Takefumi Satoh, Kitasato University School of Medicine – Urology, 1-15-1 Kitasato, Sagamihara, Kanagawa 228-8555, Japan. e-mail: firstname.lastname@example.org
To examine the incidence, timing, and magnitude of the prostate-specific antigen (PSA) level ‘bounce’ after permanent prostate brachytherapy (BT) and correlate the PSA bounce with clinical and dosimetric factors in Japanese patients with prostate cancer.
PATIENTS AND METHODS
A multi-institutional pooled analysis was carried out in 388 consecutive patients with T1–T2N0M0 prostate cancer treated with 125I-seed implant BT with no hormonal therapy or external beam radiotherapy. All patients had ≥1 year of follow-up and at least three follow-up PSA level measurements. Three definitions of PSA bounce were used: definition A, a PSA level rise of 0.1 ng/mL; definition B, a PSA level rise of 0.4 ng/mL; and definition C, a PSA level rise of 35% over the previous value, followed by a subsequent fall.
The actuarial likelihood of having PSA bounce at 24 months was 50.8% for definition A, 23.5% for definition B, and 19.4% for definition C. The median time to develop PSA bounce was 12 months for definition A, 18 months for definition B, and 18 months for definition C. There was a PSA bounce magnitude of 2 ng/mL in 5.3% of patients, and 95.3% of PSA bounce occurred within 24 months after 125I-BT. Among the before and after 125I-BT factors, clinical stage, initial PSA level, and Gleason score did not predict for PSA bounce using any definition; only being younger predicted for PSA bounce on multivariate analysis (P < 0.001).
PSA bounce is a common phenomenon after 125I-BT and occurred at a rate of 19–51% in the Japanese men who underwent 125I-BT, depending on the definition used. It is more common in younger patients, and early PSA bounce should be considered when assessing a patient with a rising PSA level after 125I-BT, before implementing salvage interventions. Furthermore, PSA bounce magnitude might be lower in Japanese than in Caucasian patients.
prostate volume receiving at least 100% dose at 1 month
prostate volume receiving at least 150% dose at 1 month
Prostate permanent brachytherapy (BT) is currently available and becoming a commonly used method for treating prostate cancer in Japan. One of the most appealing reasons for selecting this treatment is the low degree of toxicity, which may reduce the impact on quality of life in the patient [1,2]. However, PSA kinetics after this treatment is different from that after radical prostatectomy. It is generally thought that the PSA level continuously decreases after potentially successful radiation for prostate cancer and that an increase in PSA levels might reflect disease recurrence. However, physicians often observe a temporary increase in PSA levels after radiation therapy for prostate cancer, which does not reflect disease recurrence. This phenomenon is called PSA ‘bounce’ (spikes), and 30–50% of patients treated with BT have PSA bounce [3,4].
The PSA bounce is thought to be the result of compromised membrane integrity in the PSA-producing epithelium and has not proved to be of prognostic consequence in patients treated with BT . However, still unclear, are not only the mechanism of PSA bounce, but also ethnic/racial differences of PSA bounce after BT.
The present study is the first to examine PSA bounce after BT alone, with no hormonal therapy or external beam radiotherapy (EBRT), in Japanese men with prostate cancer. In addition, the effect of pretreatment and treatment-related factors on PSA bounce were analysed to determine which patients were at increased risk of developing PSA bounce.
PATIENTS AND METHODS
In all, 388 patients underwent an 125I-permanent implant. No patient received hormonal therapy or EBRT. Eligible patients for the study had to have ≥1 year of follow-up and at least three PSA measurements after treatment. The PSA level was measured at 1 month after implant and at 3-month follow-up intervals within 2 years after implant, and 6-month follow-up intervals thereafter.
The prescribed dose to the periphery of the prostate was 145 Gy using a prostate implant technique that has been previously described [1,6]. Both before and after implant analyses used a radiotherapy planning system dedicated for transperineal interstitial permanent prostate BT (Interplant version 3.4 CMS or Variseed version 7.1 VARIAN), and all doses were defined using TG43 criteria . At 1 month after implant, a CT-based dosimetric analysis was performed. Dosimetry data were available for all patients after implant. To analyse the effect of dose on PSA bounce, dose was defined as the dose delivered to 90% of the gland on the 1-month after implant dose-volume histogram (D90).
