Repeat-Reject Analysis in Digital Radiography

Q: Does AAPM TG-305 set a required reject rate?

No. AAPM Task Group Report 305 does not mandate a numerical reject-rate limit. Its purpose is to standardize the data elements and reject-reason categories so that reject data can be compared across vendors and sites, and to recommend workflows for an effective monitoring program. Specific action thresholds are set locally by the facility and its qualified medical physicist.

Q: How does the exposure index relate to reject analysis?

The exposure index (defined under IEC 62494-1) and its deviation index quantify whether a detector received an appropriate amount of radiation. Reviewing exposure index trends alongside reject data helps separate true exposure errors from positioning errors and identifies dose creep, where technologists drift toward higher exposures over time.

Q: How often should a facility review reject data?

Reject data should be collected continuously and reviewed at a regular interval — commonly monthly or quarterly — with at least an annual summary reviewed by the qualified medical physicist as part of equipment performance monitoring. The review should feed corrective actions, not just generate a number for the accreditation binder.

Repeat-reject analysis is the quality-control process of collecting every rejected radiograph, classifying why it was rejected, and acting on the pattern — because each rejected image is a patient exposure that produced dose but no diagnosis. Done well, it lowers patient dose, sharpens technologist technique, and surfaces equipment problems before they reach the radiologist.

A defensible program does three things: it standardizes how rejects are recorded so the numbers mean something, it analyzes rates by projection, room, and technologist rather than as a single facility-wide figure, and it closes the loop by feeding findings back into training and protocol changes. AAPM Task Group Report 305 provides the modern, vendor-neutral framework for the first of those.¹

Introduction

In film-screen radiography, a rejected image was a physical waste film you could hold in your hand, and reject analysis meant counting the discard bin. Digital radiography removed the bin. Rejected exposures now live in software logs, are easy to delete silently, and are easy to ignore — yet the patient was still irradiated. That is the central problem repeat-reject analysis exists to solve: making invisible dose visible again.¹⁴

The stakes are not trivial. Repeated diagnostic imaging is a recognized, avoidable source of population radiation exposure, and reducing unnecessary exposure from medical imaging is an explicit national priority.¹⁰ A radiograph that is rejected for a correctable reason — clipped anatomy, mispositioning, motion — represents dose delivered for no diagnostic benefit, plus the additional dose of the repeat that replaces it.

This guide explains what repeat-reject analysis is, how reject rates are defined and benchmarked, the technical principles behind a credible program, the clinical impact, practical optimization tips, and the regulatory and accreditation context. DRPS builds and audits these programs as part of its diagnostic radiography physics and accreditation support services across Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware.

Topic Explanation

What is a reject, and what is a repeat?

A reject is any acquired image that is not submitted for diagnosis; a repeat is the extra exposure made to replace it. The two are tightly linked because most rejected images trigger a retake, so reject counts are used as a practical surrogate for the excess patient dose caused by repeats.¹

A few distinctions matter for an honest program:

Rejected images are still patient exposures. Deleting the image does not delete the dose.
Not every reject is a repeat, and not every repeat is recorded as a reject. A patient who cannot tolerate repositioning may have a rejected image with no retake; conversely, a technologist who reshoots without flagging the first image hides a reject.
Quality-assurance and test exposures should be excluded from the clinical reject rate, or at least segregated, so they do not distort the denominator.¹

For context on how exposure adequacy itself is judged, see our guide to the digital radiography exposure index, and for the dose-control system that prevents many over- and under-exposures, automatic exposure control in radiography.

Why digital radiography changed reject analysis

In the film era, the dominant reject reasons were exposure errors — too light, too dark — and film-handling artifacts. Digital detectors have wide dynamic range and post-processing, so many of those errors are now recoverable. The result, repeatedly documented, is that positioning and patient motion have become the dominant reject categories, while pure exposure errors have shrunk.²³⁴

A classic comparison study found an overall reject rate of 27.6% in a conventional film-screen department versus 2.3% in a digital storage-phosphor department, with the dominant reason shifting from exposure to positioning.⁴ That dramatic drop is real, but it created a new risk: because digital images are cheap to delete and easy to "fix" by reshooting, departments can quietly accumulate repeats and dose creep without ever seeing a number.

