Plasma to blood ratio derived parent fractions for robust plasma input based kinetic analysis of dynamic 18F-FLT PET studies in NSCLC cancer patients

Arturo Avendano Estrada, G. Kramer, V. Frings, David Vállez García, E Smit, A. A. Lammertsma, O Hoekstra, Ronald Boellaard

Research output: Contribution to journalMeeting AbstractAcademic


Purpose/Introduction: Full quantitative analysis of dynamic 18F-FLT PET studies is important to validate the use of simplified methods and metrics, such as SUVs, for early response assessment (1). However, full quantitative kinetic analysis requires the use of a metabolite corrected plasma input function. The measurement of parent fractions is often associated with poor signal to noise ratio, measurement errors and it is labor intensive. In the present work, we investigated if parent fractions could be estimated directly from plasma and blood activity concentrations thereby obviating the need to perform metabolite analysis and/ or use estimated values for quality control purposes. Subjects & Methods: Nine NSCLC patients underwent three 60 min dynamic 18F-FLT PET scans: at baseline within 7 days before treatment and at 7 and 28 days after start of the treatment. During each PET study 6 to 7 venous blood samples were collected to obtain blood and plasma concentrations as well as parent fractions. These data were subsequently used in combination with an image derived input function to generate a metabolite corrected input function. A plasma input reversible 2-tissue compartment model (PMOD v3.7) was applied to the tumor time activity curves, yielding tumor VTs. This analysis was repeated using parent fractions calculated from plasma to blood ratios using an equation derived from the baseline scans only. VTs derived from analysis using measured and calculated parent fractions were compared as well as longitudinal % changes in VT using both approaches. Agreement was assessed with linear regression, interclass correlation coefficients (ICC) and Bland-Altman plots. Results: Excellent agreement between VTs derived from measured vs calculated parent fractions was found for both early (R2=0.983[0.981-0.996 (95%CI)]; ICC=0.992[0.982-0.996]) and late response (R2=0.994[0.992-0.998]; ICC=0.996[0.991-0.998] scans. Furthermore, Bland-Altman plots showed a strong agreement for both early (percentage bias=0.284[-6.172-6.741]) and late response (percentage bias=-1.165[-7.125-4.795) studies. Discussion/Conclusion: The strong agreement in tumor 18F-FLT VTs as well as their longitudinal changes when obtained from input functions using measured or calculated parent fractions suggest that parent fractions can be reliably estimated using plasma to blood ratios. Plasma to blood ratio derived parent fractions could be used as an alternative when blood sample data are noisy or incomplete, to discard outliers, or when it is not possible to acquire sufficient samples for metabolite analysis. The method may be applicable to other tracers as well, but this requires further studies. Reference: 1. Frings et al., JNM 2014

Original languageEnglish
Pages (from-to)S442
JournalEuropean Journal of Nuclear Medicine and Molecular Imaging
Issue numberSuppl 2
Publication statusPublished - 6-Oct-2017
Event30th Annual Congress of the European Association of Nuclear Medicine (EANM) - Vienna, Austria
Duration: 21-Oct-201725-Oct-2017
Conference number: 30

Cite this