Multi-cycle Lu-177-PSMA radioligand therapy — tracking cumulative organ absorbed dose against fixed lifetime constraints across a full treatment course
Every course of radioligand therapy (RLT) begins not with treatment but with measurement. Before — or during — the first therapeutic infusion of Lu-177-PSMA-617, quantitative SPECT/CT imaging at several timepoints establishes each patient's individual pharmacokinetics, letting physicists calculate patient-specific absorbed doses to the kidneys, bone marrow, salivary glands, and tumor rather than relying on population averages. This baseline becomes the reference against which every later cycle's contribution is added.
The Medical Internal Radiation Dose (MIRD) committee scheme, developed by the Society of Nuclear Medicine and Molecular Imaging, is the standard framework for converting a radiopharmaceutical's biodistribution into an organ absorbed dose. The core equation is:
D(organ) = Ã(source) × S(target ← source)
Where à is the "cumulated activity" — the time-integral of activity in a source organ over all time — and S is the "S-value," a pre-tabulated geometric and physical factor describing the mean absorbed dose to a target organ per unit cumulated activity in a source organ, accounting for both self-irradiation (source = target) and cross-irradiation from neighboring organs.
Cumulated activity is obtained by fitting a time-activity curve to serial quantitative SPECT/CT measurements — typically at ~4 hours, 24 hours, 48 hours, and 168 hours (one week) post-infusion — then integrating that curve (mono-exponential or bi-exponential fit, or trapezoidal numerical integration) to get the total number of nuclear transformations occurring in that organ across all time. For Lu-177 (physical half-life 6.6 days), the effective clearance half-life in kidney tissue is typically 30–50 hours, meaning the majority of the cumulated activity — and therefore dose — is delivered within the first several days after each infusion.
PSMA (prostate-specific membrane antigen) is not exclusively expressed on prostate cancer cells — it is also physiologically expressed on renal proximal tubule cells, where PSMA-targeted radioligands are reabsorbed after glomerular filtration and concentrate in the renal cortex. This off-target uptake makes the kidney the principal dose-limiting organ for PSMA-targeted RLT, receiving roughly 2–4 Gy per standard 7.4 GBq cycle — several-fold higher than most other non-target tissues.
Bone marrow toxicity arises differently: rather than receptor-mediated uptake, marrow dose is driven largely by circulating blood activity (the radiopharmaceutical bathing the highly vascularized, radiosensitive marrow compartment) plus a smaller contribution from bone surface uptake. Because marrow cannot be imaged directly with the same precision as kidney or tumor, marrow dosimetry commonly relies on blood-sample-based surrogate methods (drawing venous blood samples at the same timepoints as imaging and using them as a proxy for marrow self-dose), which carry more uncertainty than image-based organ dosimetry.
Together, kidney and marrow doses accumulate cycle after cycle in a way that tumor efficacy alone does not constrain — a tumor could keep responding while the kidneys quietly approach an irreversible-injury threshold. This is precisely why cumulative, not per-cycle, dose tracking is the clinical safeguard.
A single Lu-177-PSMA-617 cycle rarely threatens kidney function on its own. The danger is arithmetic: at 3 Gy per cycle, six cycles alone put a patient at 18 Gy — before accounting for baseline renal impairment, diabetes, hypertension, or prior nephrotoxic chemotherapy that lower the effective tolerance ceiling.
The output of Cycle 1 dosimetry is a per-organ dose "map" that becomes the baseline row of a running ledger:
• Kidneys (both, or dose-limiting single functioning kidney): absorbed dose in Gy, lifetime constraint 23 Gy (EBRT-extrapolated) up to 28–40 Gy under biologically-effective-dose (BED) adjusted models • Red bone marrow: absorbed dose in Gy, conservative lifetime constraint ~2 Gy, drawn from radioimmunotherapy-era toxicity data • Salivary glands: relevant for xerostomia risk, though rarely dose-limiting for survival • Tumor lesions (index lesions tracked individually): absorbed dose in Gy, with no upper safety limit — higher tumor dose is desired, and index-lesion dose is often used to gauge whether a cycle is delivering adequate tumoricidal effect
This ledger is updated after every cycle for the remainder of the treatment course, and it is this row-by-row accumulation — not any single cycle's dosimetry — that ultimately determines when therapy must be modified or stopped.
