Dr Narendhar Gokulanathan

Department of Medical Oncology and Hemato-Oncology,

Apollo Hospitals, Bannerghatta Road, Bangalore

Dr Vishwanath S.

Professor and Head of the Department,

Department of Medical Oncology and Hemato-Oncology

Apollo Hospitals, /bannerghatta road, Banglore

The POSEIDON (Post Operative Salvage or Adjuvant Radiotherapy and anti-androgen therapy for men with prostate cancer) individual patient data (IPD) meta-analysis represents the most granular and comprehensive evaluation to date - of hormone therapy (HT) in the postoperative (salvage) radiotherapy (PORT) setting.

Study Design

POSEIDON is an IPD-level meta-analysis of six randomized phase III trials, including GETUG-AFU 16, NRG/RTOG 0534, RADICALS-HD, and RTOG 9601. Unlike traditional meta-analyses that pool aggregate hazard ratios, this analysis used a one-stage stratified Cox proportional hazards framework, allowing direct modeling of individual interactions between PSA levels and survival outcomes (OS and metastasis-free survival).

Two statistical refinements strengthen the credibility of this meta-analysis:

• Natural cubic splines were used to model & identify non-linear associations between pre-PORT PSA and survival, avoiding arbitrary cutoffs for PSA levels and outcome effects.
• A two-stage within-trial contrast analysis was performed to decrease aggregation bias (and in-trial variations), ensuring observed effects were not driven disproportionately by any single trial.

Study Population

The analysis included 6,057 patients with a median follow-up of 9 years.

• Median pre-PORT PSA:
  1. 0.3 ng/mL in short-term HT trials
  2. 0.5 ng/mL in long-term HT trials (RADICALS and RTOG 9601)
• Gleason score 7 in ~70%
• Positive surgical margins in 54–72%
• Seminal vesicle invasion in 15–24% (worst prognostic factor)

However, the trials were conducted between 1998 and 2015, an era predating PSMA PET–CT and genomic classifiers. Thus, the biological staging relied on conventional imaging (CT and bone scan), which has important consequences regarding the interpretation of this study.

Outcomes

Overall Survival (Primary Endpoint)

Addition of hormone therapy to PORT did not significantly improve overall survival in the overall population. HR 0.87 (95% CI 0.76–1.01; p=0.06)

How did pre-PORT PSA affect OS?

A statistically significant interaction between pre-PORT PSA and OS benefit (p=0.02) emerged.

• PSA ≤ 0.5 ng/mL: No OS benefit. The number-needed-to-treat (NNT) was negative, suggesting possible harm. Routine addition of HT in the adjuvant or early salvage setting is not supported by this study
• PSA 0.51–1.0 ng/mL: NNT ~22 to prevent one death
• PSA > 1.0 ng/mL: NNT ~12.
• PSA >1.6–2.0 ng/mL: Long term HT showed significant OS benefit in this subgroup.

Metastasis-Free Survival (Secondary Endpoint)

HT significantly improved MFS - HR 0.79 (95% CI 0.70–0.89; p 0.001). However, improvement in MFS did not uniformly translate to OS benefit. For hormone therapy duration and overall survival, short-term hormone therapy had a hazard ratio of 0.93 (95% CI 0.77-1.11), whereas long-term hormone therapy had a hazard ratio of 0.79 (95% CI 0.63 – 1.00).

ICECaP (Intermediate Clinical Endpoints in Cancer of the Prostate) Surrogacy Threshold

The study applied the predefined ICECaP surrogate threshold, which requires the upper bound of the 95% CI for MFS HR to fall below 0.81 to predict OS benefit.

In early salvage (PSA ≤0.5):

• The MFS effect failed to meet ICECaP criteria.
• Metastasis delay was statistically insufficient to improve survival till PSA levels exceeded 2 ng/mL. At that level, the predefined threshold was crossed and benefit was obtained.

This is a crucial finding: delaying radiographic metastasis (detected by investigator initiated or predefined time-point bone and CT scans) is not equivalent to prolonging life.

Conclusions

POSEIDON provides the strongest evidence to date that:

• Routine addition of hormone therapy in early salvage (PSA ≤0.5 ng/mL) is not justified. Only patients with a pre-PORT PSA > 0.5 ng/mL would require hormonal therapy.
• Survival benefit of hormone therapy, especially longer duration therapy, is restricted to patients with higher PSA levels (especially >1.6–2.0 ng/mL). There was no significant interaction between the hormone therapy effect and hormone therapy duration. An exploratory analysis suggests minimal benefit to hormone therapy prolongation beyond 6 months.

