For localised prostate cancer, should I have surgery or active surveillance?

Most questions in oncology do not have a randomised answer. This one does, and it has now been followed for fifteen years.

Survival did not depend on the choice#

The ProtecT trial identified localised prostate cancer in 2,664 British men who had a PSA test between 1999 and 2009, and randomly assigned 1,643 of them to active monitoring, prostatectomy or radiotherapy. Follow-up was complete for 98 percent of them at a median of 15 years.

The first row is why this trial matters: 45 men out of 1,643 died of prostate cancer in fifteen years, and the difference between the groups was not significant (P equals 0.53). The trial's own conclusion is that mortality was low regardless of the treatment assigned, so the choice is a trade between harms.

What each option does to a man#

The same trial followed patient-reported urinary, bowel and sexual function in 1,643 men, with questionnaire completion above 85 percent for most measures. This is the part of the evidence that answers the question actually being asked.

• Prostatectomy had the greatest negative effect on sexual function and urinary continence. There was some recovery, but both remained worse than the other groups throughout the trial.
• Radiotherapy's effect on sexual function was worst at six months, then recovered somewhat and was stable thereafter. It had little effect on urinary continence.
• Bowel function was worse after radiotherapy at six months and then recovered, except for bloody stools, which became more frequent over time. Bowel function was unchanged in the other two groups.
• Under active monitoring, sexual and urinary function declined gradually — the decline of ageing rather than of treatment, and not nothing.
• No significant differences were found between the groups in anxiety, depression, or general or cancer-related quality of life.

The last item is worth pausing on. The functional differences were large and the psychological measures showed nothing. A Korean target trial emulation later found the opposite on one psychological measure — men under surveillance started antidepressants more often than men who were treated — and this publication has reported both, because they disagree and the disagreement is the current state of knowledge.

In Korea, the hospital is part of the answer#

Active surveillance is a protocol, not a decision made once, and Korean institutions run different protocols. A multicentre study followed 232 Korean men who began surveillance between 2014 and 2016 at three hospitals with distinct eligibility criteria, confirmatory biopsy policies and monitoring schedules.

The difference in treatment transition was significant at P equals 0.003. Hospital B had roughly half the hazard of transition of Hospital A; Hospital C had nearly twice the hazard of proceeding to radical prostatectomy. Median surveillance duration across the cohort was 38.5 months.

The finding that gives this its weight is the one that did not differ. Among the men who did go on to surgery, final grade group and stage were the same across the three hospitals. Protocol variation changed when men were treated, not what was found when they were.

What the evidence does not settle#

1. ProtecT randomised British men detected by PSA testing between 1999 and 2009. Diagnosis has changed since — MRI before biopsy was not standard then — and a man diagnosed today has been staged more accurately than the trial's participants.
2. More than a third of ProtecT's men had intermediate or high-risk disease, so it is not purely a trial of low-risk cancer, and its results should not be applied to high-risk disease as though they were.
3. The Korean surveillance cohort is 232 men at three hospitals followed for a median 38.5 months. It is the best Korean evidence available on this question and it is small.
4. Neither the Korean cohort nor Korea's national datasets record continence or sexual function. The trade-off at the centre of this decision remains undocumented in Korean practice.
5. Active monitoring in ProtecT is not identical to modern active surveillance, which uses MRI and confirmatory biopsy. The comparison likely understates what surveillance can achieve today.

What survives all of that is the shape of the choice. Fifteen years of randomised follow-up says the survival cost of waiting is small and the progression cost is not zero, and six years of patient-reported outcomes says the functional cost of operating is real and lasting. Which of those a man would rather carry is not a question the evidence can answer for him.