Unlike a single fraction of external beam radiotherapy, where each treatment is planned against a pre-defined total course dose, radioligand therapy is often dosed empirically at a fixed activity per cycle (e.g. 7.4 GBq every 6–8 weeks) with the cumulative organ burden only becoming clear as cycles proceed. Each infusion adds its own dose contribution on top of everything already delivered — kidneys and marrow do not "reset" between cycles, even though most of the acute radiation effect from any single cycle resolves within weeks.
After each cycle, the treating team updates a cumulative dose table: this cycle's kidney, marrow, and tumor doses are added to the running total from all prior cycles. Conceptually this resembles a stacked bar chart — each new cycle contributes a new colored segment sitting on top of the segments from previous cycles, and the total height of the stack is what gets compared against the fixed lifetime constraint.
Cumulative kidney dose after n cycles ≈ Σ(D_kidney,i) for i = 1 to n
Because administered activity is usually held constant cycle-to-cycle (unless a prior cycle's dosimetry or toxicity prompted a reduction), the per-cycle kidney contribution is often fairly stable — meaning cumulative dose grows in a nearly linear staircase across the course, making the trajectory toward the 23 Gy constraint predictable well before it is reached, if dosimetry is tracked prospectively rather than only in retrospect.
The 6–8 week interval between Lu-177-PSMA-617 cycles is not arbitrary — it is chosen to let bone marrow, the most acutely radiosensitive tissue in the regimen, recover blood counts before the next infusion. Grade 3–4 hematologic toxicity (per CTCAE) is monitored at each cycle via complete blood count, and cycles are typically delayed or dose-reduced if platelet count, hemoglobin, or absolute neutrophil count have not adequately recovered.
But recovery of blood counts is not the same as recovery of cumulative absorbed dose. A patient's platelets may return to normal between cycles 2 and 3, giving the appearance of a fully "reset" system — while the kidney's cumulative absorbed dose, which reflects essentially irreversible microvascular and glomerular injury accumulating with each cycle, continues to climb regardless of how well blood counts recover. This distinction — reversible marrow suppression that resolves between cycles versus cumulative, largely irreversible organ dose that does not — is central to why dose tracking must be cumulative rather than assessed one cycle at a time.
A patient who tolerates every individual cycle well, with normal blood counts and no acute symptoms at each visit, can still be silently accumulating kidney dose toward an irreversible threshold — cumulative dosimetry is the only way to catch this before clinical renal impairment appears.
At most experienced RLT centers, full quantitative post-therapy dosimetry (repeat SPECT/CT imaging at multiple timepoints) is performed at Cycle 1 and at least one additional cycle to establish an individual dose-per-GBq conversion factor, then simplified dosimetry (fewer imaging timepoints, or activity-based extrapolation using the established per-patient conversion factor) is used for subsequent cycles to reduce patient burden while still updating the cumulative ledger. Centers with less dosimetry infrastructure may track cumulative dose using population-average per-cycle dose estimates from trial data (e.g. VISION's reported means) rather than patient-individualized measurements — a pragmatic but less precise approach that can under- or over-estimate an individual patient's true cumulative burden.
Cumulative organ dose tracking answers "how much risk has accumulated?" — but the decision to continue treatment also requires answering "how much benefit is this patient still getting?" After 2–3 cycles, PSA kinetics and PSMA PET/CT restaging assess tumor response, and this response feeds directly back into whether continued dose accumulation in the kidneys and marrow remains justified.
Prostate-Specific Antigen (PSA) remains the most practical interim biomarker for tracking response during a radioligand therapy course, measured before each cycle. The Prostate Cancer Working Group 3 (PCWG3) criteria define categorical response:
• PSA50 response: ≥50% decline from baseline — the threshold most consistently associated with improved overall survival in trials including VISION • PSA30 response: ≥30% decline — a less stringent but still prognostically meaningful threshold • Confirmed progression: PSA rise ≥25% above nadir (the lowest value reached during treatment), confirmed by a second rise at least 3 weeks later, to distinguish true progression from normal assay or biological fluctuation
Because PSA can occasionally show a transient early rise ("PSA flare") before declining in patients ultimately responding well, single-timepoint PSA increases are interpreted cautiously and generally require imaging confirmation before treatment is stopped for presumed progression.
