Surgery, Radiation, or Active Surveillance? Findings From the ProtecT Trial for Prostate Cancer

Surgery, Radiation, or Active Surveillance? Findings From the ProtecT Trial for Prostate Cancer

Dr. Neha Vapiwala, credit David Debalko

By Neha Vapiwala, MD

Article Highlights

  • The ProtecT trial provided the first level 1 evidence comparing the relative long-term impacts of surgery, radiation, and active surveillance on outcomes and quality of life (QOL) of patients with localized prostate cancer.
  • There was no difference in metastasis-free or overall survival between the radical prostatectomy and external-beam radiotherapy groups, and a “winner” cannot be declared. However, the nature, duration, and QOL impact of the side effects were different between the arms and warrant discussion.
  • Although the results are not practice-changing, these data assist more informed decision-making by all stakeholders. The results also reinforce the now-established role of active surveillance as the third form of treatment for localized prostate cancer. 

To say that publication of the ProtecT trial results1,2 was highly anticipated is an understatement. Various stakeholders in the prostate cancer community have patiently awaited answers to the perennial question of whether surgery or radiation is “better”—and the more topical question of whether either of these is better than active surveillance—in patients with clinically localized prostate cancer.

As the first, and likely only, level 1 evidence comparing the relative long-term impact of surgery, radiation, and active surveillance on outcomes and quality of life (QOL) of patients with prostate cancer, the successful conduct of this trial is a feat in and of itself. More than 1,600 men with localized prostate cancer, derived from a larger group of more than 2,600 men diagnosed with localized prostate cancer, were randomly assigned to active surveillance (545 patients), radical prostatectomy (553 patients), or external-beam radiotherapy (RT; 545 patients). The primary outcome was prostate cancer–specific mortality; secondary outcomes included the rates of disease progression, metastases, and all-cause deaths.

With a median follow-up of 10 years, there was no statistically significant difference in either prostate cancer–specific mortality or overall mortality among the three groups. Notably, the rates of death from prostate cancer were low in all three groups1—which I will discuss in greater detail later. In terms of the patient-reported outcomes, the authors reported on multiple domains, including urinary and sexual function and general health- and cancer-related QOL. Sexual function declined in all three groups, but there was less of a deterioration in the active surveillance group. Radical prostatectomy was associated with the greatest and most persistently negative impact on sexual and urinary function, whereas patients in the RT arm reported worse bowel function and urinary voiding at 6 months, with some or near-complete recovery, respectively. Interestingly, levels of anxiety, depression, and QOL were not significantly different among the groups.2

Praise for this study as a Herculean undertaking with meticulous long-term follow-up is well deserved, but we should consider some important caveats and criticisms. The original study design was based on an assumed 10-year prostate cancer–specific mortality of 15%, and later 10%. But with only 17 deaths from prostate cancer, the study was arguably underpowered to evaluate the primary outcome.3 

Active Surveillance Considerations

The data above essentially corroborate the current practice of offering active surveillance to patients with low-risk prostate cancer, which represented the majority of patients in this population. These results support long-term safety and tolerability of active surveillance in appropriately counseled patients with localized prostate cancer when cause-specific death is the primary endpoint of interest, rather than biochemical or disease progression. But, as with any high-profile study that generates attention in both the trade and lay press, providers and patients must consider the nuances when interpreting and applying the results to formulate optimal personalized disease-management decisions.

For example, some news headlines have warned that patients randomly assigned to the active surveillance group were “more than twice as likely” to develop cancer progression and spread compared with the patients assigned to undergo immediate treatment.4 Although disease progression and metastasis development were statistically significantly worse (p < 0.001 and p = 0.004, respectively) in the active surveillance arm compared with either treatment arm, they remarkably did not result in greater cause-specific mortality. Another 5 years of follow-up may reveal this as a downside or, alternatively, demonstrate that such differences are ultimately negated by survival-enhancing agents and competing causes of death.

For now, it is certainly important to inform the appropriately selected active surveillance candidate about this relative risk compared with their counterpart who elects to have immediate prostate cancer treatment. Having said this, it is the absolute risk of a given negative outcome that fundamentally provides necessary and meaningful context to understand relative risks.

