protection tool introduced by the International Commission on Radiological Protection. They are not safety limits in the legal sense; rather, they are benchmarks, typically set at the 75th percentile of the patient dose distribution, intended to flag unusually high doses and prompt optimization where needed. In mammography, the dose metric of choice is the mean glandular dose, the absorbed dose to the glandular tissue of the breast, which the authors describe as the most accurate parameter for assessing patient exposure because glandular tissue is the most radiosensitive component of the breast. Compressed breast thickness, breast density, kilovoltage and milliampere-seconds all influence this value, which is why dose optimization must be balanced against the need for diagnostic image quality.
The study was designed as a descriptive cross-sectional survey of routine clinical practice rather than a prospective trial. The researchers collected data retrospectively from women undergoing screening and diagnostic mammography at nine EHS hospitals and healthcare centers distributed across different Emirates. All examinations were performed on full-field digital mammography or digital breast tomosynthesis systems equipped with automatic exposure control, and standard craniocaudal and mediolateral oblique projections were acquired by qualified radiographers following standardized positioning protocols. Sampling followed the recommendations of ICRP Publication 135, which advise including at least 50 patients per mammography unit to ensure reliable reference level estimation; the authors report that this criterion was fulfilled at every participating facility.
A key methodological choice was the restriction of the analysis to examinations with a compressed breast thickness of 5.0 plus or minus 0.5 centimeters, that is, between 4.5 and 5.5 centimeters. Because breast thickness strongly affects the dose required to produce an image, stratifying by thickness band makes the resulting percentiles more clinically meaningful and allows more direct comparison with international benchmarks. Exposure parameters extracted from DICOM headers and dose reports included patient age, compressed breast thickness, tube voltage, tube current-time product, half-value layer, entrance surface air kerma and mean glandular dose. The mean glandular dose values were automatically calculated by the mammography systems themselves, based on manufacturer-specific dosimetry algorithms; where the Dance et al. methodology applied, the dose was expressed as the entrance surface air kerma multiplied by conversion factors accounting for glandular energy absorption, X-ray spectrum variation and breast composition. For tomosynthesis, DBT-specific correction factors were incorporated within the vendor dose-reporting algorithms.
The headline numbers are straightforward. For two-dimensional mammography, mean glandular dose averaged 1.05 plus or minus 0.51 milligray for the craniocaudal projection and 1.15 plus or minus 0.57 milligray for the mediolateral oblique projection, with a mean tube voltage of approximately 29 kilovolts for both views and mean milliampere-second values of 72.79 and 74.94 respectively. For three-dimensional tomosynthesis, doses were markedly higher: 1.87 plus or minus 0.44 milligray for craniocaudal and 2.08 plus or minus 0.37 milligray for mediolateral oblique views, acquired at mean tube voltages of 32.61 and 33.44 kilovolts. Interestingly, the tomosynthesis examinations used lower milliampere-second values than the 2D acquisitions, reflecting the tomosynthesis technique of stacking many individually low-dose projections, acquired over a limited angular range of roughly 15 to 25 degrees depending on vendor configuration, into a single volumetric image.
Applying the 75th percentile criterion to the dose distributions produced local reference levels of 1.30 milligray for the craniocaudal projection and 1.46 milligray for the mediolateral oblique projection, applicable to examinations at the standardized compressed breast thickness. The authors report that these values fall within internationally accepted reference ranges recommended by the ICRP and the International Atomic Energy Agency. In their international comparison, the EHS 75th percentile doses were comparable to those reported in Greece in 2019 and Morocco in 2021, lower than values from several other Middle Eastern and African countries including Malaysia, Nigeria, Qatar and Turkey, and slightly higher than values reported in Dubai, Palestine, Ghana and Saudi Arabia. The authors attribute such variation to differences in mammography system generations, population breast characteristics, clinical protocols and national optimization recommendations.
The higher doses for tomosynthesis are not a surprise. Tomosynthesis acquires multiple low-dose projections over a wider angular range, so the cumulative glandular dose is inherently higher than for a conventional two-dimensional image, and the additional reconstruction steps needed to improve lesion detectability and reduce tissue overlap add to the exposure. Among all the technique-projection combinations evaluated, two-dimensional craniocaudal imaging delivered the lowest dose while three-dimensional mediolateral oblique imaging delivered the highest. The authors also note a consistent dose difference between the two projections, which they suggest may reflect differences in breast compression, tissue thickness and acquisition geometry. A statistically significant difference in mean glandular dose between full-field digital mammography and tomosynthesis was confirmed, with a p value below 0.05.
Perhaps the most reassuring finding for the health network is its internal consistency. Facility-level analysis showed uniform dose distributions across all participating sites, with no mammography unit persistently exceeding the established local reference levels and no corrective actions required during the study period. The authors interpret this uniformity as evidence of standardized imaging protocols and effective quality control, reinforced by regulatory oversight: all included units undergo annual calibration and quality-control testing under the supervision of the Federal Authority for Nuclear Regulation, including verification of tube output consistency, half-value layer and automatic exposure control performance. They also report that mean glandular dose increased with compressed breast thickness for both modalities, a pattern they describe as consistent with the appropriate functioning of automatic exposure control systems, which adapt exposure parameters to maintain image quality.
The study’s authors frame the new reference levels as a practical optimization tool rather than a verdict on their network. Radiology teams can now benchmark their own unit performance against the 1.30 and 1.46 milligray values and identify potential areas for dose reduction without compromising image quality. The findings are also positioned as baseline evidence for a larger ambition: the development of a national mammography dose registry for the United Arab Emirates, which would support patient safety, regulatory compliance and continuous monitoring of radiation exposure trends over time.
The authors are candid about the limitations. The study covered only public healthcare facilities under Emirates Health Services, which may limit generalizability to the private sector, and objective image quality metrics were not evaluated, although all examinations met clinical diagnostic standards under established quality assurance programs. Future work, they suggest, should expand to multiple categories of healthcare providers and integrate quantitative image quality assessment alongside dose metrics. Within those boundaries, the study delivers what it set out to provide: the first large multi-center assessment of mammography doses within the EHS network, covering both two-dimensional and three-dimensional systems, with compressed breast thickness stratification applied according to ICRP Publication 135, and a set of local reference levels that give the UAE a concrete starting point for national dose optimization in breast imaging.
Subject of Research: Establishment of local diagnostic reference levels for 2D and 3D mammography radiation doses in Emirates Health Services facilities
Article Title: Establishment of local diagnostic reference levels for mammography: optimizing patient radiation dose within Emirates Health Services
Article References: Al Jasmi, A., Alameen, S., Al Hajri, M., Al Msafri, S., Al Zarouni, M., Ghonim, H., El Gamal, T., Khammas, A., & Abuzaid, M. (2026). Establishment of local diagnostic reference levels for mammography: optimizing patient radiation dose within Emirates Health Services. Egyptian Journal of Radiology and Nuclear Medicine, 57(1), Article 216. https://doi.org/10.1186/s43055-026-01883-0
Image Credits: AI Generated
DOI: 10.1186/s43055-026-01883-0
Keywords: mammography, diagnostic reference levels, mean glandular dose, digital breast tomosynthesis, radiation protection, Emirates Health Services, United Arab Emirates, ICRP Publication 135, compressed breast thickness, dose optimization, full-field digital mammography, quality control
News Source: Ophelia Keating. (October 29, 2026). UAE Health Network Sets Local Dose Benchmarks for 2D and 3D Mammography. Scienmag.
