Calibration of radiotherapy ionization chambers using Co‐60 is outdated and should be replaced by direct calibration in linear accelerator beams

At a Glance

Metadata	Details
Publication Date	2015-08-05
Journal	Medical Physics
Authors	Ganesan Ramanathan, M McEwen, Colin G. Orton
Institutions	National Academies of Sciences, Engineering, and Medicine, Australian Radiation Protection and Nuclear Safety Agency
Citations	3
Analysis	Full AI Review Included

6CCVD Technical Documentation: Advanced Diamond Dosimetry for High-Energy Radiotherapy Beams

Executive Summary

This research analysis highlights the critical debate regarding reference standard traceability in medical radiation dosimetry, specifically the proposal to replace traditional Cobalt-60 (Co-60) calibration with direct calibration using medical linear accelerators (Linacs). The shift directly impacts the required performance and stability of secondary standards, creating an urgent need for advanced, high-resolution detectors—a key application area for 6CCVD’s diamond material solutions.

Core Challenge: Replacing the Co-60 reference standard (simple, stable, low cost) with Linac calibration (more complex, higher cost, but clinically relevant).
Precision Target: The primary driver is reducing reference dosimetry uncertainty. Traditional methods carry an estimated uncertainty of ~1.1%, while direct Linac calibration aims for uncertainties as low as 0.6%-0.7% (and potentially 0.15% standard deviation for k_Q factors).
Detector Requirement: The push toward Linac calibration, combined with the development of new small field dosimetry protocols, necessitates the use of novel detectors, specifically diamond and diode detectors, to handle high-energy photon and electron beams.
Stability Trade-Off: Linac beams are less stable long-term than Co-60. Reference-class secondary detectors, such as high-purity Single Crystal Diamond (SCD), must provide extreme intrinsic stability to serve as reliable QA tools in this environment.
6CCVD Solution: SCD offers unparalleled spatial resolution and stability, making it the ideal material for high-accuracy absorbed dose measurements, Monte Carlo code benchmarking, and complex small field dosimetry applications where conventional ionization chambers are inadequate.

Technical Specifications

The following parameters represent the current state-of-the-art and target goals for absorbed dose traceability and uncertainty discussed in the debate.

Parameter	Value	Unit	Context
Combined Uncertainty (Co-60/TRS-398)	~1.1	%	Estimated uncertainty using k_Q correction factor
Combined Uncertainty (Direct Linac Calibration)	0.6 - 0.7	%	Uncertainty achieved by direct calibration (e.g., ARPANSA)
k_Q Factor Relative Uncertainty (Measured)	0.3	%	Achieved for most commonly used ionization chambers
k_Q Factor Relative Uncertainty (Monte Carlo)	<0.5	%	Theoretical limit, constrained by W/e and stopping power values
Chamber-to-Chamber Variation (Reference Class k_Q)	0.15	%	Standard deviation observed in MV calibration data (NRC Canada)
Required k_Q Calibration Tolerance	Better than 0.2	%	Suggested uncertainty for application of generic calibration curves
Recommended Detector Development	Diamond / Diode	N/A	Essential for investigation of novel applications and high-energy beams
Primary Reference Standard Quantity	Absorbed Dose to Water	N_D,W,Q	Calibration target for high-energy radiotherapy beams

Key Methodologies

The debate centers on two primary methods of disseminating absorbed dose standards for radiotherapy reference ionization chambers. Advanced dosimeter development relies heavily on the capabilities of modern Linacs.

Traditional Dosimetry Chain (Co-60 Based): Calibration coefficient determined in a stable, well-characterized Co-60 reference beam, subsequently adjusted for clinical use via a calculated beam quality conversion factor (k_Q).
Direct Linac Calibration: Direct measurement of the absorbed dose to water calibration coefficient (N_D,W,Q) in a megavoltage (MV) photon beam from a linear accelerator, avoiding reliance on calculated k_Q factors or Co-60 reference data.
Monte Carlo Benchmarking: Use of high-accuracy experimental benchmarks (often requiring novel detectors like diamond) to test Monte Carlo radiation transport codes, which are essential for calculating reliable k_Q factors and commissioning Linacs.
Small and Nonstandard Field Protocol: Implementation of the new IAEA/AAPM working group formalism requiring direct calibration in machine-specific reference fields, demanding detectors with extremely high spatial resolution and negligible volume averaging effects.
Long-Term Quality Assurance (QA): Monitoring the stability of reference-class secondary dosimeters to ensure traceability over time, a challenge amplified by the inherent output variability of Linacs compared to Co-60 sources.

