Interpreting a rising prostate-specific antigen after brachytherapy for prostate cancer


  • Miren Gaztañaga,

    1. Department of Radiation Oncology, British Columbia Cancer Agency, Center for the Southern Interior, Kelowna, British Columbia, Canada
    Search for more papers by this author
  • Juanita Crook

    Corresponding author
    1. Department of Radiation Oncology, British Columbia Cancer Agency, Center for the Southern Interior, Kelowna, British Columbia, Canada
      Juanita Crook M.D., F.R.C.P.C., British Columbia Cancer Agency, Center for the Southern Interior, 399 Royal Avenue, Kelowna, BC V1Y5L3, Canada. Email:
    Search for more papers by this author

Juanita Crook M.D., F.R.C.P.C., British Columbia Cancer Agency, Center for the Southern Interior, 399 Royal Avenue, Kelowna, BC V1Y5L3, Canada. Email:


The aim of the present study was to review the English language literature on the topic of prostate-specific antigen bounce after brachytherapy and present a summary of the current knowledge. Although ultimately prostate-specific antigen is a reliable measure of success after prostate brachytherapy, it can be very misleading in the first 3 years because of the frequency with which temporary benign rises in prostate-specific antigen occur. We have reviewed the English language literature on the topic of prostate-specific antigen bounce under the following headings: prostate neoplasms, brachytherapy, biochemical definition of prostate-specific antigen failure, “benign prostate-specific antigen bounce” and “prostate-specific antigen spike”. We included brachytherapy delivered as either low dose rate or high dose rate, and either as monotherapy or as a boost combined with external beam radiotherapy. A benign self-limited rise in prostate-specific antigen after prostate brachytherapy is seen in an average of 35% of patients, but increases in frequency with younger age. In patients aged less than 55 years, it is observed in up to 68%. Other factors, such as sexual activity, dose, prostate volume and the use of high dose rate versus low dose rate have been implicated in affecting the frequency of the benign bounce. Benign increases in prostate-specific antigen are frequent after prostate brachytherapy. It is important to recognize and correctly diagnose this phenomenon in order to avoid unnecessary salvage treatment.

Abbreviations & Acronyms

androgen deprivation therapy

bNED =

biochemical “no evidence of disease”


cause-specific survival


external beam radiotherapy


high dose rate


low dose rate

NS =

not significant

OS =

overall survival


prostate-specific antigen


transition zone index (transition zone volume/total prostate volume)

V150 =

percentage of prostate volume receiving a minimum of 150% of the prescribed dose


Once a prostate cancer patient has been diagnosed and treated, management of the follow-up period can be challenging. PSA kinetics can be very variable after non-surgical treatment. The value of the nadir, the time to reach the nadir and fluctuations in the PSA en route can all be open to different interpretations. The present review addresses these issues and provides an update on the current evidence.


We reviewed the English language literature on the topic of PSA bounce under the following headings: prostate neoplasms, brachytherapy, biochemical definition of PSA failure, “benign PSA bounce” and “PSA spike”.

PSA nadir and the definition of biochemical failure

PSA nadir is defined as the lowest PSA in the follow-up period after radiotherapy. As the PSA might continue to decline for several years after treatment, three consecutive stable readings1 are required to determine the nadir. Because of assay variability and the very low levels that are achieved, especially after brachytherapy, a reading ≤0.1 ng/mL can be considered as the nadir.

Since PSA came into clinical use in the early 1990s, the radiation community has struggled with a PSA definition of “failure”. In 1997, the American Society of Therapeutic Radiation and Oncology consensus conference defined biochemical failure as three consecutive increases in PSA after the nadir with readings every 3–6 months.2 Lack of specificity1 and problems arising from the need to backdate the failure to midway between the nadir and the first increase led to a second consensus conference and adoption of the currently accepted “Phoenix” definition. Biochemical failure is specified as a PSA >2 ng/mL over the nadir.3 Both these definitions were based on PSA kinetics after EBRT and are not specific to brachytherapy.

