Dose Calibrator QC: Accuracy, Linearity, Constancy, Geometry

Introduction

Dose calibrator quality control is the structured program of four tests—constancy, accuracy, linearity, and geometry—that proves a nuclear medicine clinic can measure the activity of every patient dosage correctly before it is administered. Each test targets a different failure mode, runs on its own schedule, and ties directly to NRC or Agreement State expectations for the medical use of byproduct material. ^{1, 2}

The dose calibrator, more precisely a well-type re-entrant ionization chamber, is the single most safety-critical instrument in a nuclear medicine department. Every diagnostic dosage and every therapy administration is assayed in it. If the calibrator reads high, patients are underdosed and scans are noisy or therapies are under-delivered; if it reads low, patients are overdosed. Because the instrument sits directly in the path of patient care, regulators expect a documented QC program, not occasional spot checks. ^{1, 3}

This guide explains what each of the four tests verifies, how to perform them, how often to run them, what tolerances are typically applied, and how the whole program maps onto NRC 10 CFR 35.60, NRC Regulatory Guide 10.8, and NUREG-1556 Volume 9. It includes a reference table of the four tests and worked math for linearity, accuracy, and constancy. DRPS provides dose calibrator QC review and oversight as part of its PET/CT and nuclear medicine physics and medical physics consulting services across Florida, Maryland, Virginia, Washington DC, California, and Nevada.

Topic Explanation

What is a dose calibrator?

A dose calibrator is a well-type, sealed, pressurized re-entrant ionization chamber that measures the activity of a radioactive sample placed in its central well and displays the result in activity units after applying a radionuclide-specific calibration factor. The radiation emitted by the sample ionizes the fill gas, the resulting current is proportional to activity, and a dial-in or menu-selected calibration setting converts that current to MBq or mCi for the chosen radionuclide. ^{4, 5}

Two features of that design drive the entire QC program:

It is energy- and geometry-dependent. The chamber response per unit activity depends on the photon energy of the radionuclide and on how the sample sits in the well, which is why each radionuclide needs its own calibration factor and why geometry must be characterized. ^{4, 5}
It is a secondary instrument. A clinical dose calibrator is not a primary standard. Its accuracy is established by comparison to a reference source whose activity is traceable to a national metrology institute such as NIST. That traceability chain is what makes the displayed number defensible. ^{4, 9}

For background on the radionuclides being assayed and their emission characteristics, see Understanding Common Isotopes in PET & Radiopharmaceutical Therapy.

Why four separate tests?

No single test proves a dose calibrator is fit for clinical use, because the instrument can fail in four independent ways. It can drift day to day (constancy), be miscalibrated against the true activity (accuracy), respond non-linearly at high or low activity (linearity), or respond differently to different sample volumes and positions (geometry). The four-test framework exists because each test isolates one of these failure modes. ^{2, 3}

The four tests also operate on different timescales. Constancy is fast and cheap, so it runs daily as an early-warning check. Accuracy and linearity are more involved and run annually. Geometry is characterized once at installation and revisited only when something physical about the measurement changes. Together they give continuous, layered assurance without unnecessary daily burden. ^{2, 3}

A facility building or auditing this program should align it with its radiation safety officer procedures and its broader nuclear medicine physics oversight.

Key Technical Principles

The four dose calibrator QC tests at a glance

The table summarizes what each test verifies, how it is done, how often, and a representative tolerance. The specific acceptance limits and frequencies below are representative of common practice and historical NRC guidance; confirm the exact numbers against your radioactive material license, the calibrator manufacturer, and current NRC or Agreement State requirements.