What does TG-305 standardize?

AAPM Task Group Report 305, Guidance for standardization of vendor-neutral reject analysis in radiography (2023), addresses the core obstacle to meaningful reject analysis: every vendor's system logs rejects differently, so rates cannot be compared between rooms or institutions.¹ TG-305 recommends:

Essential data elements that should accompany every reject record (e.g., body part, projection, reject reason, technologist, device, exposure index).
A standardized reject-reason classification schema, so "positioning" means the same thing on a GE, Siemens Healthineers, Philips, Canon, Fujifilm, or Agfa system.
Workflow and reporting options for aggregating data across a multi-vendor fleet into one monitoring program.

Critically, TG-305 does not set a mandatory reject-rate number.¹ It standardizes the measurement so that local action thresholds, set by the facility and its qualified medical physicist, are defensible.

Key Technical Principles

Defining the reject rate

The overall reject rate is the fraction of all acquired exposures that were rejected:

R R = \frac{N _{reject}}{N _{total}} \times 100%, N_{total} = N_{accepted} + N_{reject}

The denominator must include the rejected exposures, because they were genuinely made. Reporting rejects against only the accepted images understates the rate.¹

Worked example. A general radiographic room acquires 12,000 exposures in a quarter. Of these, 600 are rejected. The overall reject rate is:

R R = \frac{600}{12 , 000} \times 100% = 5.0%

That single number is necessary but not sufficient. The same room might hide a horizontal-beam knee projection running at a 25% reject rate inside a healthy-looking 5% aggregate.

Projection-specific and relative reject rates

Because rejects concentrate in a few difficult projections, analysis should be broken out by body part and view. Two complementary metrics are used:¹²

Absolute reject rate for a projection = rejected images of that projection ÷ all acquired images (its share of the whole workload).
Relative reject rate for a projection = rejected images of that projection ÷ acquired images of that same projection (how often that view fails).

A chest projection can have a high absolute reject count simply because it is high-volume, while a horizontal-beam hip or knee can have a modest absolute count but a very high relative failure rate. Multi-center data illustrate this: the highest relative reject rates by body part were reported for the knee (≈18%) and pelvis (≈17%), even though chest contributed the largest absolute share.²³

Estimating wasted dose

Reject analysis becomes persuasive when translated into dose. The excess collective dose attributable to rejects can be approximated as the number of rejected exposures multiplied by the mean dose per exposure for that projection:

D_{excess} \approx i \sum N_{reject, i} \cdot \overline{D_{i}}

where $i$ indexes projection type and $\overline{D_{i}}$ is the mean incident or effective dose for that projection. Comparing $\overline{D_{i}}$ against published diagnostic reference levels helps prioritize which high-dose, high-reject projections deserve attention first.⁹

Benchmark reject rates

There is no universal mandated value, but published benchmarks give useful reference points. Multi-center digital radiography data from 44 systems across 11 hospitals reported a median reject rate of 9.1%, with a local reject reference level (the 75th percentile of observed rates) of 10.6%; patient positioning accounted for about 76% of all rejects and patient motion about 7.5%.² A separate emergency-department audit reported an overall reject rate of 10.3%, again dominated by positioning.³

Source / standard	Setting	Overall reject rate	Dominant reject reason
Peer et al. (2001) ⁴	Film-screen vs. digital	27.6% (film) vs. 2.3% (digital)	Exposure (film); positioning (digital)
Stephenson-Smith et al. (2021) ³	Emergency digital radiography	10.3%	Positioning (≈67% of multiple-reject cases)
Serra et al. (2024) ²	Multi-center public hospitals	Median 9.1% (LRRL 10.6%)	Positioning (≈76%)
AAPM TG-305 (2023) ¹	Guidance framework	No fixed target	Standardizes reason categories

