PROTEUS substudy: perioperative apalutamide + ADT vs surgery alone gives EFS HR 0.38, but much of the endpoint was measured while the drug effect had not yet worn off

Data as of 2026-10-29 (Taipei time). [Fact] statements are sourced; [Inference] statements are this article's judgment.

1. One-line conclusion

[Fact] The PROTEUS substudy (ESMO [European Society for Medical Oncology] 2026 LBA53 [late-breaking abstract], presented 2026-10-24) compared perioperative apalutamide + ADT (androgen deprivation therapy) combined with radical prostatectomy and pelvic lymph node dissection vs surgery alone: EFS (event-free survival) HR (hazard ratio) 0.38 (95% CI [confidence interval] 0.26–0.55; p<0.0001), 3-year EFS 63% vs 41.4% (J&J).

[Inference] Primary-endpoint events were mostly PSA (prostate-specific antigen) biochemical failures, while testosterone took a median 20.7 months to recover (UroToday); part of HR 0.38 was measured while hormones were still suppressed. It proves delayed recurrence; it does not yet prove that more patients are cured.

For sponsors: for a PSA-driven EFS measured during ongoing or residual castration, regulators are likely to ask to see MFS (metastasis-free survival) and landmark analyses after testosterone recovery as well; when designing perioperative hormonal trials, both should be prespecified in the statistical analysis plan.

2. What happened

[Fact] Design: randomized, open-label; 408 patients were allocated 1:3 to apalutamide 240 mg/day + ADT (6 neoadjuvant 28-day cycles and 6 adjuvant cycles, stopped 2 weeks before surgery and resumed 4 weeks after; n=102) or immediate surgery (n=306); after surgery the control arm was observed or given local/systemic adjuvant therapy per physician and institutional practice (J&J, UroToday). EFS was defined as biochemical failure (two consecutive post-operative PSA rises, the second ≥0.2 ng/mL), local/regional recurrence, distant metastasis or death; imaging was by PSMA PET (prostate-specific membrane antigen positron emission tomography) or conventional imaging with blinded central review. The design assumed HR 0.60 with 80% power (UroToday). Enrollment ran from 2022-07 to 2023-10 at 83 sites in 15 countries/regions (Urology Times).

[Fact] Results (median follow-up 37.8 months): EFS HR 0.38; median EFS not reached vs 19.1 months (J&J). First events 35/102 (34.3%) vs 172/306 (56.2%), including biochemical failure in 27 vs 92 and local/regional recurrence detected by PSMA PET or conventional imaging in 4 vs 51 (UroToday). MFS HR 0.49 (0.26–0.90; nominal p=0.0193); time to first subsequent therapy HR 0.29 (0.20–0.43), 3-year freedom from subsequent therapy 64.6% vs 32% (J&J). OS (overall survival) follow-up is ongoing (J&J). Safety: grade ≥3 TEAEs (treatment-emergent adverse events) 31% vs 22%, no fatal TEAEs; hot flushes 64% (J&J).

For sponsors: median EFS in the "surgery alone" arm in high-risk localized prostate cancer was only 19.1 months (J&J), so events accrue quickly in EFS-driven trials; but this benchmark depends heavily on the PSA failure rule and on how often PSMA PET is used.

3. Why it succeeded

[Inference] Mechanism: ADT + an ARPI (androgen receptor pathway inhibitor) directly suppresses PSA, so biochemical-failure events are inherently unlikely during treatment and before testosterone recovers. Treatment was 12 cycles, about 11 months; median time to testosterone recovery to ≥200 ng/dL was 20.7 months, about 9 months after stopping treatment (UroToday). The presenter, Briganti, also acknowledged that testosterone suppression may delay PSA rise and PSMA PET detection (Urology Times interview).

[Inference] The key evidence for reading the endpoint is the landmark analysis: the EFS-rate gap was 42 percentage points at month 20, 37 at month 24 and 22 at month 36 (all nominal p<0.001) (Urology Times). The company and presenter read "a gap that persists after testosterone recovery" as durable benefit (UroToday); but the gap roughly halved within 16 months, which also fits a pattern of "events delayed, then catching up after testosterone recovers". Telling the two readings apart needs another 2–3 years of follow-up. In addition, biochemical failure made up 77.1% (27/35) of first events in the treatment arm vs 53.5% (92/172) in the control arm (UroToday), so the treatment arm's events were more concentrated in PSA, a hormone-sensitive marker.

[Inference] Detection bias: among control-arm events detected only by PSMA PET without biochemical failure, about 70% still had PSA ≥0.1 ng/mL 3 months after surgery (persistent PSA), which looks more like incomplete resection than recurrence; in the treatment arm, preoperative ADT makes persistent PSA harder to see. After excluding, in both arms, local/regional recurrence or distant metastasis events detected only by PSMA PET without prior biochemical failure, EFS HR became 0.50 (0.33–0.76) (UroToday): the effect remains, but is smaller than the headline 0.38. PSMA PET was not done at screening (UroToday), so some "metastasis events" may be micrometastases already present at baseline, which also affects MFS.

[Inference] Comparator: this substudy compares "apalutamide + ADT" with "nothing added", so it cannot separate the contributions of apalutamide and ADT; the main study (apalutamide + ADT vs placebo + ADT) showed a smaller effect, reported as a 20% reduction in the risk of metastasis or death and a 29% reduction in the risk of recurrence or death (UroToday). With open-label design and 1:3 randomization, the treatment arm had only 102 patients, and the upper bound of the MFS confidence interval, 0.90, is close to 1 (J&J). The definition of time to first subsequent therapy includes restarting ADT, and adjuvant therapy in the control arm was at physician discretion (post-operative radiotherapy: 20% in the treatment arm vs 43% in the control arm) (UroToday), so HR 0.29 is also shaped by clinical decisions made under open-label conditions.

