Every oncologist who has worked with Lu-177 PSMA or DOTATATE therapy knows the frustration. You prescribe 7.4 GBq because that is what the registration trial used. The dose is the same for the 65 kg patient with one bone lesion and moderate renal function, and for the 95 kg patient with multiple hepatic deposits and a GFR of 42. The tumors, the kidneys, and the clearance curves are entirely different. The number in the order set is identical. Something is wrong with this picture, and the problem is not a technical gap in our imaging capabilities. It is a structural mismatch between how prescribed activity and delivered dose relate to each other across a heterogeneous population.
Reason One: Renal Clearance Varies Three-Fold Across Eligible Patients
Renal function is the primary determinant of Lu-177 clearance rate. Lutetium-177 radiopharmaceuticals that do not bind their target are cleared renally, and the rate of that clearance determines how long the radiopharmaceutical remains in circulation and available for tumor accumulation. A patient with GFR 80 mL/min/1.73m2 and a patient with GFR 35 mL/min/1.73m2 who both meet inclusion criteria for Lu-177 therapy can have clearance rates that differ by a factor of two to three for the non-target-bound fraction.
The consequence for dosimetry is direct. Slower clearance increases the time-integrated blood activity, which raises the dose to normal tissues proportionally. For the kidney itself, which filters and re-concentrates the radiopharmaceutical during excretion, slower clearance raises kidney exposure even when blood levels are only moderately elevated. Published data from dosimetry-enabled Lu-177 PSMA series show kidney absorbed dose per cycle at 7.4 GBq ranging from roughly 1.5 Gy to over 6 Gy, largely tracking with renal clearance rate. That is a four-fold spread at a fixed prescribed activity.
A fixed-activity protocol that is safe on average is therefore under-dosing some patients who could tolerate more and approaching unsafe territory for others who clear slowly. Population selection criteria (excluding patients with GFR below 30 in some protocols) do not resolve this: they reduce the tail of the distribution but do not reduce the variance within the included population. A patient at GFR 55 and a patient at GFR 80 are both included, and their kidney doses diverge substantially across a multi-cycle course.
Reason Two: Tumor Receptor Density Is Not Uniform or Predictable
PSMA expression in mCRPC and somatostatin receptor expression in neuroendocrine tumors both vary substantially between patients and, importantly, between lesions within the same patient. PSMA PET provides a qualitative signal that confirms sufficient expression for therapy eligibility, but the quantitative uptake, measured as SUVmax or SUVmean at defined scan timing, can range over a factor of five to ten between a patient with high-avidity disease and one with moderate expression who still qualifies by threshold criteria.
This receptor-density variation drives tumor-absorbed dose variance that is independent of the renal clearance variance. Two patients with identical renal function receiving 7.4 GBq of Lu-177 PSMA-617 can show tumor absorbed dose per cycle differing by three to four fold based on lesion PSMA expression density alone. When renal clearance variance and receptor-density variance are combined, the distribution of actual tumor doses delivered at a fixed 7.4 GBq is wide. Some patients receive well above the minimum effective dose. Others receive doses that are unlikely to produce the tumor kill needed for durable response.
The argument here is not that dosimetry can predict clinical response. The relationship between absorbed dose and radiobiological effect in targeted radionuclide therapy is not as well characterized as in external beam radiotherapy. What dosimetry can do is confirm that the tumor received a physically plausible dose and identify patients whose tumor dose is likely below any plausible therapeutic threshold. That is a useful clinical signal even absent a precise dose-response curve.
Reason Three: Body Composition Affects Distribution Volume
The distribution volume for a radiopharmaceutical, which determines initial blood concentration after injection, scales with body composition in a way that body weight alone does not capture. The conventional fixed-activity approach applies the same 7.4 GBq regardless of body size. A more refined approach is activity per kilogram body weight, which some protocols use. But neither accounts for the fact that lean body mass, adipose tissue, and total body water differ between patients in ways that affect initial distribution and early kinetics.
In practice, this third factor is the least influential of the three described here. Body composition effects on initial distribution are real but typically smaller in magnitude than the renal clearance and receptor-density effects, particularly for radiopharmaceuticals like Lu-177 agents that are not extensively protein-bound and distribute primarily in blood water. We mention it not because it dominates the clinical problem but because it is a reminder that "same prescribed activity" does not mean "same initial conditions" even before any pharmacokinetics have begun.
What the Absorbed Dose Window Actually Looks Like in Practice
Published dosimetry data from Lu-177 PSMA cohorts show tumor lesion absorbed doses per cycle at 7.4 GBq typically ranging from under 5 Gy to over 20 Gy, with some high-avidity lesions receiving 30 Gy or more. Kidney doses at the same fixed activity span roughly 1.5 to 7 Gy per cycle. The cumulative kidney constraint used in most protocols is 23 Gy across all cycles, inherited from the radiopeptide therapy literature. A patient receiving 5 Gy per cycle to the kidney can complete four to five cycles within that constraint. A patient receiving 7 Gy per cycle approaches the constraint after three cycles. The number of cycles that can be safely delivered before reaching kidney tolerance is patient-dependent, not protocol-fixed, and that difference has direct consequences for cumulative tumor dose across a treatment course.
This is the practical case for cycle-by-cycle dosimetry: not to optimize the dose of each individual injection in isolation, but to track cumulative kidney dose and project when the constraint will be reached for this patient at this clearance rate. That projection lets the prescribing oncologist and nuclear medicine physician decide together whether the current activity per cycle is appropriate, whether modification of the schedule is warranted, and whether additional cycles are feasible. These are clinical decisions that belong with the clinical team. Dosimetry provides the dose accounting that makes them data-driven rather than intuitive.
The Path Forward Is Not Abandoning Fixed Protocols Overnight
It would be wrong to conclude from the above that every theranostics program should immediately abandon fixed-activity prescribing. For centers without dosimetry infrastructure, the population-average approach is far better than no treatment. The fixed 7.4 GBq protocol delivers meaningful clinical benefit across a wide range of patient types, as demonstrated by the registration trials. The issue is not that fixed protocols fail all patients; it is that they underperform for a predictable subset whose individual pharmacokinetics diverge from the population mean.
The practical argument for building dosimetry capacity alongside the theranostics program is that identifying which patients are in that underperforming subset, and adjusting their treatment accordingly, is achievable without fundamentally restructuring the clinic's workflow. Serial SPECT acquisitions are already standard post-administration imaging for dosimetry-capable programs. The incremental burden is the analysis pipeline. For programs considering that investment, the clinical payoff is not speculative. It is a known quantity in the dosimetry literature, and it accrues most to the patients at the tails of the dose distribution: those who need more activity than the population average to reach a therapeutic dose, and those who need less to stay within their organ-at-risk constraint.