The pattern is consistent: a contemporary digital radiography reject rate frequently sits in the high single digits to low teens, and positioning dominates. A rate far above local reference levels signals a problem; a rate near zero usually signals under-reporting, not excellence.¹²

Reject analysis is not only a radiography problem

Although reject analysis is mandated in mammography and recommended in general radiography, the same logic extends to CT. An automated, multi-institutional study of more than 61,000 CT examinations quantified repeat acquisitions from DICOM metadata and found repeat rates that varied by protocol and site, exceeding 9% for some protocols — a reminder that "wasted" acquisitions are a cross-modality dose-management issue, not a radiography quirk.⁵

Clinical Impact

The clinical value of repeat-reject analysis is concentrated, actionable dose reduction. Because rejects cluster in specific projections, technologists, and rooms, a good program points directly at where intervention will help most.¹²

Patient dose. Every avoided repeat is avoided dose. Targeting the few projections with the highest relative reject rates yields the largest dose savings per unit of effort.²⁹
Throughput and access. Repeats consume room time and delay care. A room that retakes 1 in 10 horizontal-beam exams is losing measurable capacity.
Diagnostic quality. Reject patterns reveal recurring image-quality failures — collimation, centering, exposure factors — that, if uncorrected, eventually reach the radiologist as borderline-diagnostic studies.
Equipment health. A sudden rise in "detector artifact" or "exposure error" rejects on one unit can be the first sign of a failing detector, AEC drift, or generator problem, prompting a service call before the room produces non-diagnostic studies.

There is also a downside risk to manage: an obsessive focus on a low reject number can push technologists to under-report or to accept marginal images to avoid a retake. The goal is the right images at the lowest reasonable dose — not the lowest possible reject count.¹

Practical Optimization Tips

1. Standardize reject reasons before you trust the data

Adopt the TG-305 reason categories (or map your vendors' categories onto them) so that "positioning," "motion," "exposure," "anatomy cut-off," "artifact," and "QA/test" mean the same thing across every room.¹ Free-text or vendor-default categories make cross-room comparison meaningless.

2. Analyze by projection and technologist, not just facility-wide

A single aggregate rate hides the actionable detail. Break the data down by projection (absolute and relative rates), by room/device, and by technologist — the last handled constructively and confidentially, as a coaching tool rather than a disciplinary scoreboard.¹²

3. Pair reject data with exposure index trends

Review the exposure index and deviation index (per IEC 62494-1) alongside rejects.⁸ This separates positioning failures from exposure failures and exposes dose creep — a slow upward drift in exposures that does not generate rejects but does raise dose. See our exposure index guide for how deviation index is interpreted.

4. Exclude or segregate QA and test exposures

Phantom images, calibration exposures, and deliberate test shots should not inflate the clinical reject rate. Tag them so they can be removed from the denominator.¹

5. Close the loop

Data without action is binder filler. Each review cycle should produce specific corrective actions — targeted positioning in-services, collimation refreshers, AEC checks, or protocol adjustments — and the next cycle should test whether the targeted reject category fell.¹²

6. Watch for under-reporting

A reject rate that collapses toward zero is usually a reporting artifact, not a quality triumph. Spot-check by comparing logged rejects against exam volumes and by auditing whether reshoots are being recorded.¹

Common pitfalls

Reporting one facility-wide number and missing a high-reject projection hidden inside it.
Comparing raw rates across vendors without harmonizing reason categories.
Chasing a low number until technologists stop recording rejects.
Ignoring dose creep because it produces no rejects.
Never closing the loop — collecting data that never changes practice.

