Comparative 5-year outcomes of brachytherapy and surgery for prostate cancer
Article Outline
Abstract
Purpose
To compare the predicted outcome by radical retropubic prostatectomy (RRP) of patients actually treated with prostate brachytherapy (PB) for low- and intermediate-risk prostate cancer, using a surgical nomogram.
Methods and Materials
Data of consecutive patients treated with PB from the start of our program in 1998 to February 1, 2005 were extracted from a prospectively maintained database that records patient, tumor, and outcomes information. Prognostic features of each patient and brachytherapist experience were entered into a nomogram of surgical outcomes that incorporates surgeon experience. The definition of failure was a rising prostate specific antigen (PSA) >0.4
ng/mL or secondary intervention. Comparisons of actual PB outcome vs. predicted surgical outcome were compared using the log-rank test.
Results
One thousand two hundred fifty-four patients were treated in this era, and all are included for analysis. The median followup is 56 months. Forty-six percent have stage T1, and 54% have stage T2 cancer. Twenty-five percent have Gleason 7 disease, and the median PSA is 6.3ng/mL (range, 0.3–19.6ng/mL). Fifty-eight percent have low-risk disease, and 41% have intermediate-risk disease. Androgen deprivation therapy was used for 6 months in 92% of intermediate-risk and 46% of low-risk cases. The median PSA at last followup is 0.06ng/mL. Five-year biochemical no evidence of disease rate with PB is 90.6%, and that predicted with RRP is 86.8%, p=0.003.
Conclusions
Excellent PB outcomes are observed that exceed that predicted from the use of RRP.
Keywords: Prostate brachytherapy, Nomogram, Prostatectomy, Outcomes, Low dose rate
Introduction
Low-dose-rate prostate brachytherapy (PB) is an excellent form of treatment for the management of low- and selected intermediate-risk prostate cancer patients [1], [2], and competes with radical prostatectomy as men decide which treatment is best for them. One of the key pieces of information that 83% of men regard as essential knowledge is “When and how will you know I have been cured” (3). For almost all men with prostate cancer, the earliest indication of treatment failure is the development of a rising prostate specific antigen (PSA), although it may not be associated with any physical symptoms, nor require intervention (4).
Nomograms have been developed for retropubic prostatectomy (RRP), PB, and external beam radiation therapy (EBRT), and their utility has been reviewed (5). They are the best way of estimating the individual risk of recurrence compared with other available tools, such as risk-grouping models (6). The nomogram that is used in this study incorporates surgical experience as an independent factor (7). Surgical caseload has been described as an independent factor of outcome for radical prostatectomy, and is the most important factor predicting recurrence in those with low-risk disease (8). Similar findings have been demonstrated with PB, where the hazard of relapse is reduced by 0.89 per 100 cases (9).
This study was undertaken to compare the actual PSA control outcomes of men who were treated with PB, with the outcome predicted from the nomogram who had RRP (7). The nomogram is reproduced in Fig. 1.
Fig. 1
Nomogram used for the calculation of biochemical no evidence of disease rates with brachytherapy and prostatectomy patients (7).
Reproduced with permission from Wiley Interscience (www.interscience.wiley.com).
Methods and materials
Consecutive patients treated with PB at the British Columbia Cancer Agency from the first implant on July 20, 1998 to February 1, 2005 were eligible, so as to give a potential minimum followup of 4 years at data close-out in February 2009. Patients treated with surgery, who comprise the nomogram data were treated between 1987 and 2003. Institutional research ethics board approval was obtained for this study.
Eligibility for PB was defined as men with low-risk disease (defined as pretreatment PSA ≤10ng/mL, Gleason score [GS] ≤6, and clinical stage ≤T2b) and a “low-tier” group of intermediate-risk disease (defined as clinical stage ≤T2c and either PSA of 10–15ng/mL with GS ≤6 or GS of 7 with PSA ≤10ng/mL). Rarely (<3%), patients outside these guidelines were also treated. The 1992 tumor, node, metastasis (TNM) system was used until January 2000 (10), the 1997 TNM was used until January 2003, and the 2002 system thereafter. The nomogram inputs 2002 TNM for all the patients. As it is not always possible to convert T2 tumors between versions, a minority of patients (estimated 14%) were thus likely overstaged according to the 2002 system. For clarity, only 1997 TNM stages are quoted in this report. Patients with intermediate-risk disease and those with low-risk cancer with a prostate volume >40cm3 in the first year of the program and 50cm3 thereafter were treated with 3 months of neoadjuvant and 3 months of concurrent androgen deprivation therapy (ADT), comprising a luteinizing hormone releasing hormone agonist with 1 month of antiandrogen. Our brachytherapy technique has been previously described (11). 125I sources were implanted according to a preplan in a modified peripheral distribution to deliver a minimum peripheral dose of 144Gy. Postoperative dosimetric computerized tomography scans were routinely performed in all the patients.
