Nearly three-quarters of a million American men who have been treated with prostatectomy and/or radiation therapy experience an increasing prostate-specific antigen level known as biochemical recurrence. Although androgen-deprivation therapy remains a reasonable option for some men with biochemical recurrence, deferring androgen ablation or offering nonhormonal therapies may be appropriate in patients in whom the risk of clinical or metastatic progression and prostate cancer–specific death is low. A risk-stratified approach informed by the patient’s prostate-specific antigen kinetics, comorbidities, and personal preferences is recommended to determine the best management approach.
Key points
- •
Nearly three-quarters of a million American men who have been treated with prostatectomy and/or radiation therapy experience an increasing prostate-specific antigen (PSA) level, a condition known as biochemical recurrence (BCR).
- •
Post localized therapy, some of these men develop distant metastases with time, but many years may pass before signs of clinical progression appear.
- •
Although androgen-deprivation therapy remains a reasonable option for some men with BCR, deferring androgen ablation or offering nonhormonal therapies may be appropriate in patients where the risk of clinical/metastatic progression and prostate cancer–specific death is low.
- •
Drug development in this space is a challenge because of the heterogeneous and prolonged natural history of biochemically recurrent prostate cancer, and the lack of short-term, validated surrogate end points for overall survival.
Introduction
Approximately 239,000 men will be diagnosed with prostate cancer in 2013, but 88% of these men will ultimately die from ischemic heart disease or other nonprostate cancer causes. An estimated 60,000 to 70,000 men are diagnosed in the United States each year with biochemical recurrence (BCR), a state defined as rising prostate-specific antigen (PSA) after radical prostatectomy (RP) or radiation treatment, and overall, three-quarters of a million men are estimated to be living with rising PSA after local therapy without evidence of overt metastatic disease.
The optimal management of patients with nonmetastatic, hormone-naive, biochemically relapsed prostate cancer remains largely unestablished at this time because of the lack of prospective randomized trials designed to address standards of care. Treatment decisions remain largely intuitive at the present time. Recognizing the deficiency, this article describes a logical risk-based approach for therapeutic considerations and clinical research in this relatively common subset of patients. The approach is based on extensive data on the natural history of these patients at the Johns Hopkins Hospital (JHH). This article discusses this and other existing datasets and defines potential risk-benefit ratios of existing modalities of treatment.
Introduction
Approximately 239,000 men will be diagnosed with prostate cancer in 2013, but 88% of these men will ultimately die from ischemic heart disease or other nonprostate cancer causes. An estimated 60,000 to 70,000 men are diagnosed in the United States each year with biochemical recurrence (BCR), a state defined as rising prostate-specific antigen (PSA) after radical prostatectomy (RP) or radiation treatment, and overall, three-quarters of a million men are estimated to be living with rising PSA after local therapy without evidence of overt metastatic disease.
The optimal management of patients with nonmetastatic, hormone-naive, biochemically relapsed prostate cancer remains largely unestablished at this time because of the lack of prospective randomized trials designed to address standards of care. Treatment decisions remain largely intuitive at the present time. Recognizing the deficiency, this article describes a logical risk-based approach for therapeutic considerations and clinical research in this relatively common subset of patients. The approach is based on extensive data on the natural history of these patients at the Johns Hopkins Hospital (JHH). This article discusses this and other existing datasets and defines potential risk-benefit ratios of existing modalities of treatment.
Definition of BCR
The definition of BCR after local therapy varies based on the primary modality of treatment. After surgery, PSA levels greater than 0.2 ng/mL or greater than 0.4 ng/mL and rising are often considered evidence of BCR. However, in 2007, the American Urological Association (AUA) reported on a review of more than 13,000 citations referencing BCR in patients with prostate cancer and found 54 different definitions of BCR after surgery and 99 different definitions of BCR after radiation therapy (RT). The lack of consistently applied definitions of BCR limits the interpretation of data on natural history and some of the therapeutic considerations in these patients. Such inconsistencies are especially challenging for the interpretation and design of clinical trials.
BCR After RP
Among the 54 definitions of BCR after prostatectomy discovered by the AUA researchers, the most common was a PSA of greater than 0.2 ng/mL or a close variation. The authors, who were also members of the AUA Prostate Guideline Update Panel, recommended that practitioners use a single definition of BCR after RP as follows:
It is recommended that biochemical (PSA) recurrence following radical prostatectomy be defined as a serum PSA of 0.2 ng/mL or greater, with a second confirmatory level of PSA of >0.2 ng/mL. The first postoperative PSA should be obtained between 6 weeks and 3 months following therapy. The date of failure should be defined as the date of the first detectable PSA level once this value has been confirmed.
In establishing this recommended definition, however, the panel added two caveats. First, the higher levels of PSA (>0.4 ng/mL) would have much greater specificity for clinical and/or radiographic recurrence and progression, but the authors justified the use of 0.2 ng/mL by arguing it had “provided high sensitivity for recurrence as well as the greatest generalizability.” Second, this definition is not an effective predictor of death from prostate cancer, suggesting that prognosis should be based on nomograms that consider Gleason score and PSA kinetics, although available nomograms have not been prospectively validated.
The panel acknowledged the appropriateness of reporting biochemical outcomes using additional PSA thresholds. Although some researchers who have designed clinical trials enrolling patients with BCR have adopted this AUA definition, other researchers use 0.4 ng/mL and rising as the eligibility criterion, arguing that the higher value is more specific for future risk of clinical and radiologic progression. For trial purposes, the authors consider PSA levels greater than 0.4 ng/mL and rising as evidence of BCR after surgery.
BCR After RT
The AUA researchers found 99 definitions of BCR after RT, among which the most common was the American Society of Therapeutic Radiology and Oncology (ASTRO) definition: “the mid-point between PSA nadir and the first of three consecutive rises in PSA.” Despite a recommendation by the AUA for its adoption, the ASTRO definition is problematic because it requires backdating the time of BCR and because of its failure to use the PSA level at nadir as a risk factor. The Phoenix definition (“nadir + 2 ng/mL, with the failure date defined as the date the rise in PSA is noted”) offers greater accuracy in predicting future clinical failures compared with the ASTRO definition. However, the Phoenix definition provides substantially lower estimates of BCR at 5 years and substantially higher estimates of BCR at 10 years than the ASTRO definition. The choice of definition can impact the findings of adjuvant treatment trials; the ASTRO definition would be likely to show a greater number of people experiencing BCR after treatment. The AUA retains its recommendation to use the ASTRO definition of BCR after RT alone (no hormonal therapy), whereas the Phoenix definition can be applied after RT with or without hormonal therapy, and avoids the need for backdating. As a result the Phoenix definition has gained wider acceptance in determining BCR for clinical decisions. Patients whose PSA dynamics meet the Phoenix definition and who have evidence of rapid PSA doubling time (PSADT; less than 10 months) may initiate treatment, whereas patients with evidence of slow PSADT may defer further therapy. To ensure comparability with future trial results, it is important to continue to use the Phoenix definition to set criteria for trial entry for patients with BCR after RT.
Natural history of patients with prostate cancer experiencing BCR after surgery or RT
The natural history of patients with prostate cancer at JHH with evidence of BCR after RP has been extensively reported over the past 15 years. In 1999, Pound and colleagues first described the natural history of 304 men who had undergone RP at Johns Hopkins between April 1982 and April 1997, who demonstrated subsequent evidence of BCR (PSA≥0.2 ng/mL), and who did not receive androgen deprivation therapy (ADT) until there was evidence of metastatic disease. The researchers found that the median time from BCR to metastasis in all patients was 8 years and the median time from metastasis to death was 5 years. In this report, time to biochemical progression, pathologic Gleason score, and PSADT were significant predictors of metastasis-free survival (MFS) over 3-, 5-, and 7-year periods. Freedland and colleagues published a study in 2005 describing cancer-specific mortality of the same Johns Hopkins prostate cancer cohort, with 5 additional years of follow-up. Of 379 patients who had experienced BCR, 66 died from prostate cancer, and the overall median survival had not been reached after 16 years. This analysis found that the same three factors—PSADT (<3, 3–8.9, and 9–14.9, and ≥15 months), Gleason score (≤7 vs 8–10), and time from surgery to BCR (<3 vs >3 years)—were significant predictors of prostate cancer–specific mortality. In the most recent analysis of this same cohort, Antonarakis and colleagues provided updated information on the natural history and markers predictive of MFS in an expanded cohort of patients from the JHH database including 450 men with BCR after RP, and reported that median overall MFS was 10 years after BCR ( Table 1 ). However Antonarakis and colleagues also found that only two of the predictors reported previously were significantly predictive of MFS: PSADT (<3, 3–8.9, and 9–14.9, and ≥15 months) and Gleason score (<6 vs 7 vs 8–10) ( Table 2 ). In this updated analysis, time to BCR was not a significant predictor of metastasis. Several other analyses identified PSADT as a primary predictor of MFS and prostate cancer–specific mortality (see Table 2 ).
| Median MFS, y (95% CI) | Metastasis-free Rate at 5 y, % (95% CI) | Metastasis-free Rate at 10 y, % (95% CI) | |
|---|---|---|---|
| Pathologic Gleason Score | |||
| 8–10 | 4 (2, 6) | 43 (32, 54) | 19 (9, 33) |
| 7 | 11 (9, >17) | 71 (63, 78) | 52 (41, 62) |
| 4–6 | >15 (14, >15) | 94 (86, 98) | 94 (86, 98) |
| PSADT | |||
| <3 mo | 1 (0, 1) | 5 (1, 21) | n/a |
| 3–9 mo | 4 (2, 4) | 27 (16, 39) | 7 (1, 22) |
| 9–15 mo | 13 (6, >15) | 77 (63, 86) | 51 (34, 66) |
| >15 mo | 15 (15, >17) | 91 (85, 95) | 72 (59, 83) |
| Study | N (BCR/Total Number RP and/or RT) | End Point (Number with M1/Number PCSD) | MFS by PSADT Subgroup HR or MFS (95% CI) | PCSM by PSADT Subgroup HR or Mortality (95% CI) | OS by PSADT Subgroup HR (95% CI) | Significant Predictors |
|---|---|---|---|---|---|---|
| Primary treatment: radical prostatectomy | ||||||
| Pound et al, 1999 | 304/1997 | MFS (103 M1 pts of 304) | PSADT <10 mo predictive of MFS ( P <.001) | na | na | PSADT, Gleason score, time to BCR |
| D’Amico et al, 2003 | 611/5918 | PCSM, OS (154 ACD and 111 PCSD for RT + PR) | na | <3 mo HR 62.9 18.8–210.1) | <3 mo HR 18.2 (8.9–37.2) | PSADT |
| >3 mo HR 0.61 (0.51–0.73) | >3 mo HR 0.84 (0.78–0.90) | |||||
| Freedland et al, 2005 | 379/na | PCSM (66 PCSD of 379) | na | <3 mo: HR 27.5 (CI, 10.7–70.9) 3–8.9 mo: HR 8.8 (CI, 3.7–20.5) 9–14.9 mo: HR 2.44 (CI, 0.9–6.8) ≥15 mo: HR 1 | na | PSADT, Gleason score, time to BCR |
| Zhou et al, 2005 | 498/8669 (RT + RP) | PCSM (25 PCSD of 498) | na | <3 mo PCSM 31% (17%–45%) ≥3 mo PCSM 1% (0%–2%) | na | PSADT |
| Antonarakis et al, 2012 | 450/na | MFS (134 M1 pts of 450) | Median MFS: <3 mo: 1 mo (CI, 0–1) 3–8.9 mo: 4 mo (CI, 2–4) 9–14.9 mo: 13 mo (CI, 6 to >15) ≥15 mo: 15 mo (CI, 15 to >17) | na | na | PSADT, Gleason score |
| Antonarakis et al, 2011 | 346/na | MFS, OS (39 M1 pts of 190 with MFS data) (63 PCSD of 346) | <3 mo: HR 7.77 (CI, 2.65–22.76) | na | <3 mo: HR 27.4 (8.70–86.38) | PSADT |
| 3–8.8 mo: HR 1.95 (CI, 1.04–3.66) | 3–8.8 mo: HR 6.16 (3.00–12.64) | |||||
| ≥9 mo: HR 1 | ≥9 mo: HR 1 | |||||
| Primary treatment: radiotherapy | ||||||
| D’Amico et al, 2002 | 94/381 | PCSM (20 PCSD of 94) | na | ≤12 mo predictive of time to PCSD ( P = .003) | PSADT ≤12 mo predictive of time to ACD ( P = .02) | PSADT, delayed use of ADT |
| D’Amico et al, 2003 | 840/2751 | PCSM, OS (154 ACD and 111 PCSD for RT + PR) | na | <3 mo HR 12.2 (7.5–20.1) | <3 mo HR 4.8 (3.4–7) | PSADT |
| >3 mo HR 0.83 (0.78–0.87) | >3 mo HR 0.95 (0.93–0.98) | |||||
| Zhou et al, 2005 | 661/8669 (RT + RP) | PCSM (77 PCSD of 661) | na | <3 mo PCSM 75% (59%–92%) ≥3 mo PCSM 35% (24–47) | na | PSADT, Gleason score |
| Buyyounouski et al, 2008 | 211/1578 | MFS, PCSM (53 M1 pts of 211) (29 PCSD on 211) | <3 mo MFS HR 2.87 ( P = .001) | na | na | IBF alone for PCSM |
| ≥3 mo MFS HR 1 | IBF, Gleason score, PSA nadir, PSADT for MFS | |||||
Related posts:
Surgical Management of Prostate Cancer
Management of Patients with Biochemical Recurrence After Local Therapy for Prostate Cancer
Management of Hormone-Sensitive Metastatic Prostate Cancer
Contemporary Issues in Radiotherapy for Clinically Localized Prostate Cancer
Imaging in Prostate Carcinoma
Neoadjuvant and Adjuvant Hormonal and Chemotherapy for Prostate Cancer
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
