Prostate Cancer · Medical · Metastatic Hormone-Sensitive · ADT Strategy · 2026-09-18

SWOG 9346

In metastatic hormone-sensitive prostate cancer responding to initial ADT, can intermittent ADT preserve survival while reducing treatment burden?

Population

1,535 men with newly diagnosed metastatic hormone-sensitive prostate cancer who achieved PSA ≤4 ng/mL after 7 months of induction ADT.

Intervention

Intermittent ADT with treatment re-started according to PSA criteria.

Comparator

Continuous ADT.

Sample size / primary endpoint

1,535 randomized responders; 3,040 enrolled initially; Overall survival with a prespecified non-inferiority margin of HR 1.20

Absolute Effect

Primary efficacy result

Median survival: 5.1 vs 5.8 years; HR for death 1.10 (90% CI 0.99–1.23)

Median follow-up 9.8 years.

The confidence interval crossed the prespecified non-inferiority boundary of 1.20; non-inferiority of intermittent ADT was therefore not established.

No defensible NNT for disease control should be derived from this non-inferiority trial.

NNT / NNH

NNT: not applicable; this was a non-inferiority comparison.

NNH: not quantifiable; intermittent therapy reduced treatment exposure but could not exclude a clinically important survival disadvantage.

Treatment burden and toxicity

There was no significant difference in high-grade treatment-related adverse events. Intermittent therapy substantially reduced cumulative ADT exposure.

Quality of life: At 3 months, intermittent therapy improved erectile function and mental health, although differences diminished later.

Subgroups / clinically important modifiers

Extent of metastatic disease and PSA response after induction are important modifiers; the trial did not establish universal interchangeability.

Limitations and maturity

Only patients with a good PSA response entered randomization, and the non-inferiority criterion was not met.

Clinical interpretation

SWOG 9346 does not establish intermittent ADT as survival-equivalent to continuous ADT in metastatic disease, despite early QoL advantages.

This is an important example where a numerically modest difference cannot be relabelled as equivalence: the prespecified non-inferiority margin was crossed.

Primary / authoritative sources

SWOG 9346 / NEJM (PubMed)