Sequencing Pressure Mounts After ARPI Exposure in mCRPC
ARPI intensification in metastatic hormone-sensitive prostate cancer has changed the composition of metastatic castration-resistant prostate cancer, and the next therapeutic decision is now more complicated than a simple swap from one drug class to another. Men arriving at mCRPC after early androgen receptor pathway inhibitor exposure often carry treatment-selected resistance, which can blunt the value of later ARPI reuse and force clinicians to choose between taxanes, radioligand therapy, PARP inhibition, or emerging antibody-drug conjugates on the basis of incomplete comparative data.
No prospective sequencing trial has answered the central question. In place of a direct head-to-head roadmap, the evidence base rests on trials built for efficacy within a single line of therapy, not on the order in which therapies are delivered across multiple lines.
Cabazitaxel Still Anchors the Strongest Sequencing Signal
Among the available datasets, CARD remains the clearest sequencing trial in routine practice. Patients previously treated with docetaxel and an ARPI derived an overall survival advantage with cabazitaxel over an ARPI switch, with an OS hazard ratio of 0.64. In practical terms, the study punished the reflex to recycle the same androgen receptor class after documented resistance had already emerged.
The biological logic is straightforward. Once androgen receptor signaling has been therapeutically pressured for months or years, tumor clones often adapt through receptor amplification, ligand bypass, or lineage plasticity. A taxane such as cabazitaxel attacks a different vulnerability, making a class switch more credible than a second ARPI in many post-docetaxel patients.
PARP Inhibition Has the Cleanest Survival Signal in Selected Genotypes
For homologous recombination repair–deficient disease, the survival story is stronger, but not uniform across all genomic groups. PROfound established alpha-controlled overall survival benefit in cohort A, with an HR of 0.69. TALAPRO-2 also showed alpha-controlled OS benefit in the all-comer population, with an HR of 0.80.
The genotype matters because PARP inhibition exploits synthetic lethality in tumors that cannot repair DNA damage efficiently. BRCA-altered cancers remain the most biologically compelling group, yet the larger BRCA-specific estimates cited in this review are exploratory rather than definitive. In TRITON3, BRCA overall survival was not improved, with an HR of 0.91. That result tempers enthusiasm for assuming that every BRCA-enriched population will translate a progression endpoint into a clear survival gain.
The signal is strong enough to guide discussion, but not strong enough to justify therapeutic tunnel vision. Tissue context, prior therapy, disease tempo, and the availability of other active classes still shape whether a PARP inhibitor should come first, follow a taxane, or be reserved for a later line.
PSMA-Targeted Radioligand Therapy Improves rPFS, But OS Remains Clouded by Crossover
The radioligand data are equally persuasive on disease control and less convincing on survival so far. 177Lu-PSMA-617 improved radiographic progression-free survival versus an ARPI switch in PSMAfore, which supports the idea that PSMA-selected disease can be controlled more effectively by targeted beta-emitting therapy than by simply changing one androgen receptor blocker for another.
Overall survival, however, did not differ in the intention-to-treat analysis, with an HR of 0.91. Roughly 60% crossover blurred the comparison and makes the OS estimate difficult to interpret as a definitive absence of benefit. In a setting where most patients can eventually receive the alternative therapy, post-progression survival often dilutes the contrast between arms and pushes the field to rely on surrogate endpoints while waiting for cleaner sequencing studies.
No taxane-controlled trial has yet shown an OS advantage for radioligand therapy. That leaves an unresolved tension between strong biological rationale, favorable rPFS, and the absence of a validated survival edge against chemotherapy in a head-to-head sequence.
PLUDO Raises the Possibility That Early Docetaxel Still Has a Role
Preliminary PLUDO data complicate the enthusiasm for moving directly to radioligand therapy in every patient. In that report, OS favored docetaxel first, with an HR of 1.64. The direction of effect suggests that earlier chemotherapy may still matter in selected patients, although asymmetric crossover could have distorted the estimate.
Even with that caveat, the finding cannot be dismissed as noise. A taxane-first strategy may preserve later-line options while avoiding the risk that a radioligand is deployed after the disease has already become too heterogeneous or clinically fragile to benefit fully from a targeted approach.
One sentence captures the practical problem: sequence can change the biology of what is left to treat.
A Four-Track Framework Based on Genomics and PSMA Expression
The review proposes a hypothesis-generating four-track framework anchored to homologous recombination repair status and PSMA-PET expression. This is a sensible attempt to replace the current habit of one-size-fits-all sequencing with a more anatomically and molecularly disciplined approach. PSMA-positive, HRR-intact disease points toward radioligand therapy; HRR-altered disease points toward PARP inhibition; overlap cases may require closer attention to prior exposure, disease burden, and the pace of progression. Patients lacking either actionable feature may still be best served by chemotherapy or other non-targeted strategies.
The appeal of the framework lies in restraint. It does not pretend that a single biomarker can settle the entire treatment order in mCRPC. Instead, it acknowledges that resistance after ARPI exposure is not a single disease state but a collection of evolving phenotypes, each with different therapeutic liabilities.
What Clinicians Can Use Now, and What Still Needs a Trial
For now, treatment choice remains individualized. Cabazitaxel has the most persuasive sequencing evidence after docetaxel and ARPI exposure. PARP inhibitors have credible survival support in genomically selected populations, especially where homologous recombination repair defects are present, but the magnitude of benefit varies by molecular subgroup and prior therapy. Radioligand therapy offers an rPFS advantage in PSMA-selected disease, yet the survival effect remains harder to isolate because crossover and the lack of a taxane-controlled OS signal leave the comparative sequence unsettled.
The unanswered problem is not whether these therapies work; each class clearly does in the right setting. The unresolved issue is which class should come first when a patient has already moved through ARPI intensification and now needs a second major systemic decision. Until a prospective sequencing trial appears, clinicians will keep making that call with imperfect surrogates, molecular testing, imaging phenotype, and the realities of prior treatment exposure.