6CCVD Solutions & Capabilities

The need for highly stable, high spatial resolution detectors capable of working across high-energy photon and electron beams (MV range) and in nonstandard, small fields directly aligns with 6CCVD’s expertise in customized MPCVD diamond synthesis.

Applicable Materials

To replicate or extend this research, particularly in the realm of novel dosimeters, 6CCVD recommends:

Optical Grade Single Crystal Diamond (SCD): Required for reference-class detectors demanding the highest purity, stability, and radiation hardness. SCD minimizes noise and maximizes charge collection efficiency, crucial for achieving sub-0.5% uncertainty targets and ensuring long-term QA stability, compensating for Linac instability.
Heavy Boron-Doped Diamond (BDD): Suitable for applications requiring tunable conductivity or resistive thermal sensors within the dosimeter stack, providing versatility for internal calibration or temperature monitoring.
PCD Wafers (Inch-Size): Available up to 125mm, Polycrystalline Diamond (PCD) substrates can be used for large-area detectors or thermal management components within complex reference standards.

Customization Potential

Diamond detectors used for reference dosimetry require extremely precise geometry and clean interfaces to minimize perturbation effects, especially for small field measurements.

6CCVD Service	Relevance to Radiotherapy Dosimetry	Technical Capability
Custom Dimensions	Producing detector elements for small field and machine-specific reference fields where size is critical (down to µm resolution).	Plates/wafers up to 125mm; Thickness 0.1µm - 500µm.
High-Precision Polishing	Achieving near-perfect surface finish (Ra < 1nm) for uniform electrical contacts and minimizing surface defects that contribute to leakage current.	Ra < 1nm (SCD); Ra < 5nm (Inch-size PCD).
Metalization Services	Essential for creating electrode contacts (Ohmic or Schottky barriers) necessary for current collection in diamond detectors.	In-house application of Au, Pt, Pd, Ti, W, Cu—standard metallization used in high-energy physics and medical sensing.
Custom Shaping	Providing laser cutting and precision dicing for creating custom geometries required for advanced beam profiling and small-field array dosimeters.	Precise shaping for minimized volume averaging effects.

Engineering Support

The transition from Co-60 to Linac calibration protocols (IAEA TRS-398, AAPM TG-51 addendums) requires specialized knowledge in material science, radiation transport, and detector physics. 6CCVD’s in-house PhD team provides comprehensive support:

Material Selection: Guidance on choosing the optimal diamond material (SCD vs. PCD, doping level) to match the required spatial resolution and energy response for specific MV photon and electron beams.
Detector Design Consultation: Assistance with electrode design and metalization schemes to achieve optimal charge collection efficiency (required for minimizing measurement uncertainty) in high-flux Linac environments.
Global Supply Chain: Reliable, DDU default (DDP available) global shipping ensures rapid material delivery to standards laboratories (like ARPANSA or NRC) and manufacturing partners worldwide.

For custom specifications or material consultation, visit 6ccvd.com or contact our engineering team directly.

