ProtecT at 15 Years: Active Monitoring, Surgery, and Radiotherapy Compared

The ProtecT trial found low prostate cancer mortality at a median of 15 years whether men with localized, PSA-detected disease had been assigned to active monitoring, radical prostatectomy, or radiotherapy. Death from prostate cancer occurred in 17 of 545 men assigned to monitoring, 12 of 553 assigned to surgery, and 16 of 545 assigned to radiotherapy. The differences were not statistically significant.

That finding does not make the strategies interchangeable. Metastases and clinical progression were more common after active monitoring, while surgery and radiotherapy produced different long-term urinary, sexual, and bowel effects. The trial is most useful as a map of competing outcomes, not as a declaration that one option is universally preferable.

What ProtecT randomized#

The UK trial recruited men through a PSA-testing program between 1999 and 2009. Eligible participants had clinically localized prostate cancer and were 50 to 69 years old. They were randomly assigned to one of three strategies:

• active monitoring, centered largely on serial PSA testing with clinical review;
• radical prostatectomy;
• external-beam radiotherapy with a period of androgen-deprivation therapy.

Randomization is a major strength because men choosing monitoring in routine care can differ from men choosing immediate treatment; they may be older, have lower-risk tumors, or place different value on side effects. ProtecT balanced those factors at assignment, allowing a fair comparison of strategies.

The study was analyzed by original assignment. That intention-to-treat approach preserves randomization even when participants later receive another treatment; it estimates the effect of beginning with a strategy, not the biological effect of receiving only that treatment for 15 years.

Mortality was low in all three groups#

Across 15 years, 45 participants died from prostate cancer: 3.1% in active monitoring, 2.2% after prostatectomy, and 2.9% after radiotherapy. Overall deaths from any cause were also similar across groups.

Low event counts require careful language. A non-significant difference does not prove to you that the effects are identical. Confidence intervals permit some clinically meaningful difference in either direction. What the trial shows strongly is that absolute prostate cancer mortality was low across all three assigned strategies in this particular population and follow-up window.

Long follow-up is essential because localized prostate cancer can progress slowly. Even 15 years may not capture every later difference, especially for men diagnosed in their fifties, and competing deaths also increase over time, making prostate cancer-specific and all-cause outcomes both relevant.

Progression separated the groups#

Metastatic disease developed in 51 men assigned to active monitoring, 26 assigned to surgery, and 27 assigned to radiotherapy. Those figures correspond to 9.4%, 4.7%, and 5.0%, respectively.

Clinical progression was reported in 25.9% of the monitoring group, 10.5% of the prostatectomy group, and 11.0% of the radiotherapy group. Long-term androgen-deprivation therapy was also more common after monitoring: 12.7%, compared with 7.2% after surgery and 7.7% after radiotherapy.

These endpoints matter even when mortality is unchanged. Metastatic disease can cause symptoms and require systemic treatment. Progression can bring scans, biopsies, appointments, uncertainty, and later treatment. A mortality-only summary would hide most of the difference you would actually live through.

At the same time, progression is not a synonym for death. Treating every progression event as if it had the same consequence would overstate the result. The value you place on delayed treatment, metastatic risk, treatment adverse effects, and uncertainty is not the value the next person places on them.

Active monitoring was a pathway, not permanent non-treatment#

By 15 years, 61% of men assigned to monitoring had received radical treatment. About 39% had not undergone prostatectomy or radiotherapy, and 24% were alive without radical treatment or androgen-deprivation therapy. The monitoring arm therefore tested a policy of deferring treatment with the option to intervene, not a promise to avoid treatment forever.

This crossover tends to narrow differences between randomized groups. It is not simply a protocol failure. Switching was part of the clinical strategy when PSA patterns or other findings raised concern, and an analysis based only on treatment actually received would break the protection of randomization because reasons for switching are related to prognosis.

The appropriate reading is: what happened after men began with each policy? It is not: what happens if a tumor is never reassessed or treated?

ProtecT monitoring is not identical to current active surveillance#

The trial began before multiparametric MRI, MRI-targeted biopsy, and contemporary genomic risk tools became routine. Its monitoring protocol relied heavily on PSA change, with further evaluation triggered by concern. Modern active surveillance often uses scheduled imaging, repeat biopsy, more detailed grade classification, PSA density, and other risk measures.

That difference cuts both ways. Current surveillance may identify upgrading more accurately and select low-risk patients more precisely, potentially reducing missed progression. More intensive testing can also increase procedures, false alarms, cost, and reclassification. ProtecT cannot directly quantify the outcome of every current protocol.

The terminology should remain precise. “Active monitoring” names the randomized strategy in ProtecT. “Active surveillance” often refers to newer, structured programs. “Watchful waiting” usually describes symptom-directed management when cure is not the main goal. Treating all three as synonyms obscures their intent.

Treatment harms follow different patterns#

The 15-year mortality report should be read alongside ProtecT's patient-reported outcomes. Surgery had the largest early effect on urinary continence and erectile function, and radiotherapy was associated with bowel effects and an early effect on sexual function, with urinary continence generally less affected than after surgery. Monitoring avoided immediate treatment effects, but sexual and urinary function declined over time with aging and with later treatments.

Group averages do not predict one person's experience. Baseline function, surgical technique, radiation planning, age, comorbidity, later treatment, and outcome definitions all matter. Still, the domains are stable enough to frame a decision: cancer control is not the only endpoint, and treatment harms are not captured by a generic “quality of life” score alone.

Who was represented, and who was not#

Most participants had disease that would have been considered low risk at diagnosis, although later review using contemporary methods reclassified a meaningful minority as intermediate or high risk. The trial largely involved White men in the UK, aged 50 to 69, with PSA-detected localized cancer. Results should not be stretched automatically to very high-risk disease, clinically detected tumors, severe comorbidity, much older patients, or populations underrepresented in the trial.

Treatments also evolved. Robotic surgery, radiotherapy targeting, dose schedules, imaging, pathology grading, and supportive care have changed since recruitment. A contemporary decision can use ProtecT's durable randomized comparison while still accounting for newer treatment performance and diagnostic classification.

How to discuss the result without flattening it#

Three statements can all be true:

1. Prostate cancer mortality was low and statistically similar among assigned groups at 15 years.
2. Active monitoring produced more metastasis and progression.
3. Immediate surgery and radiotherapy produced treatment-specific functional harms.

The decision is therefore preference-sensitive when clinical risk allows more than one reasonable strategy. Relevant questions include how a person values avoiding or delaying treatment, how they view a higher chance of progression, baseline urinary and sexual function, life expectancy, tumor grade and extent, other illnesses, and willingness to continue structured follow-up.

A trial-reading checklist#

• Was the outcome death, metastasis, progression, treatment use, or function?
• Are absolute risks reported alongside relative comparisons?
• Is the analysis based on randomized assignment?
• How much crossover occurred, and was it part of the assigned strategy?
• Does the monitoring protocol resemble the current program under discussion?
• Does the patient's risk group resemble the enrolled population?
• Are patient-reported harms considered alongside cancer endpoints?
• Is uncertainty acknowledged when event counts are low?