TO THE EDITOR: In an interesting article, Dr. Huisman et al. recently investigated the optimal kinetic model for an 18F‐labeled anti–programmed cell death ligand 1 (anti–PD-L1) adnectin, namely, 18F‐BMS‐986192, to quantify PD‐L1 expression in non–small‐cell lung cancer (NSCLC) patients (1). A single-tissue-reversible (STR) compartment model, additionally including blood-volume fraction, was found to be the most preferred model for fitting the tumor time–activity curves. Its specific outcome measure is the distribution volume (VT; mL.cm−3) that is the equilibrium ratio of forward/reverse transport-rate constants, that is, Ki/kb, between blood and reversible-trapping compartment (2). VT was then used to validate simplified methods, the best correlation being obtained with body weight–normalized SUV (SUVBW) at 50–80 min after injection (R2 = 0.92–0.91), whereas a lower correlation was obtained with SUV normalized to plasma concentration (SUV/Cplasma, presumably at 50 min after injection; R2 = 0.84). The authors conclude that SUVBW at 60 min after injection is an accurate simplified parameter for uptake assessment of 18F‐BMS‐986192 baseline studies.
We would like to further analyze the latter lower correlation since, under postinjection time conditions we address in this letter: (i) VT may be assessed by the ratio of tissue/plasma tracer concentration, i.e., Ctissue/Cplasma; and (ii) the SUVBW/Cplasma ratio may be also proportional to VT since SUVBW is proportional to Ctissue. To clarify this issue, we have fitted the 18F‐BMS‐986192 input function with a triexponentially decaying function and then fitted a PD-L1–positive tumor time–activity curve by using a 3-compartment 3-parameter kinetic model (data extracted from Figs. 3 and 4 in Huisman with the Web-Plot-Digitizer software; R2 = 0.996 and 0.998, respectively) (1,3). Estimates of Ki and kb were provided, leading to the computation of VT as Ki/kb = 4.7 mL.cm−3. This analysis also allowed us to perform both decay-corrected tissue- and decay-uncorrected trapped-tracer time–activity curves (supplemental data, available at http://jnm.snmjournals.org).
Let us first consider the rate of decay-corrected trapped tracer per tissue volume unit (at steady state): dCtrapped(t)/dt = Ki ×Cplasma(t) – kb ×Ctrapped(t). At peak time of decay-corrected Ctrapped time–activity curve, dCtrapped(t)/dt = 0 and then Ctrapped(tpeak)/Cplasma(tpeak) = Ki/kb = VT. Assuming Ctissue(tpeak) ≈ Ctrapped(tpeak) (i.e., neglecting free tracer in blood and interstitial volume), tpeak was estimated to be 87 min from decay-corrected Ctissue time–activity curve, leading to Ctissue/Cplasma = 4.5 mL.cm−3 (versus Ki/kb = 4.7 mL.cm−3). Second, considering decay-uncorrected data, the differential equation becomes dCtrapped(t)/dt = Ki ×Cplasma(t) – kb ×Ctrapped(t) – λ×Ctrapped(t), where λ is the 18F physical-decay-rate constant. As a consequence, at peak time of decay-uncorrected-Ctrapped time–activity curve, Ctrapped(tpeak)/Cplasma(tpeak) = Ki/(kb + λ). The ratio Ki/(kb + λ) was calculated as 2.1 mL.cm−3, whereas, at decay-uncorrected-Ctrapped tpeak of 53 min after injection, the ratio Ctissue/Cplasma (that may involve decay correction or not) was found to be 2.2 mL.cm−3.
We therefore suggest that the SUV/Cplasma ratio (or, equivalently, the Ctissue/Cplasma ratio) is actually correlated with VT = Ki/kb when assessed within 85–90 min after injection. However, the authors acknowledged that their results were only valid within 50–80 min after injection (1). Furthermore, we suggest that the SUV/Cplasma ratio assessed within 50–55 min after injection should be correlated with Ki/(kb + λ), instead of Ki/kb (1). This alternative ratio reports on what is actually occurring at decay-uncorrected-Ctrapped tpeak, that is, an equilibrium between uptake and release plus physical decay. It is worth noting, regarding the part of postinjection time in its measurement uncertainty, a +13% increase occurs in the 50–55 min time range, whereas, for comparison, Ctissue alone, and, hence, SUVBW, shows a +3% increase.
In conclusion, investigating potential clinical biomarkers and relevant simplified metrics is of upmost importance for selecting NSCLC patients who could benefit from immune checkpoint-inhibitor treatment. In 18F‐BMS‐986192 PET imaging, Huisman et al. convincingly showed that SUVBW, at 60 min after injection, may be a relevant simplified parameter to quantify tumor uptake for baseline PET studies. We additionally suggest that the ratio SUVBW/Cplasma might be probed as a complementary possible simplified parameter, that is correlated with Ki/(kb + λ) within 50–55 min after injection.
DISCLOSURE
No potential conflict of interest relevant to this article was reported.
Footnotes
Published online May 22, 2020.
- © 2021 by the Society of Nuclear Medicine and Molecular Imaging.