For each patient, an individual radiation oncologist was identified as being the primary treating radiation oncologist. Our program was designed on an apprenticeship basis, and until an Oncologist had achieved a minimum caseload, the program was mentored by an experienced brachytherapist. Thus, early in our program each implant was performed by two oncologists, until sufficient experience was gained. For the first 2 years of our PB program, only four radiation oncologists performed PB. For the purposes of coding experience for this study, the primary radiation oncologist was defined as the oncologist named on the operative report as the surgeon. Urologists do not participate in the actual PB procedure in British Columbia. Likewise, in the surgical nomogram, surgeon experience was coded as the number of radical prostatectomies performed by the surgeon before the patient's operation.
Patients were followed up in the clinic at 6 weeks, then every 6 months for 2–3 years, and then annually. Serum testosterone and PSA were measured every 6 months. Pretreatment patient and tumor characteristics, dosimetric data, and laboratory results were entered prospectively into an institutional Access database.
Biochemical relapse was defined as a rising PSA of at least 0.4ng/mL or secondary intervention, whichever came first. This definition of biochemical relapse was used as the primary endpoint as it matched that of the comparator nomogram. Patients whose PSA “bounced” above the relapse threshold, but then declined spontaneously to <0.4ng/mL (i.e., without any intervention) were classified as nonfailing. A secondary analysis using the Phoenix (nadir+2) definition was also undertaken (12).
The following data elements were required by the nomogram (7): PSA, stage, Gleason sum, year of treatment, use of ADT, and brachytherapist experience, as defined above. The Kaplan–Meier method was used to estimate the probability of being free of PSA recurrence. Heller's method was used to calculate the expected number of recurrences at each observed failure times based on the nomogram (13). This is done by computing the difference between the observed (Brachytherapy) outcome and that expected from covariate-specific predicted outcomes of the surgical outcome. The method provides the potential for a more precise estimate of treatment activity by adjusting for patient factors, within and between study populations. The nomogram-predicted event data and the observed events provided two groups, “observed” and “nomogram predicted.” Consequently, the standard log-rank test was applied to generate significance level between the groups.
Results
One thousand two hundred fifty-four patients were eligible, according to the selection criteria. Details of their tumor and demographic characteristics are shown in Table 1. Adherence to followup was good: 6.7% of men were followed up for less than 1 year, and 11% for less than 2 years for PSA outcomes. The median followup is 55 months, which is similar to the median followup of 47 months for surgically treated men who formed the basis of the nomogram. A median of 9 followup PSA values were recorded for each patient. All but 25 patients (2%) had recovered testosterone to noncastrate levels, and of these, 17 had been followed up for less than 1 year. Eleven radiation oncologists were coded as treating the 1254 procedures, with 5 treating over 100 cases each and 4 less than 50.