Test	What it verifies	Method	Typical frequency	Representative tolerance
Constancy	Day-to-day reproducibility of the response	Assay a long-lived sealed reference source (e.g., Cs-137 or Co-57) on the same setting daily; compare to the decay-corrected expected reading	Each day of use ^{2, 3}	Within ±5% of the expected value
Accuracy	Agreement with true (NIST-traceable) activity	Assay at least one NIST-traceable reference source (commonly Co-57, Cs-137, Ba-133, or Co-60); compare to the certified, decay-corrected activity	At least annually, and at installation ^{2, 3}	Within ±5% (some programs use ±10%)
Linearity	Correct response across the clinical activity range	Decay method (assay a decaying Tc-99m source over the range) or shield/sleeve method; compare each reading to predicted activity	At least annually, and at installation ^{2, 3}	Within ±5% (some programs use ±10%) across the range
Geometry	Independence from sample volume and position	Assay a fixed activity in varying volumes and/or container positions; derive correction factors if response varies	At installation and after repair/relocation ^{2, 3}	Variation within ±5% or apply correction factors when it exceeds the limit

Constancy: the daily early-warning check

Constancy verifies that the calibrator gives the same reading today that it gave yesterday for the same unchanging source. A long-lived sealed reference source (often Cs-137, which has a 30.1-year half-life, or Co-57) is assayed each day of use, typically on the most-used clinical settings, and the reading is compared to the value expected after decay correction from a baseline. ^{2, 3}

Because the source half-life is long, the day-to-day decay correction is tiny, so any meaningful change in the reading points to an instrument problem—electronic drift, a chamber fault, contamination, or a setting error—rather than to source decay. Constancy is therefore the program's tripwire: cheap enough to do daily, sensitive enough to catch a drifting instrument before it affects patients.

A typical daily constancy evaluation computes the percent deviation between the measured reading $M$ and the decay-corrected expected reading $E$ :

% deviation = \frac{M - E}{E} \times 100%

For example, if a Cs-137 constancy source has a decay-corrected expected reading of $E = 200.0 μ Ci$ on a given day and the calibrator reads $M = 207.0 μ Ci$ :

% deviation = \frac{207.0 - 200.0}{200.0} \times 100% = + 3.5%

A $+ 3.5%$ result is within a typical $\pm 5%$ action level, so the instrument passes that day; a reading drifting toward or past the limit triggers investigation before clinical use.

Accuracy: anchoring to a NIST-traceable source

Accuracy verifies that the calibrator's reading matches the true activity of a reference source whose value is traceable to NIST. At least one NIST-traceable sealed source—commonly Co-57 (≈122 keV), Cs-137 (662 keV), Ba-133 (multiple lines), or Co-60 (1.17 and 1.33 MeV)—is assayed on its appropriate setting, and the measured activity is compared to the certificate value after decay correction to the day of measurement. ^{2, 3}

Using sources that span a range of photon energies tests the calibrator across the energy region relevant to clinical radionuclides, because chamber response varies with energy. The percent deviation is computed the same way as constancy, but here $E$ is the certified, decay-corrected source activity:

% deviation = \frac{A _{measured} - A _{certified}}{A _{certified}} \times 100%

Suppose a Co-57 source was certified at $A_{0} = 250.0 μ Ci$ on its reference date and, after decay correction to today, its expected activity is $A_{certified} = 235.4 μ Ci$ . If the calibrator reads $A_{measured} = 224.0 μ Ci$ :

% deviation = \frac{224.0 - 235.4}{235.4} \times 100% = - 4.8%

A $- 4.8%$ result is just inside a $\pm 5%$ tolerance but outside a tighter goal, so it would typically pass while prompting a closer look or a service check before it drifts further. A result beyond the tolerance requires corrective action and, depending on the program, a hold on clinical use of the affected setting.