Regulatory Considerations

Repeat-reject analysis sits primarily in the X-ray (FDA and state) regulatory world, reinforced by accreditation requirements, rather than in NRC byproduct-material rules. Radiographic units are radiation-producing machines regulated by the U.S. Food and Drug Administration and by state radiation-control programs, so reject-analysis expectations come through state QA regulations, accreditation bodies, and professional standards rather than through 10 CFR.⁶¹⁰

Key reference points:

ACR–AAPM Technical Standard for Diagnostic Medical Physics Performance Monitoring of Radiographic Equipment — establishes that radiographic equipment performance is evaluated at installation and monitored at least annually by a qualified medical physicist, the natural home for reject-rate review.⁶
ACR–AAPM Practice Parameter for Digital Radiography — addresses image quality, exposure indices, and the quality-control program expectations for DR systems.⁷
AAPM Task Group Report 305 — the consensus methodology for vendor-neutral reject analysis.¹
IEC 62494-1 — defines the exposure index and deviation index used to interpret exposure adequacy alongside reject data.⁸
NCRP Report No. 172 — reference levels and achievable doses that contextualize the dose saved by reducing repeats.¹⁰⁹
FDA's initiative to reduce unnecessary radiation exposure from medical imaging — the broader policy backdrop for minimizing avoidable dose.¹⁰

For comparison, mammography is the one modality where repeat/reject analysis is explicitly federally mandated under the Mammography Quality Standards Act program; see our guide to mammography quality control under MQSA. General radiography is held to recommended-practice and accreditation standards rather than a single federal number. Across DRPS service areas, X-ray machine QA is administered by state radiation-control programs (for example, Florida under its radiation-control rules), with Washington DC and Delaware being direct-NRC jurisdictions only for byproduct material, not for X-ray machines. Always confirm the specific QA and record-retention requirements with the authority having jurisdiction. For accreditation expectations, see our ACR accreditation physics requirements overview.

Frequently Asked Questions (FAQs)

What is repeat-reject analysis in radiography?

Repeat-reject analysis is the systematic collection and review of rejected radiographic images — exposures that delivered patient dose but were not sent to the radiologist for diagnosis. It tracks how often images are rejected, why, and where (by projection, room, or technologist), so a department can target training and equipment fixes that reduce repeat exposures and unnecessary dose.

What is a good reject rate for digital radiography?

There is no single universal target, but published multi-center digital radiography data commonly cluster around a median reject rate near 8–10%, and reject reference levels are often set near the 75th percentile of local data. A much higher rate may indicate workflow or training problems; a near-zero rate usually indicates that rejects are not being recorded honestly.

What is the difference between a repeat and a reject?

A reject is any acquired image that is discarded and not used for diagnosis. A repeat is the additional exposure made to replace a rejected image. Because most rejected images lead to a repeat, reject analysis is used as a practical surrogate for the extra patient dose caused by retakes.

Why is patient positioning the most common reject reason?

In digital radiography, post-processing can rescue many exposure errors that would have ruined a film, so the dominant remaining failure mode is geometry — positioning, anatomy cut-off, and motion. Multi-center studies consistently report patient positioning as the single largest category of rejected images, often around three-quarters of all rejects.

Does AAPM TG-305 set a required reject rate?

No. TG-305 does not mandate a numerical reject-rate limit. Its purpose is to standardize the data elements and reject-reason categories so reject data can be compared across vendors and sites, and to recommend monitoring workflows. Specific action thresholds are set locally by the facility and its qualified medical physicist.

How does the exposure index relate to reject analysis?

The exposure index and deviation index (defined under IEC 62494-1) quantify whether the detector received an appropriate amount of radiation. Reviewing exposure-index trends alongside reject data separates true exposure errors from positioning errors and reveals dose creep.

How often should a facility review reject data?

Reject data should be collected continuously and reviewed regularly — commonly monthly or quarterly — with at least an annual summary reviewed by the qualified medical physicist as part of equipment performance monitoring. The review should drive corrective action, not just produce a number.

