Access publication
Key message
Over roughly 25 years, combined advances in CT hardware, acquisition, and reconstruction have enabled approximately 2- to 10-fold radiation-dose reductions, but dose reduction should be judged against preserved diagnostic performance, not merely lower image noise.
Practical implication:
Dose optimization should be task based rather than technology based. Clinical observer studies remain the reference standard, while quantitative model-observer and virtual-imaging-trial approaches may provide more efficient ways of determining how far dose can safely be reduced for a specific diagnostic task.
Figure 1. Bar chart shows anticipated reduction in dose owing to cumulative effect of various dose-reduction approaches for single-phase CT of the abdomen and pelvis. For most multidetector CT (MDCT) systems with at least 64 detector rows, the effective dose to a reference adult (≈ 70–80 kg) is estimated to be 10 mSv. Each successive dose-reduction approach is applied to the previous dose estimate, ordered according to the likelihood of widespread clinical availability and use. After all methods discussed herein are available clinically, the effective dose associated with this examination is anticipated to have been reduced by approximately a factor of 10, from 10 to 1.1 mSv. Interior tomography has not yet been adopted due to technical challenges in modulating x-ray beam profile in real time. Deep learning–based reconstruction and postprocessing have instead become available. Reprinted with permission from McCollough et al. [9].
Fig. 8A —Bar graphs show radiation dose reduction for abdominal CT examination for two commercial systems that automatically select optimal energy behavior for final image. Data were previously published by Rajendran et al. [42]. Dose reductions are relative to technique used for noncontrast evaluation.
A, Bar graph shows radiation dose reduction for photon-counting detector CT system. Automated selection of energy level for virtual monoenergetic image (CAREkeV, Siemens Healthineers) selects energy level and dose reduction based solely on diagnostic task (bone, contrast-enhanced soft-tissue, or vascular evaluation), not on patient size (20, 30, 40, or 50 cm).
Fig. 8B —Bar graphs show radiation dose reduction for abdominal CT examination for two commercial systems that automatically select optimal energy behavior for final image. Data were previously published by Rajendran et al. [42]. Dose reductions are relative to technique used for noncontrast evaluation.
B, Bar graph shows radiation dose reduction for energy-integrating detector CT system. Automated selection of optimal tube potential (CAREkV, Siemens Healthineers) determines optimal tube potential based on both diagnostic task and patient size.
Photon-counting detector CT (PCCT) is highlighted as a major recent advance in dose efficiency. Its advantages arise from electronic-noise rejection, elimination of interelement septa, improved photon-energy weighting, and native ultra-high-resolution acquisition.
The UHR capability of PCCT is particularly relevant for high-spatial-resolution tasks: detector pixels can be as small as 0.151 mm, permitting up to 40 lp/cm and 0.2-mm slices. Importantly, smaller detector pixels do not inherently impose a noise penalty when images are reconstructed to the same target spatial resolution; the excess resolution can instead be traded for lower noise and/or lower dose.
This advantage is task-dependent. The benefit of UHR acquisition is strongest for tasks requiring high spatial resolution, whereas it is relatively small for more moderate-resolution, low-contrast tasks such as brain, liver, or pancreas lesion detection/characterization. In the temporal bone example, PCCT provided better trabecular and stapes visualization at 30% lower dose than EID-CT.
Other important dose-saving technologies include added beam filtration (40–60%), dynamic z-axis collimation (up to ~40%), automatic tube-current modulation (~25–40%), and automatic tube-potential selection (up to 60%), although the achievable reduction depends on examination and diagnostic task.
DLR requires caution. Although published studies report substantial dose-reduction potential, aggressive denoising can make images appear excellent while small or low-contrast lesions become blurred or disappear; at extreme dose reduction, DLR may also introduce false-positive structures (“hallucinations”). Therefore, apparent image quality or CNR alone is insufficient.
Source: McCollough CH, Yu L. CT Radiation Dose Reduction With Preserved Diagnostic Performance: How Far Have We Come Over 25 Years? American Journal of Roentgenology. 2026;226(6). doi:10.2214/AJR.25.34450.
