Abstract
We aim to develop tools for ground-truth phantoms for quantitative and structural MRI (qMRI). qMRI aims to
acquire maps of physical or chemical variables that can be measured in physical units and compared between
tissue regions and among subjects. In contrast, most clinical MRI acquisitions are only qualitative, i.e. “weighted
images”, and not quantitative. While qMRI has the potential to improve precision diagnostics and medicine, it
has been traditionally hampered by significant barriers such as imaging speed, computational practicalities, and
reproducibility and repeatability of MR measurements. The variability between scanners and human subjects
and the lack of ground truth in biological tissues fundamentally challenge the development, testing and
standardization of qMRI techniques. The National Institute of Standards and Technology (NIST) hosted
workshops working towards standardizing qMRI. The resulting recommendation paper highlighted a list of
outstanding needs. The proposed project aims to address these unmet needs by developing materials,
technology, tools and processes for manufacturing quantitative anthropomorphic MRI phantoms. Current state-
of-the-art solutions for manufacturing MRI phantoms often use discrete compartments or geometrical shapes
filled with chemical solutions representing a single physical parameter. In contrast, our proposed novel approach
will enable fabrication of phantoms that truly mimic the contrast heterogeneity of tissue in 3D. These will include
proton density, T1, T2, T2* relaxation times, magnetic susceptibility, diffusion, fat fraction, air-tissue field-
inhomogeneity, relative conductivity, electric permittivity and magnetic permeability. If successful, this will be the
first time that such a comprehensive set of MRI parameters is accomplished in a tissue-mimicking phantom.
Based on our preliminary work on quantitative anatomy mimicking slice phantoms, we propose two approaches:
(a) Quantitative 3D stack of thin slices. This approach is inexpensive, easy to reproduce by labs with moderate
equipment and skills. (b) An advanced approach of boundaryless fully 3D phantoms that will be fabricated via
inkjet 3D printing of hydrogels and plastics and would enable true high resolution 3D structures with
heterogeneity that mimics human anatomy. In collaboration with leading industrial partners, we will validate and
disseminate our technology. Our proposal is motivated by a rising need for quantitative measurements in MRI
driven by precision medicine and the use of data science tools for biomarker discovery. With the rise of methods
such as fingerprinting, and accelerated reconstruction, quantitative MRI (qMRI) is closer to the clinics than ever.
The proposed quantitative MRI phantom will mimic the complexity of tissue structure and contrast mechanism
that are necessary to ensure the accuracy of qMRI. If successful, the project will greatly facilitate the development
and clinical translation of qMRI, making MRI accurate, precise, and quantitative – thus enabling precision
diagnostic and discoveries that will directly improve healthcare.
