Creating Reproducible Scientific Measurement & Calibration Devices to Enable Quantitative MRI: A Phantom Manufacturer's Perspective
Aaron Oliver-Taylor1
1Gold Standard Phantoms Limited, United Kingdom

Synopsis

Keywords: Transferable skills: Commercialisation, Transferable skills: Metrology of MRI

This talk discusses the importance and application of MRI phantoms for clinical and scientific research from the perspective of a phantom manufacturer. Phantoms enable evaluation of MRI performance, accuracy, and reproducibility. These are used in the development of new techniques, harmonisation across different systems, and validation of measurements. Commercial phantoms have expectations for stability, reproducibility, and longevity, making their design and manufacture complex. There are emerging standards for MRI phantoms involving traceable measurements to national measurement laboratories. The future of this field lies in addressing challenges like system standardisation, accommodating new MRI technologies, and integrating phantoms into routine radiological workflows.

Introduction

MRI is a tremendously powerful imaging modality, that can sensitise image contrast to a wide range of phenomena, for example, intrinsic MRI properties (T1, T2), molecular diffusion, chemical environment (CEST, MRS, MT, PDFF), electromagnetic (B0, B1, susceptibility), mechanical (MR elastography), flow, and perfusion. In many of these cases, acquired images can be fitted to a mathematical model to calculate quantitative parameter maps of a particular quantity of interest, turning what is essentially a camera into scientific apparatus. This is of great value both in clinical practise, and for scientific research, however to be able to use these measures in an objective way, we need to ensure that we can reliably measure these quantities across different MRI systems, ensuring:
  • Scans are comparable between different MRI systems.
  • Data quality and objectivity in clinical trials
  • Reproducibility of new MRI methods.
In all these cases, appropriate phantoms can be used as standardised test objects, to evaluate the performance, accuracy, and reproducibility of MRI images.

What is a test object?

A test object is an object that is used to evaluate the performance or accuracy of a measurement system or process. It is also known as a phantom or a calibration object. Anything can be a test object, for example, a pencil could be used to test something that measures length, a penny to test a weighing scale, or PhD student for trialling a new version of a MRI pulse sequence. A measurement is made on the test object, and the results compared with what we expect it to be.
Good test objects should ideally provide a traceable value of the quantity for which it is a test object. In the case of a pencil to test a ruler, we could measure the length of the pencil using a precision caliper, that has been calibrated in a manner that is traceable to one of the national measurement laboratories (NIST, NPL, PTB, LNE, etc) and therefore the SI unit system, and so have the length of the pencil with known uncertainty. When we then test the ruler with the pencil, the assessment is also traceable all the way to the SI unit system. To actually demonstrate all of this in a way that satisfies any regulatory body who has an interest in measurement is quite a lot of work, so in practise it is far simpler to buy a commercially available test object that comes with a certificate, for example OIML weights[1].

Phantoms in MRI

Phantoms for MRI simulate the properties of real biological tissues and processes. They can be as simple as a bottle of water, or complex, multi-compartmental objects that simulate many different tissue types, potentially including dynamic features such as fluid flow or motion. Phantoms are used in QMRI for a variety of purposes, including:
  • Development: Phantoms are used in the development of new MRI techniques and sequences. They allow researchers to test and optimise these techniques in a controlled environment before applying them to real patients.
  • Quality Assurance: routine scanning of a phantom and analysis of the results to check that the MRI system is operating to specification. Issues with the hardware can be detected, and usually result in a visit from a service engineer. The MRI vendor’s service engineers will also use phantoms to ensure that the MR system is performing to specification, and indeed MRI systems come with many phantoms just for this purpose.
  • Acceptance testing: checking that a particular site (comprising MR system and staff) is able to acquire data of a specified quality, for enrolment into a clinical trial for example.
  • Validation: A phantom can be used as a known ground truth to check that protocols or pulse sequences are valid and produce the correct results.
  • Harmonisation and Standardisation: Phantoms are used to ensure that MRI scanners from different manufacturers and locations produce consistent and comparable results. This is important for clinical trials and research studies that involve multiple sites.
As with all models, there is a trade-off between the complexity of the phantom, and its utility as a tool to evaluate the performance of the MRI scanner. For example, different concentrations of Nickel Chloride in water can be used to control T1 relaxation times, but paramagnetic relaxation enhancement is not the same relaxation mechanism that is observed in tissue. More complex materials with macromolecular, semi-solid components could be used, but (like real tissue) will respond in different ways to different pulse sequences due to the degree of off-resonance saturation present. If the objective is to standardise T1 mapping across different MR systems then the Nickel Chloride phantom would be best. If the goal is to investigate tissue T1 due to different pulse sequences then the more complex material is a better choice.

