From PET Tracer Binding to an Occupancy Estimate
PET does not directly display the fraction of targets occupied by a therapeutic drug. It measures the distribution of a radiolabeled tracer, and a kinetic model is used to estimate how much tracer binds specifically to the target. Comparing that estimate before and after dosing can indicate how much target became less available to the tracer. The calculation is straightforward; its interpretation depends on whether the tracer signal is comparable across scans.
The central comparison is within a target region: estimate tracer binding at baseline, estimate it again after drug administration, and express the reduction relative to baseline. This produces an estimate at the post-dose scan time, not a general property of the drug independent of dose, timing, target, or measurement conditions.
The kinetic estimate may be obtained with an arterial-input model or with a reference-tissue method when that method’s assumptions are appropriate. In either case, a raw PET intensity or standardized uptake value is not itself a direct occupancy measurement. The occupancy calculation uses a binding measure such as \(BP_{ND}\), or a quantity that can be converted to it. When distribution volume ratios are reported, \(BP_{ND}=DVR-1\) under the relevant reference-tissue definition.
Deriving the Baseline-to-Post-Dose Calculation
Let \(BP_{ND,0}\) be the baseline binding potential and \(BP_{ND,1}\) the post-dose binding potential. Suppose the drug occupies a fraction \(O\) of the target sites that would otherwise be available to the tracer. If other determinants of the tracer binding potential remain unchanged, post-dose availability is \(1-O\) times baseline availability. Since \(BP_{ND}\) is proportional to availability under those conditions,
Solving for \(O\) gives the PET availability-ratio estimate:
Multiplying by 100 expresses the estimate as a percentage. For example, a ratio of post-dose to baseline binding potential of 0.40 corresponds to an estimated 60% reduction in tracer-available target. The estimate is meaningful only when baseline and post-dose measures refer to the same target region and are on a comparable scale.
Worked PET Occupancy Calculations
Worked Example: Occupancy From Binding Potential
Hypothetical data. A participant has baseline \(BP_{ND}=2.40\) in a target-rich brain region. At a specified post-dose scan time, the estimated \(BP_{ND}\) is 0.96. Assume the tracer, analysis, and target conditions are comparable across scans.
Substitute the two estimates into the availability-ratio formula:
The estimated occupancy is \(0.600\times100=60.0\%\). In context, the post-dose tracer binding potential is 60.0% lower than the baseline value, consistent with an estimated 60.0% occupancy under the stated assumptions. It is not correct to say that the PET image directly counted 60% of the receptors as drug-bound.
Worked Example: Converting Distribution Volume Ratios
Hypothetical data. A reference-tissue analysis reports baseline \(DVR=2.80\) and post-dose \(DVR=1.90\). Under the method’s definition, \(BP_{ND}=DVR-1\). Assume the reference region is suitable and its non-displaceable distribution volume is stable across scans.
First convert each ratio to binding potential:
Then calculate occupancy:
The estimated occupancy is 50.0%. Applying the formula directly to the DVR values would instead give \(1-1.90/2.80=0.321\), or 32.1%, which is not the appropriate calculation here: the non-specific component represented by the subtraction of 1 must first be removed to obtain \(BP_{ND}\).
Worked Example: Averaging Individual Occupancy Estimates
Hypothetical data. Two participants have baseline and post-dose binding potentials of 2.00 and 1.00, and 4.00 and 3.00, respectively. Calculate each participant’s occupancy, then compare the mean individual estimate with the estimate formed from the group means.
For participant A:
For participant B:
The arithmetic mean of the individual occupancy estimates is \((50.0\%+25.0\%)/2=37.5\%\). The baseline group mean is \((2.00+4.00)/2=3.00\), and the post-dose group mean is \((1.00+3.00)/2=2.00\). The ratio-of-means calculation is therefore \(1-2.00/3.00=0.333\), or 33.3%.
These values differ because the ratio of group means weights the individual binding-potential changes differently from the arithmetic mean of individual ratios. If the scientific target is the average participant’s occupancy, calculate each participant’s occupancy and average those estimates. State clearly which summary is reported.
Assumptions and Checks Before Interpreting the Estimate
The ratio is not assumption-free. It attributes the decrease in tracer binding potential to drug-related loss of availability only when other relevant influences are sufficiently stable or accounted for by the PET analysis. Important conditions include:
- Comparable tracer measurement. The tracer, acquisition protocol, target-region definition, and analysis method should be consistent across baseline and post-dose scans.
- Stable relation between availability and binding potential. Changes in tracer affinity, non-displaceable binding, or the relevant free fraction should not be mistaken for changes in available target.
