PSMA PET Natural History Study in PSMA-Positive BCR
ForBiochemically recurrent prostate ca, post-definitive + salvage RT, PSA ≥ 0.5
TL;DRBaseline data on first 130 pts: most BCR men with PSADT >9-12mo still have PSMA PET findings once PSA exceeds 5.
Reported via UroToday →
The RT-relevant point is where the disease sits: prostate bed recurrence and nodal disease are the dominant baseline patterns, all in men who already had or declined salvage RT. If most slow-PSADT pts light up above PSA 5, PET positivity alone cannot gate metastasis-directed SBRT, and the trigger has to come from kinetics or serial change.
In post-salvage-RT BCR with PSADT over 9mo and a PSMA PET showing a few nodal or bed lesions, this supports serial imaging over reflex systemic therapy; it says nothing about pts with rapid PSADT or conventional-imaging metastases.
Nodal disease and prostate bed recurrence dominate the baseline patterns in men already past salvage RT, so this is the population MDT gets offered to. If most slow-PSADT pts are PET-positive above PSA 5, lesion count reflects scan timing more than biology, which weakens PET positivity as the gate for SBRT.
Only about a third of the cohort has started any therapy, and only 5 of the first ~150 progressed on conventional imaging at ~1.5yr. That is the counterweight to reflex mCSPC-style doublet therapy triggered by PSMA PET findings in men whose PSA kinetics are slow.
11 details
Prospective observational natural history cohort at the NCI, presented at GU-ASCO 2026. Baseline data on the first 130 pts; the cohort is ongoing at a median follow-up of about 2 years with interim outcome data pending.
Men with biochemical recurrence after definitive therapy, all of whom had already received salvage radiation or declined it. PSA ≥ 0.5 required for a baseline PET. Deliberately framed as PSMA-positive BCR, a state that would not have been detectable in the historical trials these pts map onto and that excluded them from metastatic trials.
The presented analysis is cross-sectional: PSMA PET findings at baseline against PSA doubling time, the field's working predictor of metastasis in BCR. Imaging cadence is protocolized at annual if the baseline PET is negative, every 6 months if anything is seen, including equivocal findings.
Among pts with slow kinetics (PSADT > 9mo and > 12mo), the majority still had PSMA PET findings once PSA was above 5: nodal disease, prostate bed recurrence, and small equivocal bone lesions. Only about a third have started any therapy. A companion poster found 5 of the first ~150 pts progressed on conventional imaging at roughly 1.5yr, and over 95% showed no abrupt PET change out of step with PSA.
The counts are reported conversationally in an interview ("about 130", "the first 150 or so", "that number's five"), with no stratum denominators and no CIs, so the PSADT-by-PET relationship cannot be quantified from this source. Allowing any therapy under 6 months, including SBRT and intermittent ADT, means the observed group is not an untreated comparator.
The claim being staked is that PSMA PET positivity in BCR is near-ubiquitous above PSA 5 and therefore weakly discriminating, so it should not by itself trigger treatment. What the cohort has not yet shown is whether PET burden or its change over time adds anything to PSA doubling time for predicting the outcomes that matter.
Baseline cross-sectional read of an ongoing single-institution cohort, median f/u ~2yr, no outcome analysis yet. Numbers reported conversationally, not tabulated.
- Does PET burden or its change add to PSA doubling time for predicting progression?
- Which PSMA-positive BCR pts can safely be observed off therapy?
- Are small equivocal bone lesions on PSMA PET true metastases?
📚 Sources · 📄 1 paper
Abstract
The longer read
The value of this cohort is not in its numbers, which are thin and conversationally reported, but in the question it is built to answer, and that question is one the field has been avoiding since PSMA PET became widely reimbursed. Stage migration created a population that has metastatic-appearing disease on a scan and biochemically recurrent disease by every other measure, including the historical trial evidence that governs their prognosis. The practical consequence has been that men who would have been observed for years are now being started on doublet or triplet regimens on the strength of a few avid nodes. Nobody randomized anyone to establish that this helps.
What this analysis contributes is a denominator argument. If men with PSA doubling times beyond 9 and beyond 12 months, the exact group the risk calculators call low-risk, mostly have something on PSMA PET once their PSA clears 5, then a positive scan is not a discriminating test in this setting. A test that is positive in most of the population it is applied to cannot sort that population, whatever its sensitivity. The corollary is that the trigger for therapy has to remain PSA kinetics, or serial change on imaging, rather than the binary fact of PET positivity. The reported observation that over 95% of pts showed no abrupt PET change discordant from their PSA behavior points the same direction: the scan is largely tracking what the PSA already told you.
For a radiation oncologist the interesting detail is anatomic. The described patterns, nodal disease and prostate bed recurrence, are in men who already had salvage radiation or declined it, which is precisely the population where metastasis-directed therapy gets offered. The ORIOLE and STOMP experience established that SBRT to oligometastatic sites delays progression, but those trials selected pts on the basis of a limited number of detectable lesions. This cohort suggests that the number of detectable lesions is a function of when you scan and at what PSA, not a stable biological property, which complicates any framework that gates treatment on lesion count.
The design decisions cut against a strong read. Permitting any therapy shorter than 6 months, including SBRT, intermittent ADT, intermittent ARPI, and Radium-223 trial participation, means the observed group is not a natural history control in the strict sense; roughly a third have had something. A single NCI referral center enrolling men willing to be watched selects for exactly the indolent phenotype the analysis then reports. And the follow-up, around 2 years, is short against a disease state where the relevant interval is measured in the many years the investigator herself invokes.
What would make this change practice is the follow-up dataset, not the baseline one: whether baseline PET burden, or its trajectory, predicts conventional-imaging progression better than PSA doubling time alone. If it does not, the case for treating scan findings as an independent indication collapses, and observation returns as a defensible option in a group where it has quietly stopped being offered. The design was pre-EMBARK, which is worth holding in mind, because EMBARK has since made enzalutamide plus ADT the competing default for high-risk nonmetastatic BCR, and this cohort is testing whether that default is being applied to men who do not need it.