Table 1. Patient characteristics
| Number [%] | 1254 [100%] |
| Age, median [range] | 66y [45–82] |
| Baseline PSA, median [range] ng/mL | 6.3 [0.3–19.6] |
| T stage (1997 tumor, node, metastasis) | |
| 1c | 577 [46.0%] |
| 2a | 534 [42.6%] |
| 2b | 143 [11.4%] |
| Gleason score | |
| 2–5 | 88 [7%] |
| 6 | 853 [67.4%] |
| 7 | 323 [25.5%] |
| 8 | 1 [0.1%] |
| Missing | 1 [0.1%] |
| Risk group (as defined in the methods) | |
| Low | 735 [58%] |
| Low-tier intermediate | 498 [39.3%] |
| Other intermediate | 32 [2.5%] |
| High | 1 [0.1%] |
| Androgen deprivation therapy used | 823 [65%] |
| Postimplant D90, median [range] | 106% [60–150%] |
| Followup, median [range] | 55mo [0–123] |
| Number of physicians | 11 |
| Brachytherapist experience as number of prior implants, median [range] | 76 [0–318] |
| Number of oncologists | |
| Less than 10 implants | 1 |
| 10–20 | 1 |
| 20–50 | 2 |
| 50–100 | 2 |
| 100–200 | 3 |
| >200 | 2 |
The actuarial observed biochemical control rate (PSA <0.4ng/mL) at 5 years is 90.6%. No patient was treated with secondary intervention in the absence of a prior biochemical relapse. The nomogram prediction of surgical outcome, given the inputted prognostic factors is 86.8% at 5, p=0.003, Fig. 2. Using the nadir+2 definition, the corresponding figures are 96% vs. 87% (5 years, PB vs. surgery).
Fig. 2
Observed prostate brachytherapy (PB) and calculated surgical PSA control rate (≤0.4
ng/mL). Log-rank test p value=0.003. Five-year biochemical no evidence of disease rates are 90.6% vs. 86.8% (PB vs. surgery).
Discussion
It is problematic to compare outcomes between treatment modalities because of differences in known and unknown prognostic variables, as well as differing definitions of endpoint. Attempts at randomized comparison of PB and RRP, which might overcome these constraints, have not been successful, with closure of the only North American trial (American College of Surgeons Oncology Group -Z0070) because of poor accrual. In the United Kingdom, a randomized trial (ProtecT) is accruing successfully but results will not be known for many years. In the meanwhile, patients are both asking the advice of physicians as to which treatment gives best cure rates and exploring their options with the aid of online tools, such as nomograms, for example, the Prostogram (14). PSA recurrence is an early and sensitive endpoint, that predicts for subsequent tumor recurrence by several years (15). As a means of comparing outcomes, it is an alternative to the use of hard clinical endpoints such as local or metastatic failure, or death from prostate cancer. Although it could be argued that these endpoints are more likely to impact quality or length of life, the collection of clinical failure is much more dependant on the frequency of clinical and radiologic examinations. The use of death as an endpoint for localized prostate cancer is particularly problematic because of the long natural history of prostate cancer, even if recurrent, and the efficacy of secondary intervention with ADT.
However, those relapse definitions that have been adopted for radiation (12) are not suitable for use with other modalities such as RRP (16). Likewise those definitions typically used after RRP (e.g., threshold definitions of 0.2 and 0.4ng/mL) [17], [18] may not be suitable for use after PB. A definition of 0.5ng/mL was explored as a possible definition for PB (19), but has been shown to have relatively poor accuracy of 42%, because of low specificity (20). This suggests that a low threshold definition (such as 0.4ng/mL) is likely to overestimate, rather than underestimate the true recurrence rate, as any remaining functioning prostate epithelium will contribute small levels of PSA. After EBRT to conventional doses, the long-term PSA value in those with no evidence of residual cancer is 1ng/mL (21). With higher radiation doses, typical nadir values are lower and prognostic for subsequent clinical long-term control (22). The median nadir in the current cohort is 0.06ng/mL.
With PB, the bounce phenomenon (23) makes the use of a threshold definition as low as 0.4ng/mL particularly problematic because 17% of PB patients will exceed this value during a bounce; and yet apparently be cured of their cancer with subsequent declining values (24). Our own experience with bounces (25) is that almost 50% of apparent relapses that trigger the nadir+2 definition of relapse are spurious. When using the nadir+2 definition of biochemical relapse, we allow the “unfailing” of such patients if their subsequent PSA level declines without intervention to less than 0.5ng/mL with further followup. In this study where the definition of failure was 0.4ng/mL, patients were not scored as failures if they had previously had a bounce and the PSA value had declined to less than 0.4ng/mL (without a subsequent rise above that level). Of those patients with a PSA bounce, 19 have not yet declined to a level of 0.4ng/mL, although all had declining levels at last followup and were less than 1ng/mL. All of these patients lack PSA values <42 months posttreatment when bounces are commonest. Clearly, we do not know for sure if they were “bounces in resolution,” although this is likely based on our analysis of the bounce phenomenon (25). If these 19 were to be reclassified as nonfailures, the 5-year biochemical no evidence of disease rate improves by 1.6%, to 92.2%. We note that the curves in Fig. 2 cross at 9 years, but this is because of a single event and only 20 patients are still at risk. It is likely therefore that the relatively smaller number of patients in the brachy group (1254 vs. 7723) gives rise to this effect, rather than a sudden drop in PSA control at this late time point.