Linearity by the decay method

Linearity verifies that the calibrator reads correctly across the full range of activities used clinically—from a high therapy or stock activity down to a small residual dosage. The decay method exploits the fact that radioactive decay is exactly exponential and well characterized, so a single source measured repeatedly as it decays provides a built-in "known" reference at many activity levels. ^{2, 3, 6}

A high-activity Tc-99m source (Tc-99m is convenient because its 6.0-hour half-life lets it span the clinical range in about two to three days) is assayed at the start and then at intervals as it decays. Each measured value is compared to the activity predicted by exponential decay from the first reading:

A (t) = A_{0} e^{- λ t}, λ = \frac{ln 2}{T _{1/2}}

For Tc-99m with $T_{1/2} = 6.0 h$ :

λ = \frac{ln 2}{6.0 h} = \frac{0.6931}{6.0 h} = 0.1155 h^{- 1}

Take an initial assayed activity of $A_{0} = 200.0 mCi$ at $t = 0$ . The predicted activity at later times is:

A (24 h) = 200.0 e^{- (0.1155) (24)} = 200.0 e^{- 2.772} = 200.0 \times 0.0625 = 12.5 mCi

A (48 h) = 200.0 e^{- (0.1155) (48)} = 200.0 e^{- 5.544} = 200.0 \times 0.00391 = 0.781 mCi

(The 24-hour drop by a factor of 16 makes sense: 24 h is exactly four Tc-99m half-lives, and $2^{4} = 16$ .) Suppose the calibrator actually reads $13.1 mCi$ at $t = 24 h$ . The linearity deviation at that point is:

% deviation = \frac{13.1 - 12.5}{12.5} \times 100% = + 4.8%

A $+ 4.8%$ deviation is within a $\pm 5%$ acceptance window but worth watching; a systematic trend of growing deviation at low activities is a classic sign of non-linearity near the bottom of the range. The decay method is repeated until the source has decayed across the entire clinical range, and the worst-case deviation governs the pass/fail decision.

When a faster turnaround is needed, the shield (sleeve) method uses a calibrated set of graded lead sleeves to simulate decay on a single non-decaying or slowly decaying source. The sleeve set must itself be calibrated against the decay method, but once validated it lets a clinic verify linearity in a single session. ^{2, 6}

Geometry: volume and position dependence

Geometry verifies that the calibrator reads the same activity regardless of the sample's volume, container, and vertical position in the well. Because the chamber responds to a slightly different fraction of emitted radiation depending on where the activity sits and how much self-absorption occurs in the sample, a fixed activity can read differently as the fill volume or container changes—an effect that matters most for low-energy emitters. ^{2, 4, 7}

The test holds activity fixed (for example, by adding the same activity to increasing water volumes, or by measuring the same syringe at different depths) and records how the reading changes. If the variation exceeds the action level, the clinic either standardizes containers and fill volumes or applies geometry correction factors so that, for instance, a 1 mL syringe and a 10 mL vial of the same activity report the same number. Precision activity-measurement work at metrology institutes shows how strongly vial choice, fill volume, and source position influence well-chamber response, which is exactly why this characterization is required before clinical reliance on the instrument. ⁷

Clinical Impact

Patient dosing accuracy

Every value a dose calibrator displays becomes a clinical decision. A diagnostic Tc-99m or F-18 dosage assayed on a calibrator that reads 8% low means the patient is administered roughly 8% more activity than intended—an avoidable dose to the patient—while a high-reading calibrator under-administers and can degrade image quality or therapy delivery. For therapy radionuclides such as I-131 or Lu-177, calibrator error translates directly into under- or over-treatment relative to the written directive. ^{1, 3}

A study appraising two clinical radionuclide calibrators before Lu-177 internal dosimetry found that constancy, accuracy, linearity, and sample-volume (geometry) performance all fell within ±5%, which let the authors rely on the instruments for quantitative dosimetry. That is the practical payoff of a disciplined four-test program: it converts a displayed number into a defensible measurement. ⁸

Quantitative imaging and dosimetry

Calibrator QC is foundational to quantitative nuclear medicine. SUV in PET, organ activity in SPECT-based dosimetry, and absorbed-dose estimates in radiopharmaceutical therapy all start from an assayed administered activity. If the calibrator is biased, that bias propagates into every downstream quantitative result, no matter how well the scanner is calibrated. For programs doing theranostics or internal dosimetry, calibrator performance and scanner calibration are a linked chain. ⁸

This is why dose calibrator QC should be reviewed alongside imaging-equipment QC and, for therapy programs, dosimetry workflows such as those discussed in Lu-177 theranostics dosimetry.