In essence, POSEIDON reframes salvage therapy from a “treat-all intensification” approach to a PSA-calibrated strategy. It complements the DADSPORT meta-analysis for addition of HT to salvage therapy with similar quantum of benefits across MFS and OS

Author’s Perspective: Strengths, Limitations & Indian Applicability
Strengths (n=6057)

• Granularity and flexibility of IPD modeling of meta-analysis helped detect non-linear PSA thresholds and their interactions with outcome data.
• Robust statistical validation i.e one-stage and two-stage analyses confirmed the stability of results and the interaction of pre-PORT PSA levels and OS outcomes. It helped maintain the trial-level concurrent controls and consistency of trial-specific results.
• NNT analysis offers clinically interpretable, economically relevant and tangible data.
• Application of ICECaP surrogacy criteria Provides a mechanistic understanding of why MFS gains did not translate to OS in early salvage. Hormonal therapy is known to improve radiosensitivity in the definitive setting, but the effect on local control and survival is found lacking in a postoperative setting. (SAKK 09/10, PKUFH).

Limitations

• Missing or Competing Risk Variables, Statistical Juggling: Genomic or molecular data, comorbidities, PSA doubling time/velocity, and testosterone recovery were not captured, limiting personalization. No racial or ethnic stratification was available, limiting global generalizability.
  In the pre-specified one-stage analysis (ICECaP), the effect of any duration HT with any PSA levels (as both categorical and continuous variables) did not show statistical significance. The p-interaction for hormonal therapy only showed benefit in the secondary/two-stage analysis after adjusting for sensitivity
• Chronological Heterogeneity: The 17-year enrollment window spans major shifts in staging, Gleason score grade migration, imaging and treatment standards. The evolution of conformal radiotherapy may also have played a pivotal role in the results.
• Pre-PSMA PET Era: The absence of PSMA PET–CT is perhaps the most critical limitation. The observed survival benefit at higher PSA may reflect treatment of occult metastatic disease that was invisible on conventional imaging. However, this also contrasts with the fact that improved MFS did not lead to improved OS.
• Bicalutamide contribution: 760 patients forming the long-term HT data derived from RTOG 9601, used bicalutamide monotherapy. It would be less potent than the GnRH agonists/antagonists or second-generation AR inhibitors that are currently in use. This may have caused potential benefit results to underperform.

Indian Applicability and author perspectives

1. Early Salvage De-escalation

• In patients presenting with PSA ≤0.5 ng/mL:
  1. Avoiding HT spares financial toxicity.
  2. Avoids metabolic and cardiovascular complications in a population with rising comorbidity (DM, HTN, IHD, dyslipidemia) burden.
  3. Provides strong evidence for de-intensification.
• Late Presenters: In India, salvage RT is often delivered at higher PSA levels due to delayed detection. In that subset of patients with PSA >1 ng/mL, the survival benefit is clear. An NNT of 12–22 is clinically meaningful in a high-volume practice. POSEIDON offers evidence-based actionable thresholds.
• PSMA PET Divide: In urban centers with PSMA PET access, PSA-based thresholds may not be as relevant. However, a PSA of >0.5 ng/mL with negative PSMA PET provides challenges in interpretation and may warrant erring on the cautious side of providing HT.
  In rural settings without access to PSMA PET, POSEIDON thresholds remain highly practical.
• Bicalutamide - Friend or Foe?: While viewed as outdated in the West, bicalutamide remains affordable and accessible in India to a large subset of our population. The fact that survival benefit emerged at higher PSA using this inexpensive drug supports cost-conscious practice to patients without access to expensive GnRH analogues.
• Duration and Cost Efficiency: The lack of OS difference between short-term and long-term HT (p interaction 0.25) supports the adoption of a 6-month regimens in most salvage patients, which provides 75% cost reduction compared to 24-month therapy.
  We must also remember that patients with more adverse risk features and elevated PSA levels are generally enrolled in PORT trials which have long-term hormonal therapy.

Future Directions beyond PSA

Genomic tools such as PAM50 and Decipher may refine intensification decisions in the future, if accessible. The ongoing NRG GU006 trial (apalutamide) reflects this shift.

For India, genomic testing remains unevenly available, but the unmet need for predictive biomarkers is clear.

Final Perspective

POSEIDON is less about intensifying therapy and more about selecting patient populations where safe de-escalation of treatment may be considered after shared decision-making. It provides a gradient from biochemical triggers to biological relevance. In resource-variable environments like India, this nuanced PSA-based framework provides both scientific rigor and economic realism.

The future will integrate PSMA PET and genomic classifiers, but until then, POSEIDON offers a pragmatic, evidence-based roadmap, especially valuable in systems where overtreatment can carry both biological and financial consequences.