Alongside PSA, interim PSMA PET/CT (using a diagnostic tracer such as Ga-68-PSMA-11 or F-18-PSMA-1007) provides structural and molecular confirmation of response: index lesions are measured for size and PSMA-avidity (SUVmax) at baseline and again at interim restaging, analogous to RECIST 1.1 criteria but incorporating the biological information of tracer uptake rather than size alone. A drop in lesion PSMA-avidity — even before any measurable size reduction — can indicate biological response and reduced target availability for further therapeutic uptake, which is itself clinically informative: it means later cycles may deliver progressively less tumor dose per unit activity as PSMA expression falls, even as organ-at-risk doses continue to accumulate at a comparatively stable rate.
New lesions not present at baseline, or unequivocal growth of existing lesions with rising PSMA avidity, define structural progression and are weighted heavily in the continue/stop decision, particularly when combined with a rising PSA.
The clinically dangerous scenario is discordance: falling tumor PSMA-avidity (reduced potential future benefit) occurring at the same time as rising cumulative kidney dose (rising future risk) — this crossing of the risk-benefit curves is exactly the situation the interim restaging checkpoint is designed to catch before it is missed.
Interim restaging results are combined with the cumulative dose ledger from Stage 2 into a single continuation decision, typically made by a multidisciplinary team (nuclear medicine physician, medical oncologist, radiation safety physicist):
• Responding + dose well under limits → continue at full planned activity through remaining cycles • Responding + dose approaching limits → continue but flag for activity reduction in upcoming cycles • Stable/mixed response + dose under limits → continue with closer interim monitoring • Progressing (PSA and/or PSMA PET) → discontinue RLT regardless of remaining dose headroom, since further cycles would add organ risk without corresponding benefit, and alternative therapy should be considered
This is the point in the treatment course where "how many cycles are planned" (the slider control in this simulation) is most likely to be revised from its original value — a patient originally planned for 6 cycles might be stopped at 3 for progression, or a patient with an excellent response and ample dose headroom might be extended beyond an initially conservative plan.
As a patient proceeds through the later cycles of a planned course, the cumulative kidney dose bar climbs steadily toward the 23 Gy operational constraint most centers use as their hard ceiling. Once cumulative dose crosses roughly 80% of that limit, every subsequent cycle triggers a formal decision gate — continue at full activity, reduce the administered activity, or stop therapy altogether — rather than being administered on autopilot.
The 23 Gy whole-kidney cumulative dose limit widely used in RLT practice today is not derived from radioligand-specific outcome data — it is extrapolated from the Emami et al. (1991) tolerance dose tables for conventional external beam radiotherapy, which estimated a 5% risk of clinically significant nephropathy within 5 years (TD5/5) at a whole-kidney dose of roughly 23 Gy delivered in ~2 Gy fractions. This external beam figure was adopted as a conservative surrogate ceiling for internal emitter therapy in the absence of large RLT-specific dose-toxicity datasets.
Because Lu-177-based RLT delivers dose at a continuously decreasing low dose-rate over days, rather than in discrete high-dose-rate fractions, the underlying radiobiology differs meaningfully from external beam — low dose-rate irradiation generally allows more sublethal damage repair between "effective fractions," which is the rationale behind newer biologically-effective-dose (BED) adjusted models suggesting the kidney may tolerate substantially more — often cited in the 28–40 Gy BED range — before reaching comparable nephropathy risk. Clinical practice remains conservative and centers vary in which threshold they adopt, but 23 Gy remains the most commonly cited operational ceiling in guidelines and trial protocols.
At each cycle once cumulative kidney dose enters the warning zone, the treating team works through a structured decision:
Continue at full activity — appropriate when: cumulative dose remains comfortably under the ceiling even after projecting the next cycle's contribution, renal function (eGFR, serum creatinine) is stable, and tumor response remains favorable.
Reduce administered activity — appropriate when: projected cumulative dose after the next full-activity cycle would approach or cross the ceiling, but the patient is still deriving clear tumor benefit and organ function is preserved; reducing administered GBq per cycle (commonly by 20–30%) slows the rate of dose accumulation while allowing continued treatment, trading some tumor dose for extended safety margin.
Discontinue therapy — appropriate when: cumulative dose has reached or is projected to exceed the ceiling with the next cycle, renal function has meaningfully declined (e.g. eGFR drop >30–40% from baseline), grade 3–4 hematologic toxicity has occurred, or tumor progression has already been documented at interim restaging — in which case dose headroom becomes moot.
This decision is remade at every remaining cycle, not made once — a patient reduced to a lower activity at cycle 5 still has their new, slower dose-accumulation trajectory re-evaluated before cycle 6.
