New noise reference levels show lower-dose protocols are 35% more likely to fall short on image quality
Imalogix establishes first national benchmarks for CT image quality
Imalogix, the radiology operations platform that handles roughly one in five CT studies performed in the United States, has established the first national CT image quality benchmarks. Called noise reference levels (NRLs) and achievable noise (AN), they are the image-quality counterparts to the diagnostic reference levels (DRLs) and achievable dose (AD) that have benchmarked CT radiation dose for years. By comparing dose and noise together, a facility can see not only whether a protocol’s dose is high, but whether a lower dose has come at the cost of diagnostic quality. Imalogix’s 2026 data suggests it often has.
Across more than 10 million exams, lower dose protocols (below the AD) were 35% more likely to produce images with suboptimal quality (above the NRL) than protocols above the AD for the same exam type. ICRP Publication 135 warned of exactly this, stating that a low dose by itself says nothing about whether a protocol is optimized; image quality has to be considered alongside it. Without a national benchmark for noise, however, there was no practical way to heed the warning, and dose remained the only metric scored, from accreditation to Leapfrog’s pediatric CT measure.
Evaluating noise and dose together also reveals which protocols are already performing well. Analyzing exam type and scanner model combinations, Imalogix researchers identified reference protocols in use at five or more facilities operating below both the DRL and the NRL, and modeled universal adoption of that protocol across scanners. They found that the number of acquisitions exceeding the NRL would decrease by 28.8%, while the median dose fell by 3.4%, allowing for optimization opportunities across nearly all CT models and protocols. The small median dose change is a meaningful result: nearly as many protocols would need their doses increased to bring noise below the NRL as would need it lowered to bring dose below the DRL.
This logic underlies Harmonize, the Imalogix feature that benchmarks dose and image quality in tandem. Its reference levels are built from more than 5.2 million CT acquisitions performed in 2025, and its NRLs are calculated using the same size-normalized, water-equivalent-diameter methodology as the updated U.S. DRLs published this year in Radiology. NCRP Report 172 describes reference levels as dynamic values that should be reassessed as technology and clinical practice advance, and Imalogix will release new benchmarks annually, a cadence the platform’s scale makes possible.
Harmonize scores each protocol’s dose and noise against those levels and resolves it into one of four states: above NRL, above DRL, both, or neither. It also evaluates how a protocol performs across the full range of patient sizes, highlighting those that do not deliver appropriate dose and diagnostic image quality for every patient. Where a protocol falls short, Harmonize surfaces the best-performing protocol running on comparable scanners as a target for the next adjustment.
Optimization has always required both dose and image quality. Until now, only one of them could be measured against a national standard.
