Upcoming event

Ten-year mortality, disease progression, and treatment-related side effects in men with localised prostate cancer from the ProtecT randomised controlled trial according to treatment received

  • David E. Neal,
  • Chris Metcalfe,
  • Jenny L. Donovan,
  • J. Athene Lane,
  • Michael Davis,
  • Grace J. Young,
  • Susan J. Dutton,
  • Eleanor I. Walsh,
  • Richard M. Martin,
  • Tim. J. Peters,
  • Emma L. Turner,
  • Malcolm Mason,
  • Richard Bryant,
  • Prasad Bollina,
  • James Catto,
  • Alan Doherty,
  • David Gillatt,
  • Vincent Gnanapragasam,
  • Peter Holding,
  • Owen Hughes,
  • Roger Kockelbergh,
  • Howard Kynaston,
  • Jon Oxley,
  • Alan Paul,
  • Edgar Paez,
  • Derek J. Rosario,
  • Edward Rowe,
  • John Staffurth,
  • Doug G. Altman,
  • Freddie C. Hamdy

Publication: European Urology, November 2019

Background

The ProtecT trial reported intention-to-treat analysis of men with localised prostate cancer randomly allocated to active monitoring (AM), radical prostatectomy, and external beam radiotherapy.

Objective

To report outcomes according to treatment received in men in randomised and treatment choice cohorts.

Design, setting, and participants

This study focuses on secondary care. Men with clinically localised prostate cancer at one of nine UK centres were invited to participate in the treatment trial comparing AM, radical prostatectomy, and radiotherapy.

Intervention

Two cohorts included 1643 men who agreed to be randomised and 997 who declined randomisation and chose treatment.

Outcome measurements and statistical analysis

Analysis was carried out to assess mortality, metastasis and progression and health-related quality of life impacts on urinary, bowel, and sexual function using patient-reported outcome measures. Analysis was based on comparisons between groups defined by treatment received for both randomised and treatment choice cohorts in turn, with pooled estimates of intervention effect obtained using meta-analysis. Differences were estimated with adjustment for known prognostic factors using propensity scores.

Results and limitations

According to treatment received, more men receiving AM died of PCa (AM 1.85%, surgery 0.67%, radiotherapy 0.73%), whilst this difference remained consistent with chance in the randomised cohort (p = 0.08); stronger evidence was found in the exploratory analyses (randomised plus choice cohort) when AM was compared with the combined radical treatment group (p = 0.003). There was also strong evidence that metastasis (AM 5.6%, surgery 2.4%, radiotherapy 2.7%) and disease progression (AM 20.35%, surgery 5.87%, radiotherapy 6.62%) were more common in the AM group. Compared with AM, there were higher risks of sexual dysfunction (95% at 6 mo) and urinary incontinence (55% at 6 mo) after surgery, and of sexual dysfunction (88% at 6 mo) and bowel dysfunction (5% at 6 mo) after radiotherapy. The key limitations are the potential for bias when comparing groups defined by treatment received and changes in the protocol for AM during the lengthy follow-up required in trials of screen-detected PCa.

Conclusions

Analyses according to treatment received showed increased rates of disease-related events and lower rates of patient-reported harms in men managed by AM compared with men managed by radical treatment, and stronger evidence of greater PCa mortality in the AM group.

Expert's summary

By Dr. Ploussard

The ProtecT trial has randomized men aged 50-69 years with localized prostate cancer to be treated by active monitoring, radiotherapy or surgery. In the present series, the authors report outcomes according to treatment received in both randomized and treatment choice cohorts. The aim of the analysis by cohorts was to limit potential underestimation of group differences that could be caused by intention-to-treat analysis given the high proportion of patients (38%) who chose treatment and declined randomisation. Overall, it included 1643 men (62%) who agreed to be randomized and 997 men who declined randomisation and chose treatment. Differences between cohorts were estimated with adjustment for known prognostic factors using propensity scores.

In the cohort of men who declined randomisation, 51% of men chose active monitoring, 26% surgery, 13% radiotherapy and 6% brachytherapy.  Men with higher-risk grade and stage disease were more likely to undergo radical treatments than active monitoring.

Although the risks of disease-specific and all-cause death remained low, there was evidence of increased metastatic disease, disease progression, and initiation of androgen deprivation therapy in the active monitoring group compared with surgery and radiotherapy groups. The rate of prostate cancer death per 1000 person years was 1.5 in the randomised active monitoring group versus 0.75 and 0.92 in the randomised radiotherapy and surgery groups, respectively. Disease progression was also more common in the active monitoring group: 20.3% versus 5.9% in the surgery group, and 6.6% in the radiotherapy group. These findings were consistent in both randomised and treatment choice cohorts. Whereas the difference regarding death from prostate cancer remained consistent with chance in the randomised cohort (p=0.08), statistical significance was reached in the exploratory analyses performed in the randomised plus treatment choice cohort. The risk of dying from prostate cancer was reduced by 68% (p=0.026) and 66% (p=0.034) in men undergoing surgery and radiotherapy, respectively, compared with active monitoring. The combined risk of developing metastasis or dying from prostate cancer similarly decreased by 60% to 67% with radical treatments. No difference was observed for all-cause mortality according to the treatment received. With regards to the treatment-related side effects, there were higher risks of sexual dysfunction and urinary incontinence after surgery, and of sexual and bowel dysfunction after radiotherapy compared with active monitoring. Only 15-30% of men reported being potent in the radiotherapy and surgery groups at 6 years. This proportion was significantly higher in the active monitoring group but it reduced gradually over time, reflecting ageing and differed radical treatment received during follow-up.

Globally, the exploratory analyses comparing outcomes between groups defined by treatment received confirmed the initial findings reported with the intention-to-treat analysis. However, there was a stronger evidence supporting a lower risk of prostate cancer related deaths and metastasis after surgery and radiotherapy compared with active monitoring, when results of the randomisation and treatment choice cohorts were pooled. The absolute reduction of specific deaths remained low. Nevertheless, the development of metastasis (which were significantly reduced with radical treatments) may be considered as a good surrogate for specific death. Moreover, follow-up is continuing and subsequent analyses may strengthen the trends reported here regarding cancer specific survival benefits with radical treatments.