View Original Abstract

Most medical linear accelerators worldwide are calibrated using ionization chambers that are themselves calibrated by a standards laboratory, or secondary standards laboratory, in a Co-60 beam. Because these chambers are actually used to calibrate high-energy x-ray beams, it has been suggested that calibration against Co-60 is outdated and should be replaced by calibration in linear accelerator beams. This is the claim debated in this month’s Point/Counterpoint. Arguing for the Proposition is Ramanathan Ganesan, Ph.D. Dr. Ganesan earned his Ph.D. in Physics from the University of Mumbai in 2001, having previously worked for many years as Scientific Officer in the Radiation Standards Section, Radiation Safety Systems Division, BARC, Trombay, Mumbai, India. Subsequently, he spent some time working at the National Physical Laboratory, Teddington, Middlesex, UK, and NIST, Gaithersburg, MD, USA, before he moved to his current position as Senior Radiation Scientist, Radiotherapy Section Medical Radiation Services Branch, Australian Radiation Protection and Nuclear Safety Agency, Yallambie, Victoria, Australia. His major research interests are calorimetry measurements of photon and electron beams, calibration of dosimeters, small field dosimetry, development of diamond and diode detectors, and calibration of environmental radiation dosimeters. Arguing against the Proposition is Malcolm R. McEwen, Ph.D. Dr. McEwen earned his Ph.D. in Radiation Physics from the University of Surrey, UK, in 2002, having previously worked for many years at the Centre for Ionising Radiation Metrology, National Physical Laboratory, UK. He then moved to his current position as Research Officer at the Ionizing Radiation Standards Group, National Research Council of Canada, Ottawa, Canada, where he is Director of the Ottawa Medical Physics Institute and Adjunct Professor within the Department of Physics, Carleton University. Dr. McEwen is Chairman of the Consultative Committee for Ionising Radiation Section I of the Bureau International des Poids et Mesures and has been very active in the AAPM including having served as Chair of the Working Group to review and extend data in the AAPM TG-51 dosimetry protocol for radiotherapy, and he is the current Chair of the AAPM Calibration Laboratory Accreditation Subcommittee. His major research interests are improvements in reference dosimetry for radiation therapy, experimental and theoretical works on the performance and application of secondary dosimeters in high energy photon and electron beams, investigation of novel dosimeters/applications in dosimetry at radiotherapy dose levels, and development of high-accuracy experimental benchmarks for testing Monte Carlo radiation transport codes used, for example, in the commissioning of medical linear accelerators. The success of radiation therapy depends on the accuracy of the prescribed dose delivered. The starting point in the dosimetry chain deciding this accuracy is the dissemination of absorbed dose standards in the calibration of radiotherapy reference ionization chambers. Two methods of dissemination are available: one based on direct calibration of ionization chambers in megavoltage photon beams and the other based on the use of a correction factor, kQ, applied to the calibration coefficient determined in a Co-60 beam. Linear accelerators have totally replaced Co-60 for radiotherapy treatment in many countries and are becoming more common in others. The cost and difficulty in obtaining replacement Co-60 sources caused by increasing security concerns that treat Co-60 therapy sources as high risk, and the need to safely dispose of decayed sources, inhibit their use in hospitals and calibration laboratories.1 Several standards laboratories are equipped with Linacs.2 The experimental measurement of the absorbed dose to water calibration coefficient ND,w,Q and beam quality factor for the user chamber at primary standards dosimetry laboratories is the preferred option in IAEA TRS-398.3 Observed chamber-to-chamber differences, which include the effect of waterproof sleeves (also seen for Co-60), justify the recommendation in IAEA TRS-398 for kQ values specifically measured for the user chamber.4 Also, the new formalism by the IAEA/AAPM working group for reference dosimetry of small and nonstandard fields recommends the direct calibration of the dosimeter in a conventional broad beam and machine-specific-reference fields against a primary standard without the Co-60 calibration.5 The accuracy of a calculated kQ factor depends on the precision of the chamber geometry (including any deviations from the specified geometry). To use a measured kQ only requires that the chamber response be reasonably insensitive to photon energy, so that users can interpolate between beam quality indicies.6 Relative standard uncertainties for the experimental measurements of kQ factors for the most commonly used chambers have been reported as 0.3%.7 Although similar relative standard uncertainties of <0.5% have been achieved recently for Monte Carlo calculated kQ values,8 theoretical calculations are limited by the energy-dependent uncertainties of W/e and stopping power values. This is to be contrasted with the combined uncertainty for the direct calibration method such as the uncertainties achieved at the Australian Radiation Protection and Nuclear Safety Agency (ARPANSA) that are around 0.6%-0.7%, which are less than the estimated 1.1% with a reference beam quality of Co-60 and the TRS-398 energy correction.10 Furthermore, for the calibration of flattening filter free beams, the published kQ factors reported in IAEA TRS-398 have to be corrected,9 adding to the uncertainties in using Co-60 calibration. After the release of the IAEA TRS-398 and AAPM TG-51 protocols in 2000, several new ion chambers have been introduced in the market. Direct calibration may be used for these new or rare chamber types for which calculated kQ factors are not available in IAEA TRS-398 or the addendum to AAPM TG-51.11 Calibration of radiotherapy ionization chambers using Co-60 is outdated for the above mentioned reasons and should be replaced by direct calibration in linear accelerator beams. At the heart of this proposition would appear to be the fact that for reference dosimetry of linear accelerator beams, the majority of clinical medical physicists worldwide must use a chamber calibrated in a Co-60 beam together with calculated kQ factors. These factors, and the method to apply them, are provided by protocols such as AAPM TG-51 or IAEA TRS-398.11-13 The reason to change from this approach is, presumably, that these calculated factors are used “on faith” and could lead to significant dosimetric errors when combined with any particular ionization chamber. Fifteen years after the IAEA TRS-398 recommendation that users obtain calibrations in linear accelerator beams, the literature is surprisingly silent on the need to do so. Only one country in the world currently requires calibrations in linear accelerator beams. The National Physical Laboratory in the UK has carried out Linac calibration of ion chambers since 1989, and their own data14 show no significant variation in chamber kQ factors, to the extent that one can apply a generic calibration curve with an uncertainty better than 0.2%. Muir et al.15 compared Monte Carlo kQ factors to measurements and found very good agreement (0.25% or better) for a wide range of chamber types over the full range of Linac photon energies. Andreo et al.4 compared the older TRS-398 calculations to the same experimental data set and concluded that no revision of those semianalytical kQ factors was required. The National Research Council in Canada has been offering MV calibrations since 2007 and analysis of these data showed that, although one sees up to 1% variations in Co-60 ND,w coefficients, the standard deviation of kQ factors for reference-class ionization chambers was only around 0.15%.16 One can therefore reasonably ask, “What problem needs to be solved?” On a more pragmatic note, although it is perhaps easy for the clinical medical physicist to view Co-60 as “outdated,” I would argue that Co-60 is the ideal calibration beam. A Co-60 irradiation unit is simple and very reliable to use, has a very predictable output over multiple years, and is much cheaper to operate than a linear accelerator. Leaving the economics of the calibration laboratory aside (although in some areas of the world this is a very important consideration), moving from a Co-60 irradiator to a linear accelerator for calibration would immediately result in a loss of precision and a loss in the ability to monitor the long-term stability of reference-class ionization chambers. Linear accelerator beams are simply not stable enough to provide that long-term reference field. Since the accuracy gain one might achieve by moving to Linac calibrations is, based on the literature, only a few tenths of one percent, this loss in baseline QA of the detectors results in a negative cost-benefit analysis. Linear accelerators may be the obvious choice for absorbed dose delivery but Co-60 remains the best choice for absorbed dose calibration. Dr. McEwen has raised several important points regarding the calibration of radiotherapy ionization chambers. However, his comment that “Fifteen years after the IAEA TRS-398 recommendation that users obtain calibrations in linear accelerator beams, the literature is surprisingly silent on the need to do so” is not entirely correct. As I mentioned in my opening statement, a number of standards laboratories are installing Linacs, establishing primary standards, and measuring absorbed dose to water, the prime quantity needed for calibration, validated through international intercomparisons (e.g., BIPM.RI(I)-K6).17 Also, there are a number of publications on experimental measurements of kQ factors at megavoltage energies for several ionization chambers.18 Regarding his point that calibration data from the National Physical Laboratory in the UK show no significant variations in chamber kQ factors, the chambers (NE 2561/NE 2611) were designed in-house and built specifically as secondary standards. Andreo et al.4 observed significant chamber-to-chamber variations in Co-60 beams in a study of 91 NE 2571 chambers, which are the industry standard. The observation by Muir et al.,15 who compared Monte Carlo kQ factors to measurements and found very good agreement (0.25% or better), is valid only if possible variations of W/e with energy are ignored and assuming correlated uncertainties in photon cross sections. Larger deviations (∼0.5%) between measured and theoretical kQ factors occur at higher energies. Dr. McEwen argues that replacement of Co-60 with linear accelerators for calibration would immediately result in a loss of precision and a loss in the ability to monitor the long-term stability of reference-class ionization chambers. This argument is contradictory to his claim that the MV calibration data from the National Research Council in Canada since 2007 have shown the standard deviation of kQ factors for reference-class ionization chambers to be around 0.15%. The rising cost of Co-60 re-sourcing is indeed a concern for many calibration laboratories but must be considered in the context of the price tag for a linear accelerator, which is the proposed alternative. If you take the optimistic assumption of a 20-yr Linac lifetime, you can do the math and conclude that you do not come close to the same capital costs for a Co-60 irradiator over that time period, even if you re-source every half-life. And that is before accounting for maintenance costs, which are significantly more for a Linac compared to a Co-60 unit. Economics are, therefore, not a driver for a change in the calibration basis. The chamber-to-chamber variations cited are also illusory. The much larger data set from the US ADCLs analyzed by Muir19 shows very tight tolerances on ND,w coefficients, and I would reiterate that there are no data in the literature that suggest a significant chamber-to-chamber variation in kQ for any particular chamber type. Direct calibrations in MV beams do indeed result in a potentially lower uncertainty in clinical reference dosimetry, although one could argue whether the improvement is significant. However, there is one issue where MV calibrations offer a clear advantage, and I am surprised it was not brought up by my opponent. A calibration in a MV photon beam is also a precise test of the user’s chamber in a beam very similar to that in which it will be used. It answers that “What if my chamber is atypical?” question and is a QA test clearly missing from the present Co-60-based approach. Whether that, alone, is enough to warrant a wholesale change in calibration practices is for the user community to decide.

Tech Support

Original Source

DOI: https://doi.org/10.1118/1.4922710

References

2011 - Proceedings of an International Symposium on Standards, Applications and Quality Assurance in Medical Radiation Dosimetry (IDOS), Vienna