July 19, 2012 — Most men with localized prostate cancer, especially those with either low-risk disease or prostate-specific antigen (PSA) levels lower than 10 ng/mL, should be monitored initially rather than treated with surgery, according to the authors of a major randomized trial.
Overall, radical prostatectomy did not significantly reduce either all-cause or prostate-specific cancer mortality when compared with observation among men with localized disease, report the investigators from the Prostate Cancer Intervention Versus Observation Trial (PIVOT).
These results, from a median follow-up of 10 years, are published in the July 19 issue of the New England Journal of Medicine.
The absolute differences in the 2 mortality measures between the observation and surgery groups were less than 3 percentage points and were not statistically significant, report lead author Timothy Wilt, MD, MPH, from the Minneapolis Veterans Affairs Health Care System and the University of Minnesota, and colleagues.
Specifically, 47.0% of the surgery group died from all causes vs 49.9% of the observation group (hazard ratio, 0.88; 95% confidence interval [CI], 0.71 - 1.08; P = .22). In addition, 5.8% of the surgery group died from prostate cancer or treatment vs 8.4% of the observation group (hazard ratio, 0.63; 95% CI, 0.36 - 1.09; P = .09).
"Our study showed that men with early-stage prostate cancer treated with observation had similar length of life and deaths from prostate cancer," Dr. Wilt told Medscape Medical News. "Physicians can now confidently recommend observation as the preferred treatment approach for most of their patients diagnosed with [early] prostate cancer."
Notably, a PIVOT subset analysis also indicated that surgery provided no mortality benefit at all in men with early-stage (localized) disease that was low risk. Instead, observation appeared superior in low-risk men; it was associated with a nonsignificant reduction in both mortality measures.
Subset analyses also indicated that surgery did not provide a significant mortality (all-cause and disease-specific) benefit for men with PSA levels lower than 10 ng/mL.
However, the authority of these findings is undermined by the fact that the PIVOT study was insufficiently powered, suggest 2 experts in an editorial that accompanies the study.
The original design called for 2000 men to be randomly assigned to either surgery or observation, point out editorialists Ian Thompson, MD, from the University of Texas Health Science Center in San Antonio, and Catherine Tangen, DPH, from the Fred Hutchison Cancer Research Center in Seattle, Washington.
Unable to recruit enough patients, the investigators modified the design, which subsequently called for 740 patients, but only enrolled and randomized 731.
"[T]he study was thus underpowered to detect this relatively large clinical effect," the editorialists write, referring to the stated goal of detecting a 25% relative reduction in all-cause mortality, the primary endpoint.
Despite this objection, Dr. Thompson and Dr. Tangen agree with the PIVOT investigators on an all-important point: Men with low-risk, localized disease should be monitored first, not treated.
The editorialists say that subset data from PIVOT and other data from a Canadian trial "strongly support this approach [active surveillance], which should be offered to men with low-risk cancer" ( J Clin Oncol. 2010;28:126-131).
Observation is a "wise, healthy choice for the large majority of men diagnosed with prostate cancer in the United States," said Dr. Wilt, explaining that the harms related to treatment can be avoided, and that in PIVOT, the risk for death from prostate cancer in the observation group was "low."
Which Men Should Be Treated Initially?
Surgery may be an appropriate first option for men with high initial PSA values who have a risk profile that is not low, say the authors.
Radical prostatectomy was associated with reduced all-cause mortality among men with a PSA value higher than 10 ng/mL (P = .04 for interaction), and "possibly" among those with intermediate-risk or high-risk tumors (P = .07 for interaction), they report.
The editorialists were encouraged that, despite the limitations of the trial, there was a trend toward reduction in mortality among men with high-risk cancers.
The ultimate challenge moving forward in the management of prostate cancer is in accurately differentiating risk so that treatment can focus on men with "lethal cancer," say the editorialists. "Prostate cancer is not a monolithic cancer but a spectrum of disease," Dr. Thompson and Dr. Tangen write. Future clinical trials must focus on "cancers that matter," they add.
This comment could be interpreted as a criticism of PIVOT, which randomly assigned all of the men in the trial in the same manner, despite the different levels of risk.
In the trial, risk was assessed with the D'Amico scale; the tool combines the clinical TNM (tumor, node, metastasis) stage, biopsy Gleason score, and preoperative PSA level. About 40% of the participants were low risk, 34% were intermediate risk, and 21% high risk (for 5%, risk data were missing).
Enrolling these men in the trial was a tremendous task. The investigators screened 13,022 men with prostate cancer at 44 Veterans Affairs sites and 8 National Cancer Institute sites. From these patients, 5023 men met eligibility criteria and 731, with a mean age of 67 years, were enrolled.
Nearly one third of participants were black, and 85% of the patients reported they were fully physically active. Median PSA was 7.8 ng/mL. In three fourths of men, the primary reason for biopsy leading to a diagnosis of prostate cancer was a PSA elevation or rise.
Adherence: A Study Limitation?
Observation, and not active surveillance, was used as the comparator in the trial. The concept of active surveillance had not yet evolved when PIVOT began in 1994, which was, in the words of the authors, the "early era of PSA testing."
There are differences between observation and active surveillance, the study authors say. With active surveillance, if disease progression is suspected through PSA testing or biopsies, there is treatment with "curative intent." In PIVOT, patients in the observation were provided palliative treatments "if and when patients developed signs or symptoms of disease progression," but about 1 in 5 were either referred to treatment by their physician or chose it themselves, said Dr. Wilt.
According to the editorialists, 1 of the weaknesses of the PIVOT trial is that 20% of the men in the observation group ultimately received definitive treatment. In addition, 21% of the men in the surgery group never had a prostatectomy and were essentially observed during the study. The combination of these factors "further reduces the ability to discern a treatment effect," write the editorialists.
However, Dr. Wilt said that "adherence to the protocol...was considered very good and comparable to other surgical trials, including another large study of surgery vs observation in men with early-stage prostate cancer." He was referring to the Scandinavian Prostate Cancer Group 4 (SPCG-4) trial.
Results from SPCG-4, which were called a "very, very important" by an American expert, were published last year and indicated a survival benefit from surgery that was age dependent. Dr. Wilt and coauthors point out that SPCG-4 was conducted without the benefit of PSA testing and relied on enrolling patients with clinical symptoms, which created a very different population from the PIVOT trial participants, who had a more favorable prognosis overall.
The PIVOT authors champion active surveillance/observation in the discussion section of their new article, but acknowledge that at present this approach is not used much in the United States.
"Up to two thirds of men who have received a diagnosis of prostate cancer have a low PSA value or low-risk disease, but nearly 90% receive early intervention — typically surgery or radiotherapy," they write.