The optimal treatment of patients with localized prostate cancer: the debate rages on
The widely anticipated results of the ProtecT study have now been published. Unfortunately, the results do little to advance our understanding as to whether surgery or radiation provides better outcomes.
In summary
The study followed oncologic and functional outcomes of 545 patients randomized to active monitoring (surveillance), 553 to radical prostatectomy, and 545 to radiotherapy. With a median follow-up of 10 years, the authors report no significant differences in prostate cancer specific (p=0.48) or overall survival (p=0.87) among the three treatment groups. They did demonstrate an increase in disease progression and metastasis among men managed with surveillance.
In an accompanying manuscript, the authors examined patient reported outcome measures out to 6 years following treatment. The authors report worse urinary continence and erectile function following surgery and worse voiding symptoms and bowel function following radiotherapy.
What do we take from this?
The investigators and participating patients should be congratulated for successfully completing this study. Numerous authors have documented their failure to adequately accrue to randomized studies of surgery versus radiotherapy in localized prostate cancer (including MRC PR06 and SPIRIT). The failure of these trials, among others, prompted Dr. Wilt to ask “Can randomized treatment trials in early stage prostate cancer be completed?” These authors have unequivocally proven that the answer is “yes”.
However, there are many caveats in applying these results to our patients:
(1) Study power
The study was clearly underpowered to evaluate the primary outcome of prostate-cancer specific mortality. Drs. Roobol and Bokhorst eloquently described important limitations of the ProtecT study. The authors designed the study assuming prostate cancer mortality of 15% at a median follow-up of 10 years. This was later adjusted downwards to 10% based on updated UK data. In the end, rates were closer to 1%. The conclusions of the primary analysis are based on a total of 17 (17!!) deaths.
(2) Study cohort – enriched with low risk disease
Among the randomized patients, the median PSA was 4.6 ng/mL, 76% had clinical stage T1c disease, and 77% had Gleason score 6 disease. These patients would almost certainly be considered most suitable for active surveillance, rather than active therapy, if seen in clinic today. Clinically meaningful decisions between surgery and radiotherapy are in the realm of treatment of intermediate and high-risk localized prostate cancer and these comprise a small group in this study. Based on this baseline distribution, it will be unlikely that any significant differences will be found in future follow-up studies.
(3) Outcomes for active surveillance
Perhaps the most notable findings of this study involve the significantly higher rates of progression, metastasis and prostate cancer specific mortality for patients treated on the surveillance protocol as compared to those treated actively, though statistical significance was not reached for PCSM. The manuscript does not provide further details regarding the pathologic characteristics of these patients. Relevantly, what was the Gleason score for these patients? This is of particularly importance as many surveillance proponents are advocating an expanding role of AS.
(4) Treatments administered
RCTs typically require significant periods of accrual, follow-up and analysis. As a result, they may be out of date prior to completion. This is certainly true of the ProtecT study. This most prominently affects patients allocated to radiotherapy. In the study protocol, patients received 3D conformal radiotherapy at 74 Gy, not the IMRT which has now become widely used. Thus, proponents of radiotherapy will likely to discount any findings which do not favour radiotherapy.
In addition, the current day relevance of the surgical treatment provided is questionable. First, the vast majority of patients in the surgical arm underwent open RP. More concerning is the quality of surgery provided: 93 patients (24%) of the cohort had positive surgical margins. In contemporary series, the average rate is under 15% with centers of excellence approaching 5%. While PSM rates clearly affect oncologic outcomes, they likely are also a surrogate of surgical quality which may affect functional outcomes.
(5) Comparison of active treatments
In the accompanying editorial, Dr. D’Amico comments on a “trend favouring radiation and ADT over surgery” and suggests that “one may consider radiation and ADT as a preferred option”. The basis for this conjecture is 5 deaths in the surgery group and 4 in the radiotherapy group, hardly a convincing sample. In contrast to these data, there was a higher number of patients with metastasis among those treated with radiotherapy (16 vs 13). These discordant results would certainly suggest that any preference for radiotherapy is premature. Indeed, with additional follow-up one would expect the patients with metastasis to die of prostate cancer, thus favouring those treated surgically.
On a methodological note, while the inclusion of active surveillance is a strength of the study, it poses analytic difficulties. The primary analysis assesses a null hypothesis assuming equality across all study interventions. Thus, as this was non-significant, pairwise testing of surgery and radiotherapy, and each with surveillance, is inappropriate and conclusions on these comparisons should not be drawn.
(6) Functional outcomes and treatment-related complications
Most clinicians are well aware that many complications other than erectile function and urinary incontinence may affect that life trajectory of patients following prostate cancer treatments. ProtecT offers the opportunity to examine the risks of secondary malignancy, repeat urologic and gastrointestinal interventions, surgeries and hospitalizations following treatment. However, these are not currently included in the published data.
Further, the PCOS studies have clearly shown that differences in patient reported urinary, sexual and bowel function change over time with convergence after long term follow-up (15 years). With ongoing maturity, it will be interesting to see if a similar pattern emerges in ProtecT.
In conclusion
The ProtecT study may raise more questions than it answers. Among a low risk group of patients, it has shown that active treatment of PSA-detected prostate cancer can reduce the progression to metastatic disease. Assessment of prostate cancer specific and overall mortality, as well as the comparative efficacy of surgery and radiotherapy, is not possible due to power limitations.
Will you be changing you patient counselling based on these results?