Top Line: Can you trust a stable PSA in ruling out progression in patients on enzalutamide?
The Study: The short answer is perhaps not as much as you might think. This post hoc analysis of two phase III trials, ARCHES (1,150 patients with mHSPC) and PROSPER (1,401 patients with nmCRPC), examined whether radiographic progression (rPD) occurred with or without PSA rise during enzalutamide therapy. Among patients who developed rPD on ezalutamide, a striking 25% in ARCHES and 22% in PROSPER had no PSA rise whatsoever. That’s compared to just 7% and 4% of patients in each trial who were on ADT and placebo with no rise in PSA at the time of rPD. The discordance was even more dramatic when looking at PCWG defined progression (>25% above starting PSA and >2ng/mL above nadir). Over half of patients (62% ARCHES and 51% PROSPER) who had rPD did not meet PCWG criteria for PSA progression. The biological rationale is that potent AR blockade selects for AR-independent clones and lineage plasticity (including neuroendocrine transformation) that may grow without producing PSA. This is reflected in the pattern of metastatic spread where liver metastases (a hallmark of AR-independent disease) were over 5 times more common in enzalutamide patients with rPD compared with controls (11.4% vs 2.1% in ARCHES; 5.3% vs 0.9% in PROSPER). Most rPD events were asymptomatic when they were caught, however that doesn’t mean they were indolent as rPD with or without PSA changes carried significantly worse OS compared to no progression. High-volume disease and high Gleason scores at baseline predicted rPD overall, but nothing reliably predicted who would progress without a PSA signal. Similar findings were reported from the TITAN trial with apalutamide, suggesting this may be a class effect of potent ARPIs, not unique to enzalutamide.
TBL: Among patients on enzalutamide, as many as half of radiographic progression events do so without significant rise in the PSA, and as many as a quarter have no rise in PSA at all. So, don’t ignore imaging just because the PSA is well controlled. | Armstrong, J Clin Oncol 2026
The Study: The short answer is perhaps not as much as you might think. This post hoc analysis of two phase III trials, ARCHES (1,150 patients with mHSPC) and PROSPER (1,401 patients with nmCRPC), examined whether radiographic progression (rPD) occurred with or without PSA rise during enzalutamide therapy. Among patients who developed rPD on ezalutamide, a striking 25% in ARCHES and 22% in PROSPER had no PSA rise whatsoever. That’s compared to just 7% and 4% of patients in each trial who were on ADT and placebo with no rise in PSA at the time of rPD. The discordance was even more dramatic when looking at PCWG defined progression (>25% above starting PSA and >2ng/mL above nadir). Over half of patients (62% ARCHES and 51% PROSPER) who had rPD did not meet PCWG criteria for PSA progression. The biological rationale is that potent AR blockade selects for AR-independent clones and lineage plasticity (including neuroendocrine transformation) that may grow without producing PSA. This is reflected in the pattern of metastatic spread where liver metastases (a hallmark of AR-independent disease) were over 5 times more common in enzalutamide patients with rPD compared with controls (11.4% vs 2.1% in ARCHES; 5.3% vs 0.9% in PROSPER). Most rPD events were asymptomatic when they were caught, however that doesn’t mean they were indolent as rPD with or without PSA changes carried significantly worse OS compared to no progression. High-volume disease and high Gleason scores at baseline predicted rPD overall, but nothing reliably predicted who would progress without a PSA signal. Similar findings were reported from the TITAN trial with apalutamide, suggesting this may be a class effect of potent ARPIs, not unique to enzalutamide.
TBL: Among patients on enzalutamide, as many as half of radiographic progression events do so without significant rise in the PSA, and as many as a quarter have no rise in PSA at all. So, don’t ignore imaging just because the PSA is well controlled. | Armstrong, J Clin Oncol 2026