The surgical nomogram that was used in this study incorporated the novel additional factor of surgeon experience, which has been shown in numerous studies to be associated with toxicity, margin status, and PSA recurrence [8], [26]. A recent report also showed a learning curve in PB (9). Our published results show a clear trend toward higher average dose metrics (dose in gray covering 90% of the prostate [D90] and volume of the prostate covered by the 100% isodose [V100]) with increasing implant order (as a measure of institutional experience), but there was no correlation-linking implant order and the risk of biochemical or clinical relapse (11). We have not, however, specifically examined for outcome by individual oncologist experience. With EBRT, an association of biochemical outcome by provider volume has been reported (27) and has been recently verified by us (28). The surgical results are from four “world-class” institutions, and therefore might be regarded as reflecting the “best” of surgical outcomes, which might not be obtainable by a typical individual surgeon. Although that may also be true of our brachytherapy program, the results published from the British Columbia Cancer Agency are similar to others in the literature, suggesting that the outcomes we have obtained can be accomplished by others with carefully managed programs with established quality assurance procedures in place. Other centers can also make use of published nomograms to compare their outcomes. This can be carried out by individually entering each patient's prognostic features and obtaining a predicted output, or for large datasets, by arranging for the source code to be made available, in the manner that we have done in this study.
A limitation of this study is the endpoint of PSA control, which as discussed above, has challenges that cannot be easily overcome. Although a “hard” endpoint such as secondary intervention with ADT is a good alternative endpoint, it is also subject to biases such as lack of standardization of criteria for its use. In a low-risk population, it is also less sensitive as a measure of the true cure of the tumor, where PSA relapse almost always precedes other forms of relapse by several years. Late endpoints such as the development of symptomatic progressive disease are even less sensitive, although of major importance to patients. Patients treated surgically were treated a median of 4 years earlier than those with brachy (1998 vs. 2002). However, year of treatment was modeled as an independent factor in the nomogram development, and so this difference should not influence the results. Strengths of this study include the population-based program of PB and the inclusion of all sequential patients from the first implant with no exclusions. Our data had large numbers of patients, as does the use of the nomogram based on over 7000 patients from four institutions. Both employed physicians with wide variation in expertise, which was modeled by the nomogram.
In conclusion, the results of this study suggest that PSA control with low-dose-rate PB can achieve results that surpass those of RRP. Reports of quality of life after treatment suggest that the long-term impact of treatment, whether PB or RRP can be significant, with some toxicities being more common after RRP (urinary incontinence, erectile dysfunction) and others more common after PB (urinary irritation, bowel problems) (29). Individuals will continue to balance these trade-offs in their own decision making.
Acknowledgments
The nurses, physicists, and therapists who, with the physicians form the team in carrying out our program. Additional physicians not named as study authors, Drs. M. McKenzie, MD, A. Agranovich, MD, J. Wu, MD, H. Pai, MD, W. Kwan, MD, M. Liu, MD, E. Berthelet, MD, and H. Campbell MD.
References
- Population-based study of biochemical and survival outcomes after permanent 125I brachytherapy for low- and intermediate-risk prostate cancer. Urology. 2009;73:860–865discussion 865–867
- 12-Year outcomes following permanent prostate brachytherapy in patients with clinically localized prostate cancer. J Urol. 2008;179:S20–S24
- Overall information needs of early-stage prostate cancer patients over a decade: Highly variable and remarkably stable. Support Care Cancer. 2009;17:429–435
- American Society of Clinical Oncology recommendations for the initial hormonal management of androgen-sensitive metastatic, recurrent, or progressive prostate cancer. J Clin Oncol. 2004;22:2927–2941
- Prostate cancer nomograms: An update. Eur Urol. 2006;50:914–926discussion 926
- Comparison of nomograms with other methods for predicting outcomes in prostate cancer: A critical analysis of the literature. Clin Cancer Res. 2008;14:4400–4407
- Preoperative and postoperative nomograms incorporating surgeon experience for clinically localized prostate cancer. Cancer. 2009;115:1005–1010
- Surgeon experience is strongly associated with biochemical recurrence after radical prostatectomy for all preoperative risk categories. J Urol. 2008;179:2212–2216discussion 2216–2217
- Provider case volume and outcomes following prostate brachytherapy. J Urol. 2009;181:113–118discussion 118
- Cancer UIClCatAJCo, TNM classification of malignant tumors. New York: John Wiley & Sons; 1992, 1997, 2002;
- Evaluation of dosimetric parameters and disease response after 125 iodine transperineal brachytherapy for low- and intermediate-risk prostate cancer. Int J Radiat Oncol Biol Phys. 2009;73:1432–1438
- Defining biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: Recommendations of the RTOG-ASTRO Phoenix Consensus Conference. Int J Radiat Oncol Biol Phys. 2006;65:965–974
- . Improving the decision to pursue a phase 3 clinical trial by adjusting for patient-specific factors in evaluating phase 2 treatment efficacy data. Med Decis Making. 2007;27:380–386
- . Available at: http://www.mskcc.org/applications/nomograms/prostate/PreTreatment.aspx2009;Last accessed May 2010
- Time to biochemical failure and prostate-specific antigen doubling time as surrogates for prostate cancer-specific mortality: Evidence from the TROG 96.01 randomised controlled trial. Lancet Oncol. 2008;9:1058–1068
- Is it possible to compare PSA recurrence-free survival after surgery and radiotherapy using revised ASTRO criterion—“nadir + 2”. Urology. 2008;72:389–393discussion 394–395
- Defining biochemical recurrence of prostate cancer after radical prostatectomy: A proposal for a standardized definition. J Clin Oncol. 2006;24:3973–3978
- Variation in the definition of biochemical recurrence in patients treated for localized prostate cancer: The American Urological Association Prostate Guidelines for Localized Prostate Cancer Update Panel report and recommendations for a standard in the reporting of surgical outcomes. J Urol. 2007;177:540–545
- Comparison of biochemical failure definitions for permanent prostate brachytherapy. Int J Radiat Oncol Biol Phys. 2006;65:1487–1493
- . Practical application of biochemical failure definitions: What to do and when to do it. Int J Radiat Oncol Biol Phys. 2002;53:304–315
- 10-Year outcome for men with localized prostate cancer treated with external radiation therapy: Results of a cohort study. J Urol. 2004;171:210–214
- PSA nadir predicts biochemical and distant failures after external beam radiotherapy for prostate cancer: A multi-institutional analysis. Int J Radiat Oncol Biol Phys. 2006;64:1140–1150
- Caloglu M, Ciezki J. Prostate-specific antigen bounce after prostate brachytherapy: Review of a confusing phenomenon. Urology 2009;74:1183–1190.
- . Prostate-specific antigen bounce after prostate seed implantation for localized prostate cancer: Descriptions and implications. Int J Radiat Oncol Biol Phys. 2003;56:448–453
- Thompson A, Keyes M, Pickles T, et al. Evaluating the differences in prostate-specific antigen (PSA) kinetics in PSA failure versus PSA bounce after permanent I125 prostate brachytherapy. Int J Radiat Oncol Biol Phys. In press.
- The surgical learning curve for prostate cancer control after radical prostatectomy. J Natl Cancer Inst. 2007;99:1171–1177
- Validation of the contemporary Epstein criteria for insignificant prostate cancer in European men. Eur Urol. 2008;54:1306–1313
- Stokes E, Tyldesley S, Pickles T. Oncologist case load volume and disease outcomes after curative external beam radiotherapy for localized prostate cancer. In Joint ECCO 15-34TH ESMO Multidisciplinary Congress. Berlin, Germany: European Journal of Cancer Supplements; 2009.
- Quality of life and satisfaction with outcome among prostate-cancer survivors. N Engl J Med. 2008;358:1250–1261
PII: S1538-4721(10)00234-5
doi:10.1016/j.brachy.2009.12.004
Crown Copyright © 2011. Published by Elsevier Inc. All rights reserved.
