by Dr Geoffrey Modest
The 19.5 year follow-up of the PIVOT study (Prostate Cancer Intervention vs Observation Trial) confirmed their prior observations at the 12-year mark, showing that for men with localized prostate cancer, surgery was not significantly associated with decreases in all-cause or prostate cancer specific mortality (see Wilt TJ. N Engl J Med 2017;377:132).
Details:
— From 1994 to 2002, 731 men with localized prostate cancer were randomized to radical prostatectomy vs observation. This study provides patient-reported outcomes data through 2014
— Mean age at the beginning of the study was 67, PSA 7.8. All patients had to have histologically-confirmed, clinically localized prostate cancer (stage TI-T2NxM0) of any grade diagnosed in the prior 12 months, and PSA needed to be <50, age <75, negative results on bone scan for metastatic disease, and life expectancy of >10 years
— Mortality outcomes were assessed locally at the site of care, but there was blinded central histopathological reclassification after the local assessment was done
— Endpoints included all-cause mortality (primary endpoint); prostate cancer mortality (which included prostate cancer treatment related mortality); local, regional, systemic, and PSA progression; additional treatments; adverse events requiring treatment; and patient reported outcomes of urinary incontinence, erectile and sexual dysfunction, worry about health, “bother” due to prostate cancer treatment, physical discomfort, satisfaction with sexual functioning, and functional limitations due to prostate cancer or its treatment
— of the 364 men assigned to radical prostatectomy, 78.8% had surgery (281 had definitive surgery, 6 had positive nodes and surgery incomplete, 25 had radiation); of the 367 assigned to observation, 36 had surgery, 1 incomplete surgery due to positive nodes, 38 radiation (and almost all completed by 4-5 years after randomization). local progression leveled off at 5 years; systemic progression leveled off after 8 years
Results:
— 281 of 364 men assigned to surgery (61.3%) and 254 of 367 assigned to observation (66.8%) died, not quite statistically significant (p=0.06). The absolute difference in risk of death was 3.1 percentage points in 8 years, increasing to 5.5 percentage points in this study. Median follow-up from randomization to death or end of follow-up was 12.7 years.
— Deaths attributed to prostate cancer or its treatment occurred in 27 men (7.4%) assigned to surgery and 42 men (11.4%) assigned to observation, almost significant, p=0.06. Deaths that were considered to be definitively due to prostate cancer treatment occurred in 4.9% assigned to surgery and 6.0% assigned observation.
— on looking at the graphs: local progression was by 3-4 years after randomization, with level curves thereafter; regional progression was mostly between 6-15 years, again mostly level curves though these were less smooth and with smaller numbers; PSA increases began at 3 years, curves splayed with surgery benefit, then paralleled around 10 years
–Among men assigned to surgery, 41% had disease progression, and 34% received treatment for disease progression; in those assigned to observation, 68% had disease progression and 60% had treatment. Local cancer progression occurred in 34% vs 62%, in surgery vs observation groups; regional in 9 vs 14% and systemic progression in 10 vs 15%
–Subgroup analyses showed:
— age at diagnosis: no difference statistically, though trend favoring surgery was more impressive in those < 65 years old
— race: no difference between white vs black
— PSA: no difference if it was < 10 vs > 10
— risk category (D’Amico risk or one based on tumor stage, histologic score, and PSA level): of the 6 categories (low, intermediate, or high risk; all as determined locally and centrally) only locally assessed intermediate risk patients had a statistically significant 20% decreased risk with
— The Forest plot did show that for each subgroup analysis, including Gleason score above or below 7, all showed a trend showing prostatectomy benefit, however none of these reached significance, either for death from any cause or death from prostate cancer (though it should be noted that the number of patients in subgroup categories were quite low).
–adverse events:
–erectile dysfunction in 15 vs 5% (for surgery vs observation), incontinence in 17 vs 4%.
— The use of absorbent pads because of urinary incontinence was greater through 10 years in men treated with surgery, with absolute differences exceeding 30 percentage points at all times measured.
— Erectile dysfunction, decreases in sexual function/activity/interest/satisfaction were much greater through 5 years in men assigned to surgery.
— urinary incontinence and erectile/sexual dysfunction, as well as limitations in activities of daily living were each greater with surgery than observation through 10 years
— Worry about health did not differ between the 2 groups, though men assigned to surgery were more likely to report” bother” due to prostate cancer or treatment, physical discomfort, and limitations of daily activity for 2 years but not at later time points.