Design and manufacture of commercial phantoms

Commercial phantoms are designed and manufactured by companies that specialise in producing these devices. In general, they are more standardised and consistent in their properties than “home-made” phantoms built by researchers, due to the fact that there is an expectation that if two customers have the same phantom, they should perform in the same way. There is also an expectation that the phantom should last for a decent amount of time, rather than just long enough to acquire the data for publishing a paper. This being said, there are many examples of high quality, stable, and reproducible phantoms in the literature[2, 3].
As commercial products, there are expectations on the longevity, reproducibility and degree of standardisation between the same phantom model. The best examples of commercial phantoms give clear instructions to the user - how to scan it, how to analyse the images, and how to interpret the results. The ACR MRI phantom is a good example of this[4]. These factors present certain challenges in their design and manufacture:
  • The stability of the MRI visible materials, and their interactions with other materials and components that make up the structure of the phantom.
  • Materials chosen must of course be compatible for use within the imaging volume of the MRI scanner. Plastics are commonly used, but there are limitations to their mechanical strength, chemical compatibility, and available fabrication methods.
  • How to ship: hazardous materials (sodium azide for example) create paperwork.
  • Cost - despite their appearance, MR phantoms are high-tech and expensive to develop and produce. This must be balanced with market demand.
  • Significant effort must be placed on ensuring reproducible, traceable manufacturing methods, whilst also optimising and scaling production capacity to deliver phantoms at a price point that the market will accept.
  • In general, phantom manufacturers do not have their own MRI scanners, so close partnerships with researchers and clinicians are a must.

Standards for MRI phantoms

MRI phantoms have been used for as long as the technique has existed. 30 years ago it was possible to buy the Eurospin set of test objects[5, 6] set with specified T1’s. However there was no way to benchmark these other than to make reference measurements on a single MRI system and compare against these.
NIST has developed SI traceable methods to measure proton relaxation times [7] and diffusion [8], that are available as a commercial service. Phantom manufacturers simply send a sample from their batch of manufactured phantom materials for testing. NIST will then send a report with the measured values and uncertainties. Customers can buy phantoms with these materials included, and therefore have SI traceable MR quantities available to test their MRI systems.
For phantom manufacturers this provides an independent way to validate their products, and increases the value and utility of the phantoms to their customers. It is however expensive, and requires a sizeable batch quantity to be cost effective.
We are in the very early days of this field - at present the service is only available from NIST, however other national measurement laboratories are following. In other areas of measurement, SI traceable calibration or characterisation would be performed by test houses and not the NML’s directly.

Future Challenges

Some of the key challenges that phantom manufacturers will face in the future include:
  • Establishing a system of primary and secondary/field standards so that this service isn’t reliant on a handful of national measurement laboratories. If QMRI really takes off as we hope it will then this will be essential to prevent bottlenecks.
  • New MRI technologies and emerging MRI applications present new challenges and opportunities to develop phantoms. Here, researchers, the NML’s, and phantom manufacturers need to work closely to develop solutions.
  • How can phantoms be better integrated into the radiological workflow? Will we get imaging centre buy-in if they have to routinely scan a different phantom for every QMRI method they use?
  • With a QMRI phantom we can calibrate a MR system when it scans that particular phantom. But how can we use that information to calibrate a patient image? Due to patient-MR system interactions, the B0, B1 etc are slightly different for every patient.

Acknowledgements

No acknowledgement found.

References

  1. OIML R 111-1 Edition 2004 (E) https://www.oiml.org/en/files/pdf_r/r111-1-e04.pdf
  2. G. Captur et al., “A medical device-grade T1 and ECV phantom for global T1 mapping quality assurance—the T1 Mapping and ECV Standardization in cardiovascular magnetic resonance (T1MES) program,” J Cardiovasc Magn Reson, vol. 18, no. 1, Sep. 2016, doi: 10.1186/s12968-016-0280-z.
  3. H.-M. Cho et al “LEGO-compatible modular mapping phantom for magnetic resonance imaging,” Sci Rep, vol. 10, no. 1, Sep. 2020, doi: 10.1038/s41598-020-71279-1.
  4. Phantom Test Guidance for Use of the Large MRI Phantom for the ACR MRI Accreditation Program. https://www.acraccreditation.org/-/media/acraccreditation/documents/mri/largephantomguidance.pdf
  5. R. A. Lerski and J. D. de Certaines, “II. Performance assessment and quality control in MRI by Eurospin test objects and protocols,” Magnetic Resonance Imaging, vol. 11, no. 6, pp. 817–833, Jan. 1993, doi: 10.1016/0730-725x(93)90199-n.
  6. R. A. Lerski et al., “V. Multi-center trial with protocols and prototype test objects for the assessment of MRI equipment,” Magnetic Resonance Imaging, vol. 6, no. 2, pp. 201–214, Mar. 1988, doi: 10.1016/0730-725x(88)90451-1.
  7. M. A. Boss et al., “Magnetic resonance imaging biomarker calibration service: proton spin relaxation times,” National Institute of Standards and Technology, May 2018. doi: 10.6028/nist.sp.250-97.
  8. S. E. Russek, “Magnetic Resonance Imaging Biomarker Calibration Service:,” National Institute of Standards and Technology, 2022. doi: 10.6028/nist.sp.250-100.
Proc. Intl. Soc. Mag. Reson. Med. 31 (2023)