For sponsors: perioperative hormonal trials using EFS should prespecify a landmark analysis starting from testosterone recovery, a sensitivity analysis excluding events detected only by PSMA PET without prior biochemical failure, and baseline PSMA PET; otherwise the HR will be challenged on whether it measures drug exposure or treatment benefit.

4. Regulatory and commercial implications

[Fact] Apalutamide + ADT is not approved by any regulator in high-risk localized or locally advanced prostate cancer; J&J's press release says the paper has been accepted by European Urology (J&J), but Dana-Farber's press release the same day (2026-10-24) says it was published in European Urology that day (Dana-Farber), and UroToday also says it was published simultaneously with the congress (UroToday); the paper's DOI is 10.1016/j.eururo.2026.10.016 (European Urology), and we could not open the full text to check the form of publication (online-first or issue). The main study's registry lists enrollment of 2,517, with pCR (pathologic complete response) and MFS as primary endpoints and estimated primary completion in 2026-12 (NCT03767244). We could not find whether J&J has filed this indication with the FDA (US Food and Drug Administration) or the EMA (European Medicines Agency).

[Inference] On the regulatory side, the FDA has consistently emphasized MFS and OS in localized prostate cancer; in this substudy MFS is a secondary endpoint with a nominal p-value (J&J), so the substudy's EFS alone is unlikely to support approval, and the main study's MFS will be the main basis. Commercially, whether urologists are willing to delay surgery by about 6 months for neoadjuvant treatment is the real adoption threshold; this cannot be inferred directly from the data.

For sponsors: in high-risk localized prostate cancer, choosing MFS as the primary endpoint is the reliable regulatory path; EFS suits an early readout or a secondary endpoint, not a filing on its own.

5. Recommendations for the next trial and OncoLattice

[Inference] First, when designing a perioperative ARPI trial, decide whether the control arm is "surgery + ADT" or "surgery alone"; they answer different questions: the former tests the added value of the ARPI (the PROTEUS main-study design, NCT03767244), the latter the value of the whole hormonal package (this substudy's design, J&J). Second, the protocol should handle testosterone recovery explicitly for an EFS endpoint: in this substudy, median time to testosterone recovery to ≥200 ng/dL was 20.7 months (UroToday); consider requiring testosterone recovery before calling biochemical failure, or relying mainly on PSMA PET-confirmed imaging events. Third, do PSMA PET at baseline: this substudy did not do PSMA PET at screening, which the investigators themselves list as a limitation (UroToday); baseline imaging reduces stage migration and makes MFS more credible. Fourth, OncoLattice site strategy: this substudy required 6 cycles of neoadjuvant and another 6 cycles of adjuvant treatment (J&J), imaging events were read by blinded central review of PSMA PET (UroToday), and enrollment was completed across 83 sites in 15 countries/regions (Urology Times); similar trials should prioritize high-volume surgical centers where urology and medical oncology can enroll jointly and PSMA PET is accessible.

For OncoLattice: for perioperative trials in high-risk localized prostate cancer, site assessment should look at surgical volume, PSMA PET access and the referral pathway for neoadjuvant treatment together; missing any one will slow enrollment.

6. Open questions and what to watch

Event	Expected timing	Specific source
Substudy OS and longer-follow-up MFS	J&J says only that follow-up is ongoing, with no date; the registry's estimated overall study completion is 2028-10-13 (no separate date for the substudy)	J&J press release (J&J); registry page NCT03767244
Numbers at risk and confidence intervals at each landmark, definition of persistent PSA	Published simultaneously with the congress on 2026-10-24 (per Dana-Farber and UroToday); we could not open the full text	European Urology paper (DOI 10.1016/j.eururo.2026.10.016); Dana-Farber
Primary completion of the main study (apalutamide + ADT vs placebo + ADT)	Registry estimate 2026-12	Registry page NCT03767244
Whether FDA/EMA filings have been made	Not disclosed	The only official statement so far is J&J's press release, which says it is not approved by any regulatory authority (J&J); watch the J&J newsroom

For sponsors: until OS is available, perioperative apalutamide should not be treated as the standard comparator in high-risk localized prostate cancer trials.

The next thing to watch is the EFS and MFS gap more than a year after testosterone recovery: if the gap keeps narrowing, HR 0.38 mainly reflects delayed events; only if it holds steady does it support a higher cure rate.

Evidence limitations

Efficacy and safety figures here come from J&J's press release (J&J); event composition, testosterone recovery, the sensitivity analysis and subsequent therapy come from UroToday's congress report (UroToday); the landmark gaps and the number of enrolling sites come from Urology Times (Urology Times); the presenter's comments on testosterone suppression come from the Urology Times interview (Urology Times interview). We did not obtain the LBA53 slides and could not open the European Urology full text (DOI), so the numbers of patients and confidence intervals at each landmark are not available.

Advice for readers: when citing EFS HR 0.38, also cite HR 0.50 (0.33–0.76) after excluding events detected only by PSMA PET without prior biochemical failure (UroToday), and the median 20.7 months to testosterone recovery (UroToday), so readers can judge the effect size.

Note: differences from the original relay

EFS HR 0.38 (0.26–0.55), 3-year EFS 63% vs 41.4% and n=408 match J&J's press release (J&J). On the relay's "J&J press release date unconfirmed": the press release is dated 2026-10-24 (US time), the same day as the ESMO presentation (J&J). The relay mentioned skepticism on X about how to read EFS; this article does not cite X posts directly and instead discusses the presenter interview and the landmark data.