It is challenging but critical to provide patients with a clinical perspective on the statistics so that they can determine their personal priorities and exercise their prerogative accordingly. When the absolute risk of prostate cancer disease progression is exquisitely low with treatment—as in this case at about 3%—a doubling to 6% without treatment may be quite acceptable to some patients for deferring the treatment-related morbidity described above, recognizing that neither radical prostatectomy nor RT are guaranteed cures, even in early-stage, organ-confined disease.

Conversely, for other patients, any intervention to minimize the risk may be preferable to no intervention; 15- and 20-year data may be more relevant to properly weigh the ramifications of an active surveillance decision made at age 50 on disease-free survival. The logistical and financial costs should also be factored, along with psychosocial stressors on loved ones—none of which are captured here, but they are factors that will matter to some individuals. Unfortunately, the published reports do not provide details regarding the Gleason score or other pathologic characteristics of the patients whose disease progressed, which would be useful when counseling individual patients.

Two other aspects must be considered when adapting the ProtecT trial’s active surveillance data to today’s clinic. First is the very real-world crossover rate: approximately one-third of patients in the active surveillance arm elected to undergo treatment secondary to anxiety and suspicion, and an additional one-fifth pursued treatment for disease progression, leaving 44% of the original group successfully on active surveillance for 10 years. Second, modern-day active surveillance protocols in the United States are more stringent in terms of prostate-specific antigen and biopsy requirements than the protocols used in the ProtecT trial, and emerging molecular and imaging advances were, of course, not available.5 As with any clinical trial, the benefits of long-term follow-up are inherently tempered by the inability of the original protocol to account for every advance during the course of its conduct.    

Immediate Treatment Considerations

The statement above also understandably applies to the two treatment arms, primarily with respect to the relatively high rates of patient-reported toxicities. Taking the positive resection margin rate as a surrogate for quality, modern-day series report an average of 15% (markedly lower in high-volume, high-quality settings) and raise some concern regarding the 24% reported in the surgical arm of the study. Urologic surgeons also nowadays commonly perform robotic-assisted radical prostatectomy, whereas nearly all patients in this study underwent the more traditional open retropubic nerve-sparing radical prostatectomy. Similarly, some radiation oncologists would argue that 3D conformal RT and the use of 3 months to 6 months of androgen deprivation therapy (not medically indicated in this group6) are contributors to greater toxicity that would not be used in the United States for these patients today. In the case of radiation technique, the use of intensity-modulated radiation has been associated with both reduced recurrence rates and toxicity, albeit not with overall survival differences.7  

As mentioned earlier, the ProtecT study population is enriched with patients who would be prime active surveillance candidates by today’s standards (median prostate-specific antigen 4.6 ng/mL and more than three-quarters with clinical stage T1c and Gleason score 6 disease). To determine clinically meaningful differences between surgery and RT, one ideally needs a population prone to more disease-related events (i.e., intermediate- and high-risk localized prostate cancer), which, unfortunately, comprises a relatively small proportion of the trial. As far as what we can glean from this dataset, there is no difference in terms of metastasis-free survival or overall survival between the two treatment groups, and a “winner” cannot be declared.

To many radiation oncologists accustomed to hearing that surgery is the gold standard for prostate cancer treatment, the equal footing—as of now, anyway—is a small victory in itself. The nature, duration, and QOL impact of the side effects, however, are quite different between radical prostatectomy and RT, and they warrant healthy and honest discussion. Additional follow-up will reveal whether everything equals out in the end, as it did in the Prostate Cancer Outcomes Study,8 and if differences in provider-reported grade 3 to 4 late toxicities and secondary malignancy rates emerge. Ultimately, both reports from the ProtecT trial humbly remind us of the therapeutic limitations of immediate treatment and of the potential for treatment to do harm in a population that is unlikely to die of their disease.


The ProtecT investigators should be commended for providing us with what no one else could or likely ever will. Although the results are not practice-changing and the roles of radical prostatectomy and RT remain vital in the overall management of prostate cancer, including for some men with early-stage disease, these data assist more informed decision-making by all stakeholders. The results also reinforce the now-established role of active surveillance as the third—and for many, perhaps, the preferred—form of treatment for men with localized prostate cancer. 

About the Author: Dr. Vapiwala is an associate professor of radiation oncology and vice chair of education in the Department of Radiation Oncology at the Hospital of the University of Pennsylvania.