A PSA bounce was defined as an elevation of the PSA level (at any time an elevation occurred, excluding the 1-month PSA value) from the previous value with a subsequent fall. Three descriptions of a PSA bounce were used: definition A, a rise of 0.1 ng/mL; definition B, a rise of 0.4 ng/mL; and definition C, a rise of >35% over the previous value .
Associations between PSA bounce and the before treatment and treatment-related factors were examined using logistic regression analysis.
The patient clinical disease characteristics and dosimetric variables are shown in Table 1. The median percentage of the prostate volume receiving a minimum of 100% of the prescribed dose (V100) was 94.7%, and the median minimum D90 was 158.9 Gy. The median (range) follow-up for the entire group was 30 (12–42) months. The actuarial likelihood of having PSA bounce at 24 months was 50.8% for definition A, 23.5% for definition B, and 19.4% for definition C (Fig. 1).
Table 1. The patients’ characteristics before treatment and dosimetric variables after treatment
Initial PSA level, ng/mL
Prostate volume, cm3
The distribution of the time of onset of PSA bounce is shown in Fig. 2A, based on definition A. The median time to PSA bounce was 12 months for definition A, 18 months for definition B, and 18 months for definition C; 95.3% (164/172) of PSA bounce was seen within 24 months after 125I-BT for definition A.
The median (range) before bounce PSA level was 1.09 (0.11–5.74) ng/mL, and the median (range) bounce height (increase above the before bounce level) was 0.40 (0.1–6.62) ng/mL. The magnitudes of the bounce were 0.1–0.2 ng/mL in 26.7%, 0.2–0.4 ng/mL in 22.1%, 0.4–1.0 ng/mL in 35.4%, 1.0–2.0 ng/mL in 10.5%, 2.0–5.0 ng/mL in 4.1%, and >5 ng/mL in 1.2% (Fig. 2B).
The effects of patient age, initial PSA level, Gleason score, clinical stage, prostate volume, PSA density, D90, V100, and the prostate volume receiving at least 150% dose at 1 month (V150) on developing a PSA bounce were tested using univariate analysis with each of the three definitions (Table 2). Using definition A, age, prostate volume, and PSA density (PSAD) significantly affected the incidence of bounce (P < 0.001, P = 0.018, P = 0.009, respectively). Using definition B, patient age and prostate volume significantly affected the incidence of bounce (P < 0.001, P = 0.041, respectively). Using definition C, patient age also significantly affected the incidence of bounce (P = 0.001). The multivariate Cox regression analysis determined predictors for PSA bounces are listed in Table 3. On multivariate analysis, only patient age predicted a PSA bounce (P < 0.001) using definition A. Using definition B, patient age and prostate volume predicted PSA bounce on multivariate analysis (P < 0.001, P = 0.029, respectively).
Table 2. Univariate analysis of factors potentially affecting PSA bounce
Initial PSA, ng/mL
Prostate volume, cm3
Table 3. Multivariate analysis of factors potentially affecting PSA bounce
It is well known that PSA is a kallikrein III, seminin, semenogelase, ã-seminoprotein, and P-30 antigen and is a 34-kDa glycoprotein manufactured by the cells of the prostate gland. PSA is present in small quantities in the serum of normal men and is often elevated in the presence of prostate cancer and other prostate disorders. The Hybritech Tandem-R PSA test, released in 1986 , is a blood test to measure PSA concentration and is widely used not only for detection of prostate cancer, but also as a follow-up tool after treatment.
It is also well known that PSA level changes after prostate BT show a gradual decline, unlike radical prostatectomy. The PSA values may take >5 years to reach a nadir . The intact prostate, with its normal and abnormal epithelial cells, is the source for these PSA fluctuations. A PSA bounce after prostate BT causes anxiety not only in patients, but also in physicians.
PSA bounce after permanent seed BT has been reported by several investigators, and it is common, occurring in 20–40% of men after prostate BT. Various definitions have been reported, and Critz et al. originally reported the phenomenon based on 779 men treated with simultaneous radiotherapy with a 125I-prostate implant followed by EBRT. They defined PSA bounce as an increase of ≥0.1 ng/mL above the preceding PSA level after simultaneous radiation followed by a subsequent decrease below that level. PSA bounce occurred in 35% of the men, and the median time to PSA bounce was 18 months from the time of implant, with 92% of bounce occurring within 36 months. Stock et al. reported that PSA bounce is more common in patients with higher implant doses and in younger patients. The patients receiving an implant dose of ≤160 Gy had a PSA bounce rate at 5 years of 24% vs 38% for those receiving an implant dose of >160 Gy, and younger patients (aged ≤65 years) had a PSA bounce rate at 5 years of 38% vs. 24% for older patients. Crook et al. reported that PSA bounce occurred in 40% of men and that the only clinical or dosimetric factor predictive of PSA bounce in multivariate analysis was being younger. Bostancic et al. reported that the incidence of PSA bounce was substantially different in patients aged <65 years vs ≥65 years (38.5% vs 16.3%), and patient age and isotope were significant predictors for PSA bounce.
