PET imaging of thin objects: Measuring the effects of positron range and partial-volume averaging in the leaf of Nicotiana tabacum

Introduction: PET imaging in plants is receiving increased interest as a new strategy to measure plant responses to environmental stimuli and as a tool for phenotyping genetically engineered plants. PET imaging in plants, however, poses new challenges. In particular, the leaves of most plants are so thin that a large fraction of positrons emitted from PET isotopes (¹⁸F, ¹¹C, ¹³N) escape while even state-of-the-art PET cameras have significant partial-volume errors for such thin objects. Although these limitations are acknowledged by researchers, little data have been published on them. Methods: Here we measured the magnitude and distribution of escaping positrons from the leaf of Nicotiana tabacum for the radionuclides ¹⁸F, ¹¹C and ¹³N using a commercial small-animal PET scanner. Imaging results were compared to radionuclide concentrations measured from dissection and counting and to a Monte Carlo simulation using GATE (Geant4 Application for Tomographic Emission). Results: Simulated and experimentally determined escape fractions were consistent. The fractions of positrons (mean±S.D.) escaping the leaf parenchyma were measured to be 59±1.1%, 64±4.4% and 67±1.9% for ¹⁸F, ¹¹C and ¹³N, respectively. Escape fractions were lower in thicker leaf areas like the midrib. Partial-volume averaging underestimated activity concentrations in the leaf blade by a factor of 10 to 15. Conclusions: The foregoing effects combine to yield PET images whose contrast does not reflect the actual activity concentrations. These errors can be largely corrected by integrating activity along the PET axis perpendicular to the leaf surface, including detection of escaped positrons, and calculating concentration using a measured leaf thickness.