The current definition (nadir + 2) does not specify the frequency of PSA testing required for outcome assessment, although the National Cancer Comprehensive Network recommends at least one PSA per year.4 Ciezki et al. reviewed the records of 5616 patients treated with a curative intent with radical prostatectomy, EBRT or brachytherapy. The frequency of PSA testing was an independent predictor of outcome and the optimal testing frequency was recommended to be twice yearly.5

Brachytherapy is unique amongst radiation modalities in that it is an ablative treatment. Unlike EBRT, brachytherapy frequently eliminates all PSA-producing tissue, thus driving the PSA to an undetectable level. Unfortunately, because of the frequency of the benign PSA bounce, PSA is an unreliable indicator of success in the first 3 years. Most of the “bounce” literature comes from experience with permanent seed LDR prostate brachytherapy, although bounces are also reported after high dose rate brachytherapy. The mechanism of the PSA bounce remains unproven, although several theories have been proposed. In the case of LDR brachytherapy, a foreign body response as a result of the presence of the seeds might cause a temporary increase in PSA. This can be seen as diffuse inflammation on magnetic resonance spectroscopy in patients experiencing a PSA bounce.6 For HDR patients, it is postulated that the very high biologically effective dose (typically 10 Gy in 8–10 min) might induce a greater radiation-induced inflammatory response.7

The association between lower PSA nadir and improved freedom from biochemical failure is well established.8,9 However, any arbitrary PSA threshold to define success or failure is discouraged. Despite the lack of a clear demarcation in the range between 0.1 and 1.0 ng/mL, Grimm et al. found that the lower the PSA nadir, the higher the probability of success.9 A PSA nadir <0.5 ng/mL is associated with significantly better long-term freedom from biochemical failure (95.2% vs 71.5%).8 A PSA level at 5 years after brachytherapy is highly predictive of outcome, with 97.4% 10-year biochemical control in a cohort of 921 men with PSA ≤0.2 ng/mL at 5 years post-brachytherapy. Hayden et al. found only one late biochemical relapse amongst 762 patients with a PSA ≤0.2 ng/mL at 48–60 months postimplant.10

Bounce definition

After brachytherapy, approximately 35% of patients will experience a benign temporary increase in their PSA levels that resolves spontaneously without therapeutic intervention. This is called a benign “bounce” or “spike”. The first study published on PSA bounce was reported in 2000 by Critz et al.11 Although physiological fluctuation in PSA can follow instrumentation, infection, ejaculation12 or even activities such as bicycle riding, these are quite distinct from the postbrachytherapy benign bounce.

The most common definition of a PSA bounce is an increase ≥0.2 ng/mL, followed by a spontaneous decrease to the prebounce level or lower,1,13,14 but other definitions have been reported. The smaller the increment in PSA used to define the bounce, the greater the frequency. Smaller bounces have fewer consequences in terms of patient/physician anxiety and avoidance of inappropriate intervention. However, some bounces can be very disturbing and 15% will have a magnitude >2 ng/mL,1 and some even exceed the baseline pretreatment PSA. Double bounces can also occur, once again increasing anxiety in patients and clinicians. Median bounce height varies depending on the bounce definition, but it generally ranges from 0.315 to 0.76 ng/mL.1 Some authors report different median PSA values depending on whether the rise is a bounce (0.6 ng/mL) or a recurrence (1.2 ng/mL),16 but this distinction is artificial and misleading. Many bounces exceed this threshold and a rising PSA from a recurrence will continue to rise until therapeutic intervention, such as androgen ablation, is undertaken. A bounce will not do this, as by definition it will spontaneously decline to the prebounce level or lower. As a result of the frequency with which bounces occur, patients should be aware of this possibility before undergoing brachytherapy, so as to minimize subsequent anxiety. An example of a benign bounce is shown in Figure 1 for a young man, aged 46 years at the time of brachytherapy, who had a stage T2a, Gleason 6 adenocarcinoma.

Figure 1.

PSA trend for a 46-year-old man with a T2a Gleason 6 adenocarcinoma of the prostate treated with permanent seed iodine-125 prostate brachytherapy. The PSA bounce peaked at 24 months after the implant at a level of 7.5 ng/mL, but decreased spontaneously without intervention and remains undetectable at 6.5 years.