Catching contamination and instrument failure

The QC program also detects problems unrelated to calibration. A sudden shift in the daily constancy reading can reveal chamber contamination, a liner leak, electronic failure, or a wrong setting before any patient is affected. Because constancy runs daily, it is often the first place a developing instrument fault shows up, which is precisely why regulators emphasize the daily check. ^{2, 3}

Practical Optimization Tips

A reliable dose calibrator QC program follows a consistent operational rhythm.

1. Standardize the daily constancy routine

Use a dedicated long-lived constancy source (commonly Cs-137 or Co-57) and the same source-holder geometry every day.
Check constancy on every clinically used setting, not just one.
Record the reading, the expected decay-corrected value, and the percent deviation against a defined action level.
Trend the results so slow drift is visible before it crosses the limit.

2. Treat accuracy and linearity as scheduled annual events

Use NIST-traceable sources that span the relevant photon-energy range for accuracy.
For linearity, choose a top activity at or above the highest clinical activity and follow the decay down to at or below the lowest dosage you assay.
If you use the sleeve (shield) method, validate the sleeve set against the decay method before relying on it.

3. Characterize geometry once—then control it

Determine correction factors for the containers and fill volumes you actually use.
Standardize syringes, vials, and fill volumes so corrections stay small and consistent.
Re-evaluate geometry after any repair, relocation, or change in routine containers.

4. Define action levels and corrective action in writing

Document the pass/fail tolerance for each test and what happens on a failure (repeat, service, remove from clinical use).
Keep records retrievable for inspection per your license retention requirements.

Common pitfalls to avoid

Running constancy on only one setting. A drift can affect one radionuclide setting and not another; check the settings you actually use.
Skipping decay correction. Comparing today's reading to an uncorrected baseline produces a false deviation and erodes trust in the check.
Ignoring the low end of linearity. Non-linearity is most common at very low activities—exactly where residual-dose and pediatric measurements live.
Forgetting geometry after a change. Switching syringe brands or vial sizes can shift the reading; re-verify rather than assume.
No defined action level. A QC number with no pass/fail criterion is data, not a control. Define the limit and the response in advance.
Poor record retention. Undocumented QC is, for inspection purposes, equivalent to no QC.

Regulatory Considerations

Dose calibrator QC sits squarely inside the NRC medical-use framework, and the program must be defensible against the facility's radioactive material license. The core requirement is that a licensee possess and use instrumentation to measure the activity of unsealed byproduct material before it is administered to patients, and calibrate that instrumentation appropriately. ¹

Key frameworks to reference:

10 CFR 35.60 — Possession, use, and calibration of instruments used to measure the activity of unsealed byproduct material. It requires the licensee to possess and use such instrumentation, to measure dosages of photon-emitting radionuclides before administration, to calibrate the instrument in accordance with nationally recognized standards or the manufacturer's instructions, and to retain the records. ¹
NRC Regulatory Guide 10.8 — the historical source of the four-test framework (constancy, accuracy, linearity, geometry) and the methods and tolerances widely adopted in practice. A 1996 Health Physics analysis of the NRC's amended dose-calibrator rule documents the regulatory history, including the raising of the lower linearity-test activity limit to 1.1 MBq (30 µCi) to align with NRC program requirements. ^{2, 3}
10 CFR Part 35 — Medical Use of Byproduct Material — the broader rule governing authorized use, written directives, dosage determination, medical-event reporting, and the RSO's responsibilities. A calibrator that produces an administered activity outside the prescribed range can implicate written-directive and medical-event provisions. ¹⁰
NRC NUREG-1556, Volume 9 — program-specific licensing guidance for medical use, including expectations for instrument calibration, QC procedures, and recordkeeping that inspectors use as a benchmark. ¹¹