Commentary:
–These findings are consistent with those of the Scandinavian Prostate Cancer Group Study as well as the Prostate Testing for Cancer And Treatment trials, the latter compared surgery or radiation with active monitoring and delayed radical intervention based primarily on PSA results
— as a perspective, only 9.4% of these patients died from prostate cancer, the absolute difference in all-cause mortality was 5.5 percentage points and for prostate cancer mortality was 4.0 percentage points over almost 20 years, suggesting that the long term mortality attributable to prostate cancer with the above screening criteria/methodology is quite low
— however, despite these low numbers in patients with localized prostate cancer (less than 3% for metastatic progression, less than 1% for prostate cancer specific mortality), studies suggest that observation, PSA-based monitoring, and active surveillance with delayed radical intervention are used infrequently, even among older men.
–There are several issues with this study.
–This clearly reflects radical prostatectomy as the tested intervention. Radiation therapy in its many forms was not evaluated.
–how does one explain the fact that prostate cancer mortality is quite high (3rd leading cause of death in men, 1 in 39 die from prostate cancer) yet screening does not seem to help so much?? Perhaps there really is large diversity in what we call or see under the microscope as “prostate cancer”. One patient with localized prostate cancer may go on for decades never receiving treatment, yet another with the apparent same microscopic findings and even D’Amico risk may progress quickly and perhaps no treatment would help. There was a study done many years ago which (as I remember was in JAMA and from the Prostate Cancer Prevention Trial database) found that those who had more rapid increase in their PSA (PSA velocity), even initially if the PSA was “normal”, did just as poorly if they had surgery or not. Lending some further support to this possibility was that for those who did progress in the above study, this happened early (within the first few years) with little apparent increase subsequently. Given the frequency of prostate cancer, it would be very important to have a more refined and accurate risk stratification, so that the nonprogressors do not have the aggressive treatment (with the clear and pretty immediate adverse effects), those with aggressive disease who do not benefit from surgery don’t get it, and a middle group that perhaps does benefit, does get it.
–though they did look at Gleason grade, this is a little difficult to assess here since there have been subsequent revisions in the grading system which make it hard to compare the old with the new, the newer system attributing higher Gleason scores.
–The study also did not control for potential confounders, though they do note that these would likely decrease any observed benefit from active intervention
–And, we do not have information about what happened to those on observation who had evidence of cancer spread. did they have surgery? radiation? nothing? did it matter?
–It is a little unnerving that there was such a difference in the histopathology scoring from the local hospitals to the central site (as in prior blogs, there are quite profound differences in radiologic interpretations as in mammograms as well as histopathology in several studies, and these studies are typically done by the foremost academic centers where one might expect more consistency). Leaves us in primary care in a significant quandry: when do we trust/act on/submit patients to awful further procedures or treatments if there is a real possibility that another radiologist or pathologist would have a totally different read? and how many bad things are missed because of an error in the opposite direction??. for example, see blog
so,
There are a few points here:
— the study adds impressive long-term follow-up that localized prostate cancer does not seem to increase either all-cause mortality or prostate-specific mortality significantly. And one might imagine that some of the 3% who have metastatic progression might have less so with regular PSA monitoring and biopsy vs active surveillance programs
— surgery does decrease the risk of disease progression, either by PSA or by local/regional/systemic progression. this suggests that followup is important, but that there are still a pretty large group (30+%) in those in the observation group who did fine, and the risk of erectile dysfunction (15%) and incontinence (17%) requiring treatment was pretty high. And, the long-term difference in even prostate-cancer specific mortality as well as all-cause mortality (the really important clinical outcomes) was small and not statistically significant
— The study clearly reflected surgery and not other modalities of prostate cancer therapy. However many patients still receive surgery, and it certainly seems appropriate that they be aware of the findings of the study, which seems to reflect a trend to benefit but quite real adverse effects.
— my sense from the literature is that there is more impetus to do PSA screening, after a lull of many years. I certainly do not know the right answer here. My guess is that there is a real subset of patients who have early prostate cancer who do better with surgery, and this is reflected in the fact that all of the subgroup analyses and the overall results do show a trend to better outcomes with surgery. But the fact that these surgical benefits are only a trend in this really long study strongly suggests that a large % of people getting surgery will get no benefit and the exposure to significant risks. so, it is hard for me to recommend screening without a better tool to risk-adjust patients and select those truly at increased risk for prostate cancer or all-cause mortality who might benefit.