Bounce incidence and timing

After LDR brachytherapy, the frequency of benign bounces has varied from 9.5% when defined as an increase of 0.8 ng/mL, to 84% when defined as any increase in the PSA. The average PSA bounce rate using a definition of an increase of ≥0.2 ng/mL is 35% (Table 1). The usual time of onset of a bounce is between 12–24 months, but can be as early as 6 months after brachytherapy and occasionally as late as 4 years. The later the occurrence, especially beyond 3 years, the more likely it is to represent a true failure. The mean duration of a bounce is usually less than 1 year,1 and the majority will have resolved within 3 years after treatment.11,26

Table 1. Summary of the literature on PSA bounce after brachytherapy
ReferenceYear n Definition%Onset (months)TreatmentMedian follow up (months)Association
Cavanagh et al.132000591≥0.23624.8Brachy +/− EBRTNot availableNS
Das et al.162002186≥15%6226.4Brachy +/− EBRT27NS
Merrick et al.1720022180.22419.5Brachy +/− EBRT46.2Young age, stage, TZI V150 <55%, 3-mo PSA
Critz et al.18200310110.14118Brachy + EBRT72Young age
Stock et al.1920033730.13119.5Brachy48Young age
0.41719.5  High volume
Patel et al.2020042950.22819.4Brachy38Young age
Morita et al.1520042000.14013BrachyNot availableNS
Toledano et al.2120062950.15519Brachy +/− ADT40.3Young age, D90 >200 Gy
Ciezki et al.1420061620.24615Brachy +/− ADT73Young age
Pickles et al.222006449Any increase8413Brachy +/− ADT48Androgen ablation
Crook et al.120072750.24015.6Brachy44Young age
Mitchel et al.2320082950.23714.8Brachy45NS
Satoh et al.2420093880.15112Brachy30Young age
Thompson et al.25201010062.0 falling to ≤0.54449Brachy +/− ADT54Young age
Caloglu et al.2620118200.230.117.4Brachy +/− ADT58Young age
0.612.816.2PSA per year
Hinnen et al.2720119750.23219.2Brachy +/− ADT78Improved bNED, CSS, OS

Regarding HDR brachytherapy, there are fewer data about the onset and duration of the bounce. Bachand et al.28 reported the median time to PSA bounce was 15.2 months, and a median duration of 18.7 months.

Bounce and prognosis

In patients treated with brachytherapy, PSA bounces do not have a negative impact on prognosis.1 In some series, a PSA bounce is predictive of increased freedom from biochemical failure,14,20,27 and improved disease/specific and overall survival.27 A recent report from Hinnen et al. analyzed 975 brachytherapy patients with a median follow up of 6 years and observed bounces in 32%, with a median time to bounce of 1.6 years. There was a significant negative association between bounce and death from disease, with just one prostate cancer death (0.3%) in the group experiencing a bounce, compared with 40 (6.1%) in the no-bounce group. This observation fits with the theory of Rosser et al.29 that a PSA bounce is related to the transition of sublethal to lethal cellular damage which results in a sudden release of PSA into the bloodstream.27 The absence of bounce would correlate with the absence of this transition to lethal damage, and, therefore a higher rate of disease progression.29

Predictive factors


Young age is the factor most frequently associated with the bounce phenomenon.1,14,17–19 Various age break-points have been described, the most common ones being 6519,20 or 70 years.21 The importance of young age in the etiology of this phenomenon is illustrated by Gomez et al. in a cohort of 96 men with a median age 53 years (range 45–55) treated with brachytherapy alone where bounces (>0.2 ng/mL) were seen in 68%.30 It is important to recognize this, as younger patients are more likely to be offered “salvage” treatment for an early rising PSA.

Sexual activity

The study by Crook et al. confirmed the association with younger age and potency; however, in multivariate analysis, only age was independently associated with PSA bounce.1 The explanation for the association of younger age and bounce might relate to higher androgen production in younger men, which could translate into more reactive epithelial cells.19 More frequent sexual activity could induce more of an inflammatory response because of the presence of the seeds acting as a foreign body irritant.1 This is supported by the presence of diffuse inflammation in the prostate at the time of the bounce as shown by magnetic resonance imaging spectroscopy.6

Other factors

Other factors that have been reported, but that are not consistently observed, include dose21 and prostate volume.

