TO THE EDITOR: In a recent paper, Dr. Gowrishankar and colleagues have demonstrated that 6″-18F-fluoromaltotriose, which targets the bacterial maltodextrin transporter, is taken up by a variety of pathogenic bacterial strains in vitro and in vivo (1). This new tracer might thus play a major role in diagnosis and, potentially, in assessing response to antibiotic therapy. In particular, in a simple Escherichia coli–induced myositis model, the authors compared two 1-h dynamic PET time–activity curves that were obtained in mice bearing both viable and heat-inactivated bacteria injected in left and right thigh muscle, respectively (Fig. 2A in Gowrishankar et al. (1)). These decay-corrected time–activity curves showed 2 remote peaks at about peak time (tpeak) = 4.5 and 27.5 min after injection, respectively, thereby indicating that tracer trapping was reversible in each muscle (2).
We thought of interest to further investigate the comparison between these 2 time–activity curves, focusing on their common input function (IF), for which the time constant α can be assessed from their peak time. Previous studies have shown that in each time–activity curve, tracer release rate constant kB can be obtained from a monoexponentially decaying fit of its decreasing part, and when tpeak and kB are known, the value of α can be obtained from the equation tpeak = Ln [α/kB]/[α− kB] (assuming IF decay correction and monoexponential decay) (3,4). Fitting the last 5 data points in each time–activity curve provided the following kB values 0.011 and 0.018 min−1 and R = 0.996 and 0.991, hence leading to an IF time constant estimate of α = 0.883 versus 0.085 min−1 for control versus infected muscle (using solver in Microsoft Excel software), respectively. This 10-fold discrepancy in α does not make sense because the tracer IF must be exactly the same for any tissue in a mouse and, more specifically, whereas the value of α in infected muscle may be plausible that in control muscle is just not realistic. In an attempt to explain this major discrepancy, we would like to suggest that the issue of a time-decaying uptake rate constant for the control muscle, in other words, an uptake saturation, may be considered. Indeed, it has been previously shown that a time decay of the tracer uptake rate is equivalent to an apparent increase in the IF time constant α, leading to a peak time of the tissue time–activity curve earlier than without saturation (Appendix in Laffon et al. (5)). In this connection, the uptake rate constant of the control muscle could be written as: Ki(t) = Ki × exp(−0.798 × t) where 0.798 min−1 is the difference between the 2 α values “0.883–0.085.” That is, the number of tracer molecules that could be potentially trapped in control muscle was very likely too small in comparison with that of injected ones. We therefore suggest that the lower the expected number of injected tracer molecules to be trapped in a tissue of interest, the lower the activity to be injected. Otherwise, the so-called tracer dose assumption usually made in molecular PET imaging, that is, radiotracer is injected in a small amount that does not affect its own kinetics, may be ruled out. Because of a too large amount of injected tracer molecules leading to a saturation situation, tracer uptake may be hard to quantify because of its time-varying nature. Furthermore, we suggest that the above-proposed reasoning for identifying a saturation situation might apply to the framework of the radiopharmaceutical use for therapeutic purpose, in an effort to limit adverse effects and to optimize costs.
To conclude, Dr. Gowrishankar and colleagues have convincingly demonstrated that 6″-18F-fluoromaltotriose is able to image bacterial infections in preclinical models and have shown that the pharmacokinetic properties of this novel tracer make it suitable for future clinical studies. On the basis of their results (illustrated in Fig. 2A of Gowrishankar et al. (1)), we suggest that uptake saturation might occur in PET imaging, as assessed by using the above-proposed rationale.
Footnotes
Published online Feb. 15, 2018.
- © 2018 by the Society of Nuclear Medicine and Molecular Imaging.