Agreement States administer their own equivalent programs. Of the states DRPS serves, Florida, Maryland, Virginia, California, and Nevada are NRC Agreement States that license medical use of byproduct material under their own radiation-control rules, while Washington, DC is regulated directly by the NRC. A facility must verify which authority issues its license and which calibration, test-frequency, tolerance, and recordkeeping requirements apply before relying on any QC assumption. ^{10, 11}

For broader licensing context and common inspection findings, see the NRC radioactive material license guide, and connect the QC program to documented procedures with medical physics consulting and RSO support.

Frequently Asked Questions (FAQs)

What are the four dose calibrator QC tests?

The four classic dose calibrator quality control tests are constancy, accuracy, linearity, and geometry. Constancy checks day-to-day reproducibility, accuracy compares the reading to a NIST-traceable source, linearity verifies the response across the clinical activity range, and geometry confirms the reading does not change with sample volume or position.

How often does a dose calibrator need QC?

Constancy is typically performed each day of use, accuracy and linearity at least annually, and geometry at installation and after any repair or relocation that could affect the chamber. Follow the schedule in your radioactive material license, manufacturer guidance, and applicable NRC or Agreement State requirements.

What does NRC 10 CFR 35.60 require for dose calibrators?

10 CFR 35.60 requires that a licensee possess and use instrumentation to measure the activity of unsealed byproduct material before administration, calibrate that instrumentation per nationally recognized standards or the manufacturer's instructions, and retain the records. The historical four-test framework comes from NRC Regulatory Guide 10.8.

What is an acceptable dose calibrator accuracy tolerance?

A commonly used acceptance limit is that the measured activity of a NIST-traceable reference source agrees with the certified, decay-corrected value within about ±5% (some programs and standards use ±10%). Confirm the exact tolerance and action level against your license, the calibrator manufacturer, and the applicable standard.

How is dose calibrator linearity tested by the decay method?

A high-activity source, usually Tc-99m, is assayed repeatedly as it decays over the clinical activity range, often a couple of days. Each measured value is compared to the activity predicted by exponential decay from the first reading. Deviations are typically required to stay within about ±5% (or ±10%) across the range.

Why does dose calibrator geometry matter?

A well-type ionization chamber can read differently for the same activity depending on sample volume, container, and vertical position because of variable self-absorption and detection geometry. The geometry test characterizes these effects so the clinic can apply correction factors or standardize containers and fill volumes.

Can a dose calibrator failure be a reportable event?

A dose calibrator that is out of tolerance can lead to administered activities that differ from the prescribed dosage, which may trigger written-directive, medical-event, or reporting requirements under 10 CFR Part 35. Prompt QC, action levels, and corrective action help prevent dosing errors and support compliance.

Key Takeaways

The four tests each catch a different failure. Constancy catches drift, accuracy catches miscalibration against truth, linearity catches range-dependent error, and geometry catches volume- and position-dependence; you need all four.
Frequencies are layered by cost and risk. Constancy is daily, accuracy and linearity are at least annual, and geometry is at installation and after physical changes—confirmed against the facility license.
Accuracy depends on NIST traceability. A clinical calibrator is a secondary instrument; its defensibility comes from comparison to certified, decay-corrected reference sources (Co-57, Cs-137, Ba-133, Co-60).
Linearity uses exact physics. The decay method leverages $A (t) = A_{0} e^{- λ t}$ so a single decaying Tc-99m source provides known reference activities across the whole clinical range.
Geometry is characterized once, then controlled. Standardizing containers and fill volumes—or applying correction factors—keeps the reading independent of sample geometry.
QC is a compliance and patient-safety control. It maps to 10 CFR 35.60, Reg Guide 10.8, and NUREG-1556 Volume 9, and a documented program with defined action levels protects patients and supports inspection readiness.