Stock et al. reported that higher doses (>160 Gy) were more likely to result in a bounce (38% vs 24%), postulating that higher doses might increase the likelihood of a radiation-induced inflammatory reaction.19 Merrick et al. found the opposite association, and suggested that higher radiation doses that obliterate both malignant and benign tissue decrease the risk of a PSA bounce.17

Regarding prostate volume, Stock et al.19 reported that larger prostates were associated with a 23% increased risk of bounce, whereas Merrick et al.17 found that larger transition zone volumes were predictive of a bounce.

When neoadjuvant hormonal therapy is used before radiotherapy, testosterone recovery several months after treatment might produce a temporary increase in PSA.1,22 Pickles et al. analyzed 2030 patients treated with EBRT or brachytherapy, with or without androgen deprivation, and reported that a temporary PSA increase was more common after androgen ablation.22 Toledano et al. did not find an association between neoadjuvant hormonal therapy and PSA bounce in a brachytherapy population.21 In general, the PSA rise related to testosterone recovery begins 6–9 months after the androgen ablation is suspended, and parallels the testosterone rise, whereas a PSA bounce typically occurs 12–24 months after brachytherapy, when the testosterone level should be stable, even if androgen deprivation has been used.

Characteristics of the PSA bounce amongst different ethnic groups have not been extensively reported in the literature. However, PSA bounces have been described in Japanese patients after brachytherapy at a median time of 12–18 months and, as in North American publications, are associated with younger age, although the magnitude of the bounce might be lower.24

Bounce phenomenon in HDR versus LDR

Classic LDR brachytherapy usually delivers approximately 50–60 cGy/h, although the initial dose rate for an iodine-125 brachytherapy seed implant is only approximately 10 cGy/h. In contrast, HDR brachytherapy is delivered at a dose rate of approximately 100 Gy/h (1000 times higher than a seed implant). A dose rate difference of this magnitude clearly impacts on the mechanism of cell death, at the very least resulting in more double strand DNA breaks. Post-treatment PSA kinetics would not be expected to be similar.

Nonetheless, PSA bounces have been described after HDR brachytherapy. Bachand et al., in a series of 153 patients treated with EBRT followed by a HDR boost, reported PSA bounces in 9.8%, using a high-threshold definition of an increase of 2 ng/mL.28 McGrath reported on patients treated with LDR monotherapy, HDR monotherapy or HDR boost, and found more frequent bounces after HDR monotherapy (40% for the >0.3 ng/mL definition) than after LDR (25%) or HDR boost (29; P = 0.04).7 A possible explanation is the potentially increased biologically effective dose delivered through HDR alone compared with the other treatment modalities, in turn generating a greater radiation-induced inflammatory response.

Post-radiotherapy prostate biopsies

The fact that most PSA bounces occur 12–24 months after implantation makes investigation problematic. Prostate biopsies carried out within 2 years of radiotherapy are frequently misleading, showing “residual cancer with treatment effect”.31 Because of long tumor doubling times and because radiotherapy causes post-mitotic cell death, histological resolution of prostate cancer often takes 2–3 years. By 3 years, most benign PSA bounces will have resolved, obviating the need for a biopsy.

After EBRT, the rate of positive rebiopsies depends on EBRT dose and the timing of the biopsy.32 There are fewer data on postbrachytherapy prostate biopsies. Stone et al. described a 21.5% incidence of positive biopsies 2 years after prostate brachytherapy, but noted that many resolved with continued surveillance and repeat biopsies, with an ultimate positive biopsy rate of just 7.7%.33,34 Clearly, biopsies carried out before 3 years after brachytherapy will not help to elucidate the cause of a rising PSA and, on the contrary, often incorrectly identify “residual disease” and lead to inappropriate intervention. Reed et al. reported on the dilemma of “positive