3D-printable phantoms for quantitative dynamic contrast-enhanced MRI

This dataset contains the experimental data and code required to reproduce results and figures in the article "3D-printable phantoms for quantitative dynamic contrast-enhanced MRI" (final published version). Abstract ----------- Purpose: A novel 3D-printed phantom design and methodology is proposed, addressing important requirements for technical validation, quality assurance and multi-site harmonisation of quantitative dynamic contrast-enhanced (DCE-) MRI measurements. Methods: Phantoms were produced by 3D-printing gels incorporating channels and pores as proxies for blood vessels and extravascular extracellular space, respectively. A flow circuit was designed to reproduce clinically relevant arterial input functions (AIF). Using nine gels with variable porosity and channel size we evaluated the effect of 3D-printing parameters on DCE-MRI parameters obtained using the extended Tofts (ET) model. Physical gel and fitted model parameters were correlated by multiple linear regression. Results: All phantoms generated realistic AIFs and tissue-like signal enhancement curves were accurately modelled by the ET model. As hypothesised, vp was positively associated with vchan (B = 0.86, p < 0.001) and showed a weaker, negative association with vpore (B = -0.18, p = 0.006); PS was positively associated with both vchan (B = 0.13 min-1, p < 0.001) and vpore (B = 0.051 min-1, p < 0.001). ve was positively associated with vpore (B = 0.90, p < 0.001) but not vchan. Fitted parameters were repeatable (coefficient of variation 2.1 - 3.2 %). Conclusions: Tailorable 3D printed porous gel phantoms generating tissue-mimicking DCE-MRI signals have the potential to support validation, quality assurance and multi-site harmonisation of quantitative DCE-MRI.