Conclusion

Dose calibrator quality control is not a box-checking ritual—it is the measurement-assurance backbone of a nuclear medicine program. The four tests work together: constancy gives a daily tripwire, accuracy anchors the instrument to NIST-traceable truth, linearity proves the response across every activity the clinic assays, and geometry ensures the answer does not depend on how the sample is presented to the chamber.

The RSO and medical physicist should treat the program as a connected system, with defined sources, defined tolerances, defined action levels, and retrievable records aligned to the facility's license. A clinic that runs this program well can trust every number its calibrator displays—and can defend that trust during an inspection. The few minutes spent on a daily constancy check and the annual accuracy and linearity events are a small price for confidence that every patient dosage is right.

How DRPS Can Help

Diagnostic Radiation Physics Services helps nuclear medicine clinics build, run, and document defensible dose calibrator QC programs. This may include establishing daily constancy and annual accuracy and linearity procedures, characterizing geometry corrections, selecting NIST-traceable reference sources, defining action levels and corrective-action workflows, and reviewing records for inspection readiness as part of PET/CT and nuclear medicine physics, medical physics consulting, and Radiation Safety Officer support.

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

Strong calibrator QC is not just about passing inspection. It is about making sure every patient dosage is measured right, every time.

Related Resources

References

U.S. Nuclear Regulatory Commission. 10 CFR 35.60: Possession, use, and calibration of instruments used to measure the activity of unsealed byproduct material. ecfr.gov
U.S. Nuclear Regulatory Commission. Regulatory Guide 10.8: Guide for the Preparation of Applications for Medical Use Programs. nrc.gov
Hung JC. A comparison of current regulations and regulatory guide governing quality control of dose calibrators. Health Physics. 1996;71(2):206-210. doi:10.1097/00004032-199608000-00012. PubMed
Bobin C, Thiam C, Chauvenet B. A radionuclide calibrator based on Cherenkov counting for activity measurements of high-energy pure β⁻-emitters. Applied Radiation and Isotopes. 2017;119:60-65. doi:10.1016/j.apradiso.2016.10.018. PubMed
International Atomic Energy Agency. Quality Assurance for Radioactivity Measurement in Nuclear Medicine (Technical Reports Series No. 454). IAEA; 2006. iaea.org
National Council on Radiation Protection and Measurements. NCRP Report No. 58: A Handbook of Radioactivity Measurements Procedures. NCRP. ncrponline.org
Juget F, Nedjadi Y, Buchillier T, Durán T, Bochud F, Kottler C, Bailat C. A portable precision ionization chamber: The transfer ionization reference chamber. Applied Radiation and Isotopes. 2017;134:95-99. doi:10.1016/j.apradiso.2017.10.022. PubMed
Van Wyk BP, Hasford F, Nyakale NE, Vangu MM, Oelofse B, Leboea HM. Critical Appraisal of Radionuclide Calibrators and Gamma Cameras Prior to Lutetium-177 Internal Dosimetry at Two South African Hospitals. World Journal of Nuclear Medicine. 2022;21(1):44-51. doi:10.1055/s-0042-1746173. PubMed
National Institute of Standards and Technology. Radioactivity measurements and standards — traceability for nuclear medicine. NIST. nist.gov
U.S. Nuclear Regulatory Commission. 10 CFR Part 35: Medical Use of Byproduct Material. ecfr.gov
U.S. Nuclear Regulatory Commission. NUREG-1556, Volume 9, Revision 3: Consolidated Guidance About Materials Licenses — Program-Specific Guidance About Medical Use Licenses. nrc.gov
National Nuclear Data Center, Brookhaven National Laboratory. NuDat decay data (Tc-99m, Co-57, Cs-137, Ba-133, Co-60 half-lives and emissions). nndc.bnl.gov