- Appropriate kinetic model. The model must adequately represent tracer delivery and binding. A reference-tissue method additionally requires a suitable reference region with negligible specific binding and comparable non-displaceable distribution.
- Tracer does not materially occupy the target. Tracer mass and specific activity should be such that the tracer itself does not substantially displace target ligand or alter target availability.
- Comparable target state apart from the drug effect. Target density, endogenous ligand competition, and other physiological influences should not change enough between scans to explain the observed reduction independently.
- Appropriate scan timing. The estimate describes occupancy at the time represented by the post-dose scan. If drug concentration or binding is changing during acquisition, a single equilibrium-style interpretation may not describe the full scan.
Some of these conditions are supported by protocol design; others are assessed from the fitted time-activity data and model diagnostics. A plausible-looking occupancy percentage is not evidence that the model is adequate. Review fit quality, parameter plausibility, regional time-activity curves, and whether the selected model can represent the observed tracer kinetics. Large differences in tracer delivery or unstable reference-region behavior warrant caution, even if a numerical \(BP_{ND}\) estimate is available.
Use a justified PET kinetic method to obtain baseline and post-dose \(BP_{ND}\), or convert an appropriate \(DVR\) estimate to \(BP_{ND}\).
Confirm that target region, tracer protocol, model, and relevant reference or input assumptions are suitable across the scans.
Compute \(1-BP_{ND,1}/BP_{ND,0}\) and retain the scan time and target region in the interpretation.
Describe the result as an occupancy estimate conditional on the assumptions, not as a direct count of bound targets or proof of downstream benefit.
Diagnostics, Alternatives, and Limits
A post-dose \(BP_{ND}\) greater than baseline produces a negative value from the formula. That result should not automatically be set to zero or described as negative biological occupancy. It indicates that the observed binding potential increased, and can arise from measurement variability, a genuine change in conditions, or model misspecification. Investigate comparability and model fit, report the observed estimates, and explain why an occupancy interpretation is uncertain.
A very low baseline \(BP_{ND}\) is another warning: the denominator is small, so even a modest absolute difference can produce a large percentage change. If post-dose binding potential is near zero, the formula may yield an estimate near 100%, but the precision and reliability of that estimate depend on how well the binding measure can be distinguished from zero. Do not treat a boundary-looking percentage as proof of complete target occupancy.
The baseline-to-post-dose approach is useful because each participant serves as their own reference. Its main vulnerability is that any change affecting tracer binding can be confounded with the drug’s effect. A placebo-controlled or repeated-baseline design can help assess time-related or procedural shifts. Arterial-input and reference-tissue kinetic methods are alternatives for estimating binding, not alternative definitions of occupancy: each has different data requirements and assumptions. A simple comparison of raw uptake is less suitable when delivery, metabolism, or non-specific distribution could differ between scans.
Finally, the calculation is relative to the measured baseline availability. Baseline tracer binding can already reflect endogenous competitors or other determinants of availability. Thus, the estimate is generally a drug-associated change relative to baseline, not necessarily an absolute fraction of every target molecule occupied in the tissue. PET occupancy also does not, on its own, determine the size or direction of a pharmacological response.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- Baseline \(BP_{ND}\) is 3.00 and post-dose \(BP_{ND}\) is 1.20. What is the estimated occupancy, and what does it mean?
- A study reports baseline \(DVR=2.40\) and post-dose \(DVR=1.90\). Assuming \(BP_{ND}=DVR-1\), calculate the occupancy estimate.
- Two participants have occupancy estimates of 20% and 80%. What is their mean individual occupancy? Why might a ratio calculated from group-average binding potentials not equal it?
- Post-dose binding potential is higher than baseline. Should the calculated negative estimate be silently changed to zero? Explain.
Solutions. (1) \(\hat{O}=1-1.20/3.00=0.600\), or 60.0%. The post-dose tracer binding potential is 60.0% below baseline, interpreted as estimated occupancy only if the comparison assumptions hold. (2) Convert first: baseline \(BP_{ND}=2.40-1=1.40\), post-dose \(BP_{ND}=1.90-1=0.900\). Then \(\hat{O}=1-0.900/1.40=0.357\), or 35.7%. (3) The mean individual estimate is \((20\%+80\%)/2=50.0\%\). A ratio of group means uses the binding-potential values in its numerator and denominator, so it does not generally equal the unweighted mean of individual ratios. (4) No. A negative result signals increased measured binding potential and may reflect variability, changed conditions, or model problems. Report and investigate it rather than silently truncating it.