CT is the only clinical imaging modality whose gray scale is numerically meaningful: every pixel carries a Hounsfield unit (HU), and HU is defined relative to water by definition — so water = 0 HU, air ≈ −1000 HU.2 This numerical meaning is a privilege but also a responsibility: if water does not read 0, every diagnostic value has shifted. Quality control (QC) is the process that regularly checks this — plus noise, resolution and dose. QC also catches artifacts early, because a ring or loss of uniformity usually shows up first in the QC phantom.
Why QC in CT?
In CT one trigger affects three things at once: diagnostic accuracy, image quality and patient dose. If calibration drifts a lesion may be missed; if noise rises the technologist tends to raise dose; if resolution drops fine structures vanish. So QC is not just "is the machine working?" but a foundation of dose optimization. The tests below form the measurable backbone of CT performance.
CT number accuracy and uniformity
The most basic test: does a known material read its expected HU? Scanning a uniform water phantom should give a mean of about 0 HU; other known materials (air, low-density polyethylene, PMMA, Teflon) should land at their expected values too.1 Uniformity means the same water reads the same HU at the center and the periphery; a center-to-periphery difference signals a deviation (e.g. cupping). Together these two tests secure the "zero point" and flatness of the entire gray scale.
Noise
Noise is the standard deviation of HU values in a uniform region, read from the same water-phantom ROI. Noise is directly tied to dose: fewer photons → more noise (the √N relationship). So the noise test is also a dose-consistency indicator — noise higher than expected may signal a problem or an under-exposed output. Detail: Dose and Noise (√N).
Spatial resolution
The ability to resolve small high-contrast structures is assessed two ways: subjectively with a line-pairs pattern, and objectively with the modulation transfer function (MTF) computed from the point spread function (PSF) of a small tungsten bead.1 Resolution along z is defined by the slice sensitivity profile and its full width at half maximum (FWHM). On current 64-slice systems this is roughly 0.6–0.9 mm in all three dimensions.1 MTF is usually measured by a medical physicist at acceptance/commissioning and after major upgrades.1
Low-contrast detectability
If spatial resolution asks "how small," low-contrast detectability asks "how faint": the ability to see structures whose HU differs only slightly from their surroundings. Phantoms like the CatPhan use acrylic inserts of varying diameter around the center to measure the effect of object size on low-contrast visibility.1 This test is very sensitive to noise; as noise rises, low-contrast structures disappear — so it is directly tied to dose.
Slice thickness
The nominal slice thickness must match what is actually measured. This is verified with the FWHM of the slice sensitivity profile along z.1 Slice thickness is a critical QC parameter because it affects resolution (thinner → better z resolution, less partial volume) as well as noise and dose.
Dose and geometry
Alongside image quality, QC also checks dose: CT's standard dose index is CTDIvol, verified regularly with phantom measurements (detail: Dose in CT). Geometric tests cover laser alignment, table position/increment accuracy and gantry alignment; for radiotherapy-planning CT, laser alignment and the HU→electron density calibration are additionally critical and are done by a medical physicist.1
Who, how often?
QC is layered. Technologists do the frequent (often daily), quick checks: CT number and noise on a water phantom, a visual artifact scan, warm-up/calibration. Medical physicists handle the broader, less frequent tests: MTF/resolution, slice thickness, low contrast, dose (CTDI) and geometric checks — especially at acceptance/commissioning, periodic review, and after major upgrades.1
References
- IAEA. Diagnostic Radiology Physics: A Handbook for Teachers and Students (STI/PUB/1564), 2014. Bölüm 11 (CT): CatPhan fantomu — HU değerleri (hava, düşük yoğunluklu polietilen, PMMA, Teflon), düşük kontrast akrilik insertler, yüksek kontrast çizgi çiftleri, tungsten boncuk PSF'i ve homojenlik (s.282–283); uzaysal çözünürlük PSF→MTF ve z ekseni için kesit duyarlılık profili/FWHM, 64 kesitli sistemlerde 0,6–0,9 mm (s.282); MTF'in kabul/devreye alma ve büyük yükseltmelerde medikal fizik uzmanınca ölçülmesi (s.282); zamansal çözünürlük ~rotasyonun %50'si (s.283); RT planlamada lazer hizalama ve HU→elektron yoğunluğu kalibrasyonu (s.279). iaea.org
- Bushberg JT, Seibert JA, Leidholdt EM, Boone JM. The Essential Physics of Medical Imaging, 3rd ed. Lippincott Williams & Wilkins, 2011. Hounsfield birimi tanımı: HU = 1000·(μ − μsu)/μsu (Denklem 10-2, s.325) — bu tanım gereği su = 0 HU, hava ≈ −1000 HU. BT görüntü kalitesi (gürültü, çözünürlük) Böl.10.
- Kesin tolerans değerleri ve test sıklıkları ulusal düzenlemeler ile uluslararası protokoller (IAEA, AAPM, IEC, EFOMP) tarafından belirlenir; kurum prosedürleri ve yetkili otorite rehberleri esas alınmalıdır. Doz göstergesi CTDIvol için bkz. BT'de Doz.