Key Takeaways

Every rejected image is patient dose with no diagnostic return — and the repeat that replaces it adds more. Reject analysis makes that invisible dose visible.¹⁴
Standardize before you measure. TG-305 harmonizes reject-reason categories and data elements so rates are comparable across vendors and rooms.¹
One facility-wide number is not enough. Analyze absolute and relative reject rates by projection, room, and technologist.¹²
Benchmarks cluster in the high single digits to low teens for digital radiography, with positioning the dominant reason — but the right local target is set with your medical physicist, not copied from a paper.²³
Pair rejects with exposure-index trends to catch dose creep that produces no rejects.⁸
Close the loop. Data must drive training, protocol, and equipment action — and the next cycle must verify the fix.

Conclusion

Repeat-reject analysis is one of the highest-yield, lowest-cost quality tools in diagnostic radiography. It converts software logs that are easy to ignore into a concrete map of where patient dose is being wasted and where image quality is failing. The modern program, framed by AAPM TG-305, standardizes how rejects are recorded, analyzes rates by projection and operator, contextualizes the result against benchmarks and dose, and — most importantly — closes the loop with targeted corrective action.¹²

A reject rate is not a grade to be minimized at all costs; it is a diagnostic for the imaging operation itself. Read it honestly, break it down, act on it, and re-measure. That is what turns a number in an accreditation binder into real dose reduction for patients.

How DRPS Can Help

Diagnostic Radiation Physics Services helps imaging facilities build and audit repeat-reject programs that stand up to inspection and actually reduce dose: standardizing reject-reason categories across mixed-vendor fleets, setting defensible local reject reference levels, integrating exposure-index monitoring, and translating the data into technologist training and protocol fixes. This work is delivered through our diagnostic radiography physics testing, accreditation support, and medical physicist consulting services.

DRPS supports facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware.

A strong reject-analysis program is not about hitting a magic percentage. It is about making the lowest-dose, highest-quality image the routine result of everyday practice.

Related Resources

References

Little K, Reiser I, Apgar B, et al. AAPM task group report 305: Guidance for standardization of vendor-neutral reject analysis in radiography. J Appl Clin Med Phys. 2023;24(5):e13938. doi:10.1002/acm2.13938. aapm.onlinelibrary.wiley.com
Serra D, Neep MJ, Ryan E. Multi-centre digital radiography reject analysis for different clinical room use types: the establishment of local reject reference levels for public hospital departments. J Med Radiat Sci. 2024;71(3):412-420. doi:10.1002/jmrs.796. doi.org
Stephenson-Smith B, Neep MJ, Rowntree P. Digital radiography reject analysis of examinations with multiple rejects: an Australian emergency imaging department clinical audit. J Med Radiat Sci. 2021;68(3):245-252. doi:10.1002/jmrs.468. doi.org
Peer S, Peer R, Giacomuzzi SM, Jaschke W. Comparative reject analysis in conventional film-screen and digital storage phosphor radiography. Radiat Prot Dosimetry. 2001;94(1-2):69-71. doi:10.1093/oxfordjournals.rpd.a006482. doi.org
Rose S, Viggiano B, Bour R, Bartels C, Szczykutowicz T. A multiinstitutional study on wasted CT scans for over 60,000 patients. AJR Am J Roentgenol. 2020;215(5):1123-1129. doi:10.2214/AJR.19.22604. doi.org
American College of Radiology, American Association of Physicists in Medicine. ACR–AAPM Technical Standard for Diagnostic Medical Physics Performance Monitoring of Radiographic Equipment. acr.org
American College of Radiology, American Association of Physicists in Medicine, Society for Imaging Informatics in Medicine. ACR–AAPM–SIIM Practice Parameter for Digital Radiography. acr.org
International Electrotechnical Commission. IEC 62494-1: Medical electrical equipment — Exposure index of digital X-ray imaging systems — Part 1: Definitions and requirements for general radiography. iec.ch
National Council on Radiation Protection and Measurements. Reference Levels and Achievable Doses in Medical and Dental Imaging: Recommendations for the United States. NCRP Report No. 172. Bethesda, MD: NCRP; 2012. ncrponline.org
U.S. Food and Drug Administration. Initiative to Reduce Unnecessary Radiation Exposure from Medical Imaging. fda.gov