There is little agreement on risk factors associated with PSA bounce; however, several articles have reported that age is a predictor for PSA bounce after BT. In the present study, there was a significant correlation between age and PSA bounce in a Japanese setting, and it is a common phenomenon not only among Caucasians, but also Japanese. Crook et al. reported that the median magnitude of PSA bounce was 0.76 ng/mL and that 64% had a magnitude of <1 ng/mL, 21% had a magnitude of 1–2 ng/mL, and 15% had a magnitude >2 ng/mL, surprisingly. In the present study, the median magnitude of the PSA bounce was 0.40 ng/mL and 84.3% had a magnitude <1 ng/mL, 10.5% had a magnitude of 1–2 ng/mL, and 5.3% had a magnitude >2 ng/mL.
Compared to our current study, frequency and onset of PSA bounce were closed to our results. However, PSA bounce height in Japanese patients was lower than in Caucasian patients. The reason for this difference is not clear; however, PSA bounce height may also be related to the duration of follow-up. Therefore, this should be considered in the present study where the follow-up was of limited duration.
The underlying mechanism of PSA bounce and the reasons for the influence of age on this bounce are still unknown. The first potential hypothesis is that PSA bounce might be due to a late-developing radiation reaction such as radiation prostatitis. Cesaretti et al. reported on a late complication of 125I-BT, which they have termed urinary symptom flare. The time to develop a ‘flare’ ranged from 5.8 to 64 months (median 23.9 months). A transient late exacerbation of urinary symptoms is common and can occur in up to half of all patients by 5 years after 125I-BT. Lehers et al. also reported that after BT for prostate cancer, some patients had recurrent LUTS after an asymptomatic period; this secondary exacerbation of symptoms (‘symptom flare’) occurred at ≈2 years after implantation, and this time-frame might correspond to PSA bounce.
However, Rosser et al. evaluated the effect of patient age on the occurrence of PSA bounce after EBRT for prostate cancer. In all, 12% of the patients developed a PSA bounce, and age was not associated with the occurrence of PSA bounce (P = 0.63) after EBRT alone. Therefore, PSA bounce after BT might involve a different mechanism compared with EBRT.
The second hypothesis is that younger patients might have more prostatic epithelium compared with older patients because serum PSA levels correlate significantly with prostatic epithelial volumes [14,15]. The present data suggest that prostate volume might be correlated with PSA bounce, and therefore younger patients are more likely to have inflammation and transient rises. However, the correlation between prostate volume and PSA bounce was weak and only definition B was significant in the multivariate analysis. Further analysis of the mechanism of PSA bounce is needed.
Another hypothesis involves the influence of ejaculation on PSA bounce. Crook et al. reported that potency before implantation was significantly predictive of PSA bounce. In addition, Tchetgen et al. reported that ejaculation causes a significant increase in the serum PSA concentration in men aged 49–79 years that may persist for up to 48 h. This change appears to correlate with age and baseline PSA levels, and it is recommended that men abstain from ejaculation for 48 h before having a serum PSA determination. However, sexual function correlated with age is usual, and it is still unclear that a correlation exists between the influence of ejaculation and the timing of PSA bounce (the median time to PSA bounce was 18 months).
Finally, we should be aware of the limitations of the present study, because of retrospective data collection and the short follow-up period. In the interim, we think that PSA bounce height might be lower in Japanese patients than in Caucasian patients.
In conclusion, the present study is the first to examine the incidence, timing, and magnitude of the PSA bounce after permanent prostate BT in Japanese patients. PSA bounce is a common phenomenon after 125I-BT in the Japanese population. Physicians should be aware that PSA bounce is more common in younger patients and must be excluded before implementing salvage interventions. Further analysis of the mechanism of PSA bounce is needed.