Radical Surgery versus Radiotherapy for Adenocarcinoma of the Prostate
Reviewed by: Drew Moghanaki MD, MPH
Source: The Journal of Urology 1982. Sep;128(3):502-4.
This report by the Uro-Oncology Research Group is the first and only purported prospective randomized study comparing prostatectomy to external beam radiotherapy in the management of localized prostate cancer. Unfortunately, this paper has been strongly criticized since its publication, even by one of its own authors.1
It is unclear whether prostatectomy or external beam radiation is superior in the treatment of localized prostate cancer. This study represents an attempt to compare the two treatments in a prospective, randomized fashion.
Between 1974 and 1978, patients with newly-diagnosed biopsy-proven prostate adenocarcinoma were enrolled in a prospective multi-institutional study, and randomized to either prostatectomy or external beam radiotherapy. All patients had locally confined tumor (T1-T2 or A2-B), normal pretreatment bone scans, normal serum prostatic acid phosphatase (PAP) levels, and pathologically negative pelvic nodes. One hundred-six patients were enrolled and assigned to treatment at each institution via a scheme of 4 envelopes containing assignment to either surgery or radiation (2 envelopes to each modality per institution). Surgical patients underwent suprapubic or perineal prostatectomy, while radiotherapy patients received 6,500-7,000 cGy to the prostate. The total dose consisted of 4,500-5,000 cGy to the whole pelvis followed by a 2,000 cGy boost to the prostate.
Initially, 59 and 47 patients were randomized to radiation and prostatectomy, respectively. However, there were patients in both groups who failed to receive their assigned treatments, and cross-over to the opposite treatment arm was allowed.
Analysis was performed by treatment given, and not by treatment intention.
The primary endpoint of failure was recorded as any 2 consecutive rises in PAP levels, or abnormal chest x-ray or bone scan.
Follow-up consisted of periodic physical examination, PAP levels, CXR, and bone scans. The trial was closed early due to inadequate funding, and after study closure, follow-up appointments were only encouraged, but not required.
Although 66% and 38% of prostatectomy and radiotherapy patients, respectively, were lost to follow-up, the authors reported a statistically significant (p=0.037) increased rate of failure in the radiotherapy arm via Kaplan-Meier analysis.
Prostatectomy provides a statistically significant increase in controlling organ-confined prostate cancer as compared to external beam radiotherapy.
There are multiple flaws in this paper that are raised in subsequent editorials.1,2 Perhaps the single largest problem is highlighted in an editorial by the collective group of radiation oncologists participating in this study. They list criticisms that were raised during manuscript preparation, but never incorporated into the final version. The majority of criticism is directed toward the manner of statistical analysis that was utilized. Interestingly, the analyses were not conducted by the UORG as originally planned, but at Duke University, without any oversight by the radiation oncologists involved.
Other problems include the 4-envelope randomization scheme, which is fraught with potential selection bias, as it allows the opportunity to know the randomization in up to 50% of patients. In addition, not all institutions completed the 4-patient randomization, raising concern for physician selection bias. The analysis was not intent-to-treat and permitted cross-over, neither of which is generally accepted practice in randomized studies. Pathologic stage C patients were excluded from the surgical arm but allowed in the radiation arm, and the rate of failure in the radiotherapy arm resembles a pattern more consistent with stage C patients. No median follow-up time is given, and no overall survival data are reported.
The authors do attempt to describe the radiotherapy technique; however, the information provided suggests a non-standard fractionation scheme during the whole-pelvis component of treatment. Additionally, although unpublished, it has been suggested that the vast majority of patients received only 6,500 cGy, which by today's standards is known to be sub-therapeutic.
There was no tissue validation reported for relapse, and no quality assurance was maintained for abnormal bone scans. Multiple reports have demonstrated a >30% rate of false positive bone scans in patients treated for prostate cancer. Additionally, it has become clear over time that elevated PAP levels can result from radiotherapy to bone, confounding this variable as a marker of failure in the radiotherapy group.
Regarding the current utility of this study in the treatment decision-making process for patients with locally confined prostate adenocarcinoma, it should be emphasized that the surgical procedure used, prostatectomy, has since been replaced with radical prostatectomy as the standard uro-oncologic approach for prostate cancer. Additionally, a 40% rate of incontinence among patients undergoing surgery is considered unacceptably high by today's standards.
It is unfortunate that the methodologic flaws present within the study preclude any meaningful conclusions regarding prostatectomy versus radiotherapy in organ-confined prostate cancer. Modern retrospective analyses have repeatedly suggested equivalency of surgery versus radiation in this patient population.3 Ultimately, however, it would be useful if an appropriately designed prospective randomized trial addressing this question could be conducted.
1 Re: Radical surgery versus radiotherapy for adenocarcinoma of the prostate. Byhardt RW, Greenlaw RH, Jensen R, Nag S, Roswit B, Stephani S, Woodward K. Journal of Urology 1983 Dec;130(6):1205-6.
2 More on the Uro-Oncology Research Group report of radical surgery vs. radiotherapy for adenocarcinoma of the prostate. Int J Radiat Oncol Biol Phys. 1988 May;14(5):1053-4.
3 Pretreatment PSA velocity and risk of death from prostate cancer following external beam radiation therapy. JAMA. 2005 Jul 27;294(4):440-7.