Mammography is the toughest test in imaging physics: the findings sought are sometimes microcalcifications smaller than a millimeter, sometimes low-contrast masses nearly invisible within fatty tissue. And they must be shown while giving a healthy breast the lowest possible dose. This delicate balance — high image quality + low dose — can only be held by very rigorous quality control. That is why mammography has the strictest legal QC framework in imaging (e.g. MQSA in the US).2
Why the most rigorous QC?
The margin for error in mammography is very narrow. Microcalcifications are among the most important signs of early breast cancer; if noise or resolution degrades even slightly, these tiny specks can vanish. At the same time the breast is a radiosensitive organ, so "rescuing" the image by raising dose is not an acceptable solution. QC is critical precisely because it holds both ends together: the image must be both good enough and low-dose enough.
Accreditation phantom
At the heart of mammography QC is the accreditation phantom. As Bushberg describes it, this phantom mimics the attenuation of a "standard breast" — 4.2 cm compressed, 50% adipose / 50% glandular — and contains a PMMA block, a wax insert and a contrast disk on top.1 The wax insert holds three kinds of structure: six nylon fibers (mimicking fibrous structures), five speck groups (six aluminum-oxide specks each — mimicking microcalcifications) and five disks simulating masses.1 The test counts how many of these are visible; the pass criterion depends on system type (e.g. 5 fibers/4 speck groups/4 masses for direct flat panel; 4/3/3 for CR and indirect flat panel).1
SNR and CNR
The visual count is completed by quantitative measures. Placing regions of interest (ROI) on the phantom image, two ratios are computed: SNR (signal-to-noise ratio) = the background ROI's mean signal ÷ its standard deviation; CNR (contrast-to-noise ratio) = the difference of the object and background means ÷ the background standard deviation.1 The technologist measures these weekly and verifies they are within the manufacturer's limits; if a value falls out of tolerance, the system may not be used clinically until the cause is found and corrected.1
kVp and HVL
Beam quality is critical in mammography, because too-low energies raise dose while too-high energies cut contrast. So the half-value layer (HVL) is kept within tight bounds: MQSA sets a minimum HVL to ensure adequate effective energy (about kV/100 + 0.03 mm Al with the compression paddle in place); ACR recommends a maximum HVL to preserve contrast.1 A beam "harder" than optimal can indicate too much filtration, a pitted anode or an aged tube, and reduces both output and image quality.1
Compression and AEC
Compression, applied firmly, separates overlapping tissue and reduces breast thickness — and hence scatter and dose; a thinner breast means less noise and better contrast.1 Automatic exposure control (AEC) sets the exposure according to breast thickness and density; that the AEC gives a consistent, reproducible image across thicknesses is tested regularly. Compression force and AEC consistency are core QC items for both dose and image quality.
Average glandular dose
Dose in mammography is expressed as the average glandular dose (MGD) — because what matters for cancer risk is the dose to the radiosensitive glandular tissue. QC verifies, with phantom measurements, that the MGD stays in the expected range. How MGD is defined and computed is covered separately: Dose in Mammography (MGD).
Who, how often?
Mammography QC is an explicitly shared duty.1 Technologists do the frequent tests: the weekly phantom evaluation and SNR/CNR, visual artifact checks, compression and daily device checks. Medical physicists handle the broader, less frequent tests (typically annual and after installation/repair): HVL/beam quality, AEC performance, resolution, MGD measurement and geometric checks. Mammography demands not only equipment performance but the technical competence of the radiologist, technologist and physicist.1
References
- Bushberg JT, Seibert JA, Leidholdt EM, Boone JM. The Essential Physics of Medical Imaging, 3rd ed. Lippincott Williams & Wilkins, 2011. Bölüm 8 (Mamografi): mamografi akreditasyon fantomu — mum insert içinde altı naylon lif, beş alüminyum oksit speck (mikrokalsifikasyon) grubu (her grupta altı taneç) ve kitleyi taklit eden beş disk; standart meme 4,2 cm sıkıştırılmış, %50 yağ/%50 glandüler (s.279–281, Şekil 8-32). Geçme ölçütü üreticiye göre değişir (ör. 5 lif/4 speck grubu/4 kitle ya da 4/3/3). SNR = arka plan ROI ortalaması / standart sapması; CNR = (nesne − arka plan ortalaması) / arka plan standart sapması; teknikerin haftalık ölçmesi gerekir (s.281). HVL: MQSA asgari sınırları (≈ kV/100 + 0,03 mm Al, kompresyon paleti ile), ACR azami sınırları (Tablo 8-2/8-3). Sayfa numaraları bu baskıya aittir.
- Mammography Quality Standards Act (MQSA), 21 CFR Part 1020.30 ve ACR Akreditasyon Programı — mamografi ekipman performansı, HVL sınırları, fantom değerlendirme ölçütleri ve onarım toleransları. accessdata.fda.gov
- Kesin tolerans değerleri ve test sıklıkları ulusal düzenlemeler ile uluslararası protokoller (EUR/EUREF, IAEA, AAPM TG-282) ve kurum prosedürlerince belirlenir. Ortalama glandüler doz (MGD) için bkz. Mamografide Doz (MGD).