Fact check: Can a patient have a zero PSA level after prostatectomy and still experience cancer recurrence?

1. A straightforward claim: Can “zero PSA” equal zero risk?

Clinical analyses consistently report that an undetectable PSA reduces but does not eliminate the risk of later biochemical recurrence. A 2025 cohort study that tracked men with at least 10 years of undetectable PSA after radical prostatectomy found that 5.2% subsequently developed biochemical recurrence, demonstrating that long-term undetectable PSA is strongly protective but not an absolute safeguard ^[1]. Earlier work using ultrasensitive assays reached similar conclusions: very low or undetectable postoperative nadirs predict relapse-free survival, yet recurrence remains possible, especially when microscopic residual disease or later metastasis occurs ^[4].

2. What the salvage-radiotherapy literature adds to the picture

Studies of salvage radiotherapy after prostatectomy show undetectable PSA post-salvage is associated with better biochemical progression-free and metastasis-free survival, and these results have prompted discussions about de-escalating additional systemic therapy for selected patients who achieve undetectable PSA after salvage treatment ^[2]. Those findings emphasize that treatment response status—undetectable PSA after salvage therapy—carries prognostic weight and can guide management decisions, though the data do not imply that such patients face zero future risk of recurrence ^[2].

3. How long-term natural history studies shape expectations

Longitudinal studies tracking men with initially undetectable PSA who later developed detectable PSA show a varied natural history: many patients experience slowly rising PSA with excellent long-term outcomes even without immediate salvage therapy, while others progress to clinically significant disease. This heterogeneity illustrates that PSA kinetics and clinical context matter—a single undetectable PSA value cannot substitute for serial monitoring and risk stratification based on pathological features and time since surgery ^{[5] [4]}.

4. The role of prognostic models and machine learning

Recent diagnostic systematic reviews report that machine-learning models achieve robust discrimination for predicting biochemical recurrence (pooled AUC ~0.82), particularly over short-term horizons, and integrating multimodal data improves performance ^[3]. These models underscore that combining PSA trajectories with clinical, pathological, and imaging data refines recurrence risk estimates beyond a single undetectable PSA value; however, models are probabilistic and cannot certify individual outcomes, leaving a residual risk even when predicted risk is low ^[3].

5. Reconciling apparent contradictions across studies

Apparent contradictions—undetectable PSA predicting favorable outcome versus documented late recurrences after long undetectable intervals—are reconciled by recognizing differences in study design, follow-up duration, assay sensitivity, and patient selection. Cohorts with extended follow-up reveal small late-recurrence rates that shorter studies or salvage-treatment-focused cohorts may not capture, while studies emphasizing early post-treatment undetectability show strong short-term prognostic value. Consequently, both messages are true: undetectable PSA is highly prognostic for favorable outcomes but not an absolute guarantee against later recurrence ^{[1] [2] [5]}.

6. Practical implications for patients and clinicians

For clinicians and patients, the evidence supports continued serial PSA monitoring and individualized risk assessment even after prolonged undetectable readings. Undetectable PSA can inform decisions about adjuvant therapy or surveillance intensity, and achieving undetectability after salvage radiotherapy may justify withholding androgen-deprivation therapy in selected patients, given the prognostic association with better outcomes; nonetheless, a small absolute risk of later biochemical recurrence persists and should be part of shared decision-making ^{[2] [1]}.

7. Gaps, uncertainties, and what to watch next

Key uncertainties include the drivers of late biochemical recurrence after long undetectable intervals, the impact of ultrasensitive assays on long-term risk estimates, and how evolving predictive algorithms will alter follow-up strategies. Ongoing work combining longer follow-up cohorts, standardized ultrasensitive PSA use, and validated machine-learning tools will clarify how to balance reassurance from undetectable PSA against the documented small residual risk of recurrence; until then, practice should reflect both the prognostic strength and the limitations of an undetectable PSA ^{[4] [3] [1]}.