Kidney dose alone rarely forces discontinuation before marrow toxicity or disease progression intervenes in most published series — but for patients receiving 6 full-activity cycles with above-average per-cycle kidney uptake, cumulative dose can realistically approach the 23 Gy ceiling, which is exactly why prospective cycle-by-cycle projection, not retrospective tallying, is the safeguard.
While the kidney ceiling is usually the more dramatic long-horizon constraint, cumulative bone marrow dose (limit ~2 Gy) and its associated hematologic toxicity is frequently the constraint that intervenes first, cycle-to-cycle, in real practice — because marrow suppression manifests acutely (thrombocytopenia, anemia, neutropenia visible on routine bloodwork at each visit) rather than as a silently accumulating structural dose. CTCAE grade 3–4 hematologic toxicity is the single most common reason for dose reduction or treatment delay in published Lu-177-PSMA-617 cohorts, occurring in roughly 10–15% of VISION trial patients, more often than kidney-driven modifications. Both constraints are tracked in parallel at every cycle, with whichever is closer to its ceiling driving the decision.
When the planned course concludes — whether at the originally intended cycle count, or earlier due to progression, toxicity, or dose-limit constraints — the final step is a formal end-of-treatment dosimetry summary: every organ-at-risk's cumulative absorbed dose is tabulated against its lifetime constraint and flagged as within-limit or exceeded. This record then becomes the foundation for a multi-year follow-up surveillance plan, because the risks of cumulative internal radiation exposure do not end when the last infusion is given.
At course completion, the cumulative dose ledger maintained since Cycle 1 is finalized into a summary record: cumulative kidney dose (Gy) versus the 23 Gy operational ceiling, cumulative red marrow dose (Gy) versus the 2 Gy ceiling, cumulative dose delivered to each tracked index tumor lesion, and the total number of cycles actually completed versus originally planned. Each organ-at-risk is flagged — within constraint, or exceeded — and any exceedance is documented with its clinical context (e.g., dose reduction implemented at cycle 5, discontinued one cycle early for renal function decline).
This record travels with the patient's oncology chart indefinitely: because Lu-177-PSMA-617 may not be the last radionuclide therapy a patient ever receives (some patients are later considered for re-treatment courses, or for other radiopharmaceuticals), the cumulative dose ledger from this course becomes the starting baseline — not a reset to zero — for any future radionuclide therapy decision.
Because both kidney and marrow injury from internal emitter therapy can manifest with a delay — subclinical nephron loss progressing over years, or clonal hematopoietic changes emerging long after the acute cytopenias of active treatment resolve — end-of-treatment does not mean end of monitoring:
• Renal function: serial eGFR and serum creatinine measurements, typically every 3–6 months for at least the first 1–2 years post-treatment and periodically thereafter, watching for a gradual decline pattern distinct from any acute on-treatment fluctuation • Hematologic recovery and surveillance: complete blood counts at similar intervals, watching not just for count recovery but for any emerging cytopenias or abnormal blood smear findings that could signal therapy-related myelodysplastic syndrome (MDS) — a rare but recognized late effect of cumulative marrow irradiation from radionuclide therapy, reported in a small minority of long-term survivors across radioligand and radioimmunotherapy literature • Salivary and lacrimal gland function: symptom-based monitoring for xerostomia, which can be a persistent quality-of-life issue even though it is rarely dose-limiting for the treatment course itself
The clinical logic of this entire tracking framework is continuity: a fixed lifetime organ constraint only has meaning if every cycle's contribution is added to a single, uninterrupted running total, from Cycle 1 dosimetry through the final follow-up visit years later — never re-zeroed, never assessed in isolation.
Not every patient completes the originally planned cycle count, and the end-of-treatment summary looks different depending on why the course ended:
• Completed as planned: full planned cycles delivered, doses tracked throughout, no organ constraint approached — the most common outcome for patients tolerating therapy well with a full response • Stopped early for dose-limit approach: cumulative kidney or marrow dose reached the decision-gate threshold before all planned cycles were given; remaining planned cycles are foregone in favor of preserving organ function, even if tumor response was still favorable • Stopped early for progression: interim or ongoing restaging showed the tumor was no longer responding, so remaining dose headroom became irrelevant to the treatment decision • Stopped early for toxicity: grade 3–4 hematologic or renal adverse events forced discontinuation independent of the cumulative dose ledger's numeric status
In VISION, the median RLT-arm patient received 5 of the 6 planned cycles — illustrating that early discontinuation for one of these reasons is a common, expected part of real-world dosing courses, not a departure from protocol.