The role of standards laboratories in reducing uncertainty in clinical dosimetry: A Canadian perspective

As a National Metrology Institute, the National Research Council Canada (NRC) provides confidence in measurement results, and data traceable to SI units. This paper outlines some of the ways that the NRC contributes to reducing measurement uncertainty in clinical medical physics dosimetry. These activities include, (i) the improvement of existing primary standards and traceability for established beam modalities, such as MV photon and electron beams; (ii) improving measurement accuracy for new beam modalities through the development of transportable systems which permit operation at the user’s facility; (iii) contributing to new dosimetry protocols, best practice reports and educational outreach; and (iv) supporting the verification of clinical implementation by offering dosimetry auditing capabilities through coordination with the clinical medical physics community.


Introduction
The National Research Council Canada (NRC) was created though an Act of Parliament in 1916 as a Federal applied research and technology organization.A clause in the original act states that the NRC shall act as the National Metrology Institute (NMI) for Canada, and NRC continues to fulfil that role today.In fact, Canada had already become a signatory to the Metre Convention (Convention du Mètre) in 1907 -this is the international treaty that forms the basis for international agreement on units of measurement -and the NRC is responsible for maintaining that international participation.
For some, the NRC is just "that place in Ottawa that calibrates our instruments", and is perhaps also known for the longest-running radio show on the CBC (1 pm every day!) where you can check your wristwatch or other timepiece.However, an NMI plays a much larger role in the application of measurements and measurement science, and in this paper, we will outline some of the ways that, primarily for Canada but ultimately internationally, the NRC contributes to improving accuracy in measurements in medical physics.
Before discussing improvements, it is important to note that the concept of maintenance is essential in the field of measurement standards.A calibration certificate is issued with the underlying assumption that the measurements underpinning the calibration are consistent with those made the previous day/week/year and can be reproduced in the equivalent times in the future.That requires the active maintenance of equipment and procedures that is supported by i) a recognized quality system, and ii) an international framework of comparisons, thus demonstrating competence and equivalence [1].

Improved primary standards with reduced uncertainty
In parallel with the essential maintenance activities noted above, a core activity of an NMI is to improve current primary standards to achieve increased accuracy and reduced uncertainty.For most standards, accuracy is limited by the measurement uncertainty and therefore reducing uncertainty is often the primary goal for established standards.An example of this at the NRC is the recent review carried out by Mainegra-Hing of the Co-60 air kerma standard [2].In that work, initiated by the requirement to implement the recommendations of ICRU Report 90 [3], an analysis was carried out to determine each component that is required to determine air kerma using a cavity ionization chamber, resulting in an improved realization (even though no change was made to the physical standard).

Improved traceability for existing beam modalities
The development of primary standards is not always in phase with user applications, and this is often seen in radiation therapy.Ideally, one would have the necessary measurement infrastructure in place ready for any new treatment technique or beam modality, but it can be challenging to develop measurement standards without access to those radiation beams, and therefore standards development tends to lag clinical practice.Interim methodologies can be developed to provide adequate traceability but the aim should be to develop as direct a measurement standard/method as possible for each application or beam modality.User take-up and the required uncertainty for treatment delivery will impact the timescale for development of a new standard (and, of course, resourcing is always a constraint).In 1999, the AAPM TG-51 dosimetry protocol was published, detailing a procedure to realize absorbed dose to water in clinical photon and electron beams produced by linear accelerators [4].The basis for traceability was an ionization chamber calibrated in a Co-60 beam and calculated conversion factors were required to transfer this calibration to the user beam.This basic formalism was also used by the IAEA TRS-398 Code of Practice, and therefore projects were initiated at a number of NMIs to provide direct traceability for MV photon and electron beams [5].In 2009, the first comparison organized by the BIPM of MV photon beams was carried out at the NRC and currently around ten countries have the capability of providing direct traceability, and reduced uncertainty, for such beams [6].The improvement in uncertainty is significant; The TG-51 route, as detailed in its Addendum, yields a standard uncertainty of around 1%, while the direct calibration route reduces that to less than 0.5% [7].Similar initiatives have been completed, or are ongoing, by NMIs for protons, heavy ions and Ir-192.

Improved traceability for new beam modalities
As a more direct means of establishing traceability in modern beam modalities, new dose to water standards, most often calorimetry-based, are being developed as transportable systems which permit operation at the user's facility.This trend towards developing portable standards purpose-built for clinical use, potentially coordinated within the framework of a wider consortium [8], is likely to become a default approach, as it is unfeasible for NMIs to possess, or even have local access to, every specialized radiotherapy modality brought to market.In the long run, it will be more practical to perform at least some dose calibrations in the clinic using transfer standards, reducing the uncertainty in the beam qualities used by NMIs and those by the end user.Recent examples include water and graphite calorimeters adapted by NMIs for direct use in MR-guided linac [9] and proton beams [10].In this approach, calorimeter measurements can form the basis of a direct dose calibration of an ionization chamber in the clinically-relevant field as a more accurate alternative to determining correction factors for those beams.ICRU Report 78 and 91, which provide prescribing, recording, and reporting guidelines for protons and small photon beams recommend the direct use of calorimetry in the user's beam where possible [11][12].
At the NRC, an easy-to-use, but accurate aluminum calorimeter system has been commissioned, successfully shipped (at the height of the pandemic!)and operated remotely at a BC Cancer Agency radiotherapy clinic.By leveraging a proven design [13], simplifying the data acquisition process, and engineering the system to be lightweight and robust, it was possible to construct a truly mail-able calorimeter that does not require support staff on-site for operation.In the same vein, a bespoke calorimeter was designed at the NRC and was recently used to measure the dose rate of a 2-mm wide 140 keV synchrotron beam at the Canadian Light Source in Saskatoon [14].Preliminary data indicate that the overall uncertainty on dose determination can be significantly improved from the air kerma-based methods currently in use.Perhaps the best current example of coordinated dosimetry research in response to an emerging beam modality is the joint research project, UHDpulse, where NMI collaborators are trialing several new calorimeter systems for direct use in ultra-high dose rate (i.e., FLASH) and laser-driven particle beams [15][16][17].At the NRC, work is underway characterize pyroelectric-based calorimeters capable of FLASH dosimetry that can be read out with only an electrometer [18].

Improved protocols to make use of new understandings
There is a long history of NRC involvement in the development of reference dosimetry protocols.One of the lead authors of the AAPM TG-51, and an advocate for this transition to dose standards [19], was David Rogers, long-time NRC researcher and group leader at the time.Also, many of the key advances that allowed dissemination of dose standards based on water calorimetry were made by NRC scientists [20].
An Addendum to TG-51 for photon beam reference dosimetry was published in 2014, and the lead author, Malcolm McEwen, is based at NRC.The Addendum provides new k Q data, recommendations to improve the accuracy and consistency of the implementation of TG-51 (including the extension to flattening-filter free beams), uncertainties and how the clinical physicist can influence uncertainties, and specifications for a reference-class ionization chamber [8].The new kQ data recommended in the Addendum is from the work of Muir and Rogers [21], and the lead author of that publication, Bryan Muir, is also now based at NRC. Muir is also the lead on an Addendum to TG-51 for electron beam reference dosimetry, which is under review in the AAPM committee structure.The electron beam Addendum provides new data and recommendations based on a review of the literature on electron beam reference dosimetry published in the decades since TG-51 was released.
The AAPM TG-351 on clinical reference dosimetry in MR-guided radiotherapy is extending the guidelines of TG-51 for use in MR-linacs.A draft is in the final stages of preparation.Canadian researchers are leading this effort to address the growing user base for MR-linacs.
NRC is also contributing to AAPM TG-359, on FLASH dosimetry.The goal of this TG is to publish a guidance report on determining the dose and reviewing the need for standardization in dosimetry for UHDR beams to be used in research experiments and in pre-clinical applications.General guidelines will be provided on calibration, dosimetry, operation, and monitoring of beams in FLASHtype modes.

Improved guidance to reduce errors in clinical implementation
Although the TG-51 protocol and its Addendum clearly specify the procedure for reference measurements of external beams, the protocol lacks implementation guidance.Muir is the lead author and Renaud, also based at NRC, is a contributing author on an AAPM report that provides procedural guidance to achieve the most accurate reference dosimetry measurements.Topics include: (i) measurement of depth-ionization curves to obtain beam quality specifiers for the selection of beam quality conversion factors, (ii) considerations for the dosimetry system and specifications of a referenceclass ionization chamber, (iii) commissioning a dosimetry system and frequency of measurements, (iv) positioning the water tank and ionization chamber, (v) requirements for ancillary equipment needed to measure charge and to correct for environmental conditions, and (vi) translation from dose at the reference depth to that at the depth required by the treatment planning system.Commonly-used simplified procedures are described and the impact on the uncertainty quantified to aid the physicist on where to allocate resources.This report is in the final stages of Medical Physics Journal review.
In addition to contributing to protocols and reports, researchers at NMIs provide guidance through educational outreach.There have been several educational sessions presented by NRC researchers that focus on improving clinical understanding of reference dosimetry protocols with the goal of achieving better accuracy and reducing errors in clinical reference measurements [22][23][24][25][26][27][28][29][30].

Development of audit capabilities to verify clinical implementation
Alanine dosimetry capabilities have been established at the NRC since the early 2000s [31][32].The system had been inactive for around 10 years but was re-established for postal dosimetry audits by Mansour in 2018 [33].Alanine dosimeters are used to measure dose to water at therapy dose levels (10 Gy -100 Gy) in Co-60 and MV photon beams with an uncertainty in the calibration of a batch of alanine dosimeters irradiated in Co-60 and in a clinical MV beam of 0.5 % and 0.6 %, respectively.To date, six Canadian cancer clinics have participated in postal dosimetry audits for MV photon beams.The agreement between stated and measured doses were typically within 1 % (k=1).
In addition to TG-51 verifications for conventional linac photon beams, two audits were performed for Canadian clinics with Elekta Unity MR-linacs [34].The results show that combining the uncertainty budget for NRC measurements of alanine dosimeters irradiated in the MR-linac beam in terms of dose to water with the data given in the TG-51 Addendum and removing the common component of the Canadian primary standard of absorbed dose to water, then the level of agreement between stated and measured doses is within the overall k=1 uncertainty.
The alanine dosimetry system is also being used for reference dose measurements for FLASH beams from the IntraOp Mobetron system at CHUM, along with a portable beam current monitor [35], since conventional real-time dosimeters (i.e., ionization chambers) perform poorly in beam with ultrahigh doses per pulse.