Repository logo

Combining MRI with PET for partial volume correction improves image-derived input functions in mice.

Accepted version

Repository DOI



Change log


Evans, Eleanor 
Buonincontri, Guido 
Izquierdo, David 
Methner, Carmen 
Hawkes, Rob C 


Accurate kinetic modelling using dynamic PET requires knowledge of the tracer concentration in plasma, known as the arterial input function (AIF). AIFs are usually determined by invasive blood sampling, but this is prohibitive in murine studies due to low total blood volumes. As a result of the low spatial resolution of PET, image-derived input functions (IDIFs) must be extracted from left ventricular blood pool (LVBP) ROIs of the mouse heart. This is challenging because of partial volume and spillover effects between the LVBP and myocardium, contaminating IDIFs with tissue signal. We have applied the geometric transfer matrix (GTM) method of partial volume correction (PVC) to 12 mice injected with 18F-FDG affected by a Myocardial Infarction (MI), of which 6 were treated with a drug which reduced infarction size [1]. We utilised high resolution MRI to assist in segmenting mouse hearts into 5 classes: LVBP, infarcted myocardium, healthy myocardium, lungs/body and background. The signal contribution from these 5 classes was convolved with the point spread function (PSF) of the Cambridge split magnet PET scanner and a non-linear fit was performed on the 5 measured signal components. The corrected IDIF was taken as the fitted LVBP component. It was found that the GTM PVC method could recover an IDIF with less contamination from spillover than an IDIF extracted from PET data alone. More realistic values of Ki were achieved using GTM IDIFs, which were shown to be significantly different (p<0.05) between the treated and untreated groups.



Arterial input function, Geometric transfer matrix, MRI, partial volume correction, small animal PET

Journal Title

IEEE Trans Nucl Sci

Conference Name

Journal ISSN


Volume Title



Institute of Electrical and Electronics Engineers (IEEE)
Medical Research Council (G0001354)
Medical Research Council (G1000183)
E. Evans was funded by an MRC studentship and travel to PSMR 2014 was funded by the EU COST action for PET/MR.