This article is mentioned in:

Introduction

In 2012, prostate cancer was the most common cancer among European men, comprising 12% of all new cancer cases and 5% of all cancer deaths. This incidence was the highest in high-income countries (1). Treatment of localized disease varies, as there is no evidence that one treatment is more effective compared with another (2). According to the current guidelines of the American Urological Association (3) and the European Urological Association (4), the treatment of low-risk, localized prostate cancer may include active surveillance, prostatectomy, or either interstitial or external-beam radiation therapy. Permanent seed brachytherapy (BT) is mostly used for patients with low prostate-specific antigen (PSA) concentrations and low Gleason scores. For these patients, studies have shown no significant difference in clinical effectiveness among treatments (5). According to previous studies, the 5-year disease-free survival is ~80–90% (5,6).

The aim of this study was to evaluate risk factors for biochemical failure (BF) after permanent seed 125I BT for prostate cancer among patients treated in Oulu University Hospital. The study consisted of a retrospective chart review conducted to evaluate clinical characteristics at diagnosis, treatment-related details and follow-up data.

Materials and methods

Chart review

Between March, 2001 and December, 2014, 607 patients received treatment for early-stage (T1/2N0M0), histologically confirmed prostate cancer using 125I seed BT. Study data were collected retrospectively from their medical records at Oulu University Hospital (Oulou, Finland) and at other hospitals for those who were followed up elsewhere, including local hospitals in Kajaani, Kemi, Kokkola, Oulainen and Rovaniemi. The following data were recorded: Patient age, former diagnosis, Gleason score, TNM stage, PSA concentration prior to treatment, PSA nadir concentration, possible increase in PSA concentration during follow-up, all additional treatments for prostate cancer before or after BT, and possible radiological progression. Treatment outcome was defined in terms of time to PSA nadir concentration, BF, defined as PSA concentration progression >2 µg/l from the PSA nadir concentration according to the Phoenix criteria (7), treatment of BF, and overall and prostate-cancer specific survival. PSA nadir concentration was defined as the lowest PSA concentration observed after BT. After treatment, the patients were followed up at 3, 6 and 12 months, then biannually for 4 years, and annually thereafter. More frequent follow-ups were possible based on the decisions of the urologists/doctors responsible for the follow-ups. There were no strict treatment protocols following BF.

Ethics

According to Finnish legislation and directions from Finnish Ethics Committees, this chart review was exempted from formal approval by the Institutional Review Board. However, the study was conducted according to the principles of the Helsinki Declaration.

BT technique

The dose plan was based on ultrasonic (US) images taken at 0.5-cm intervals and used stranded seeds (125I IsoCord®; Bebig GmbH, Berlin, Germany). The implantation technique was intraoperative and used sagittal images. VariSeed 8.0 (Varian; Palo Alto, CA, USA) was used for dose planning and for calculating the dose-volume histogram. The actual dose plan was performed using 125I IsoCord S06 seeds (Bebig GmbH). The prescribed dose was 145 Gy using seed activity of 17.46 MBq (0.472 mCi).

Statistical analysis

Unless otherwise stated, the summary statistics included the mean, range and standard deviation (SD) or, if biased, the median with the 25–75th percentile. Survival analyses and analyses of time to BF were conducted using the Kaplan-Meier method, and the statistical significance of the differences between groups was analysed using the log-rank test. The data were analyzed using SPSS statistical software, version 22.0 (IBM SPSS; Armonk, NY, USA). Two-tailed P-values are reported, and P-values <0.05 were considered to indicate statistically significant differences.

Results

Patient characteristics

Follow-up details were available for 605 patients, and survival data were available for 606 patients. The mean age at BT was 64 years (range, 44–78; SD, 6.1). The median PSA concentration prior to BT was 7.7 (25–75th percentiles, 5.5–10.0). The median prostate volume was 29.0 cm3 (25–75th percentiles, 23.3–35.9). The median number of seeds used was 55 (25–75th percentiles, 47–63). The median number of needles used was 21 (25–75th percentiles, 19–23). Of the 607 patients, 537 (88.5%), 69 (11.4%) and 1 (0.2%) had Gleason scores of 6, 7 and 8, respectively, on prostate biopsy. The clinical T-stage according to the TNM classification (8) was cT1 and cT2 in 400 (65.9%) and 207 (34.1%) patients, respectively. Prior to BT, 91 patients (15%) received neoadjuvant hormonal therapy with the intention of decreasing the prostate volume to <50 cm3. The median duration of this neoadjuvant hormonal therapy was 5 months (25–75th percentiles, 3–9).

The median follow-up time was 81 months (range, 2–161; SD, 39). After a mean of 46 months (range, 8–136; SD, 32), 117 (19.3%) patients developed BF and 91 (15.0%) were treated for BF. For those patients, the mean time to the first treatment after BT was 57 months (range, 12–149; SD, 34). The first treatment selected was antiandrogens (n=47, 51.6%), chemical castration (n=29, 31.9%), external-beam radiation therapy (n=11, 12.1%) or radical prostatectomy (n=2, 2.2%). In addition, as an initial treatment for BF, 2 patients (2.2%) were treated with a 5-α reductase inhibitor. During follow-up, 32 (5.3%) patients received external-beam radiation therapy. Eventually, 21 (3.5%) developed metastatic disease, the mean time to which was 66 months (range, 18–149; SD, 43). Of the 606 patients. 12 succumbed to prostate cancer and 47 to other causes. The overall survival and prostate-cancer specific survival rates were 90.3 and 98.0%, respectively. The mean overall survival time and prostate-cancer specific survival time were 147 and 158 months, respectively.

BF was significantly more common among patients with PSA nadirs ≥0.5 µg/l compared with those with PSA nadirs <0.5 µg/l (Fig. 1). The mean time to BF was 55 [95% confidence interval (CI): 47–63] and 114 (95% CI: 112–116) months for patients with PSA nadirs ≥0.5 µg/l and <0.5 µg/l, respectively (P<0.001).

Figure 1. Association of PSA nadir value after permanent seed BT with BF-free survival among prostate cancer patients. BF was common among patients with PSA nadir concentrations ≥0.5 µg/l after BT (P<0.001). PSA, prostate-specific antigen; BT, brachytherapy; BF, biochemical failure.

We further evaluated the potential association of underlying diseases with BF. BF tended to develop more rapidly among patients with hypertension or diabetes at diagnosis. For patients without and with hypertension, the mean time to BF was 104 (95% CI: 100–107) and 98 (95% CI: 93–103) months, respectively (P=0.035). For patients without and with diabetes, the mean time to BF was 103 (95% CI: 100–106) and 89 (95% CI: 77–102) months, respectively (P=0.006). Other comorbidities, including coronary artery disease, previous myocardial infarctions, other cancers, or obstructive pulmonary disease, did not affect the time to BF. Due to the small number of deaths, further survival analyses for overall survival and prostate-cancer specific survival were not performed.

Discussion

BT is an option for treating low-risk prostate cancer, and it has achieved promising results in terms of disease recurrence (9–18); however, certain patients still develop disease recurrence. In search of the risk factors for BF after BT, PSA nadir was identified as a strong predictor of BF. For patients with PSA nadirs ≥0.5 µg/l, the mean time to BF was <6 years, whereas it was >9 years for patients with PSA nadirs <0.5 µg/l. Recently, McLaren et al (19) published their institutional results, reporting that PSA concentrations of >0.4 µg/l over the PSA nadir predict disease recurrence. In addition, their analyses demonstrated that, if their PSA nadir was >0.8 µg/l (19), approximately half of the patients experienced disease relapse within 4 years, a finding that is in line with the results of the present study.

Comorbidities, including hypertension and diabetes, have previously been associated with decreased overall survival following BT (20,21). Despite the promising results of lower mortality among prostate cancer patients exposed to metformin (22), an analysis of 270 men with diabetes, with and without metformin use, revealed no association between metformin use and progression-free, disease-free, or overall survival following treatment with BT (23). Similar results were found in another cohort of 199 diabetes patients, in which diabetes did not affect cancer-specific survival or biochemical progression following BT (24). However, increased blood glucose levels have been associated with increased risk of disease recurrence following radical prostatectomy or radiation therapy as a radical treatment for prostate cancer (25). The finding of the present study regarding a tendency toward more rapid BF among patients with underlying diabetes or hypertension is partly supported by previously published results in cases with diabetes (25). To the best of our knowledge, hypertension has not been associated with progression-free survival in prostate cancer. However, the findings of the present study require evaluation in other cohorts. In addition, the presence of a clinically meaningful association with BF remains obscure, as the mean time to BF was long (89–104 months).

Our study was limited by its retrospective, single-center nature. In addition, the study did not evaluate diabetes medications or the use of statins, although they may have a positive prognostic effect (26). Furthermore, the small number of deaths did not enable reliable survival analyses. Moreover, BF was defined according to the Phoenix criteria, which have been shown to be more sensitive and specific in defining BF in patients treated with BT. However, it has been demonstrated that the ASTRO and Phoenix criteria (7) have a ~8% difference in the rate of biochemical control, with the latter achieving lower values (27). Finally, comparing among various studies is difficult due to the differing definitions used.

However, the results of the present study indicate that low PSA nadir reliably predicts BF after BT. Patients with underlying hypertension or diabetes tended to exhibit shorter times to BF, emphasizing the need for more attentive follow-up of such patients after BT. However, these findings require further investigation.

Acknowledgements

We are grateful to Leena Heikkilä for assisting with data collection, and to Pasi Ohtonen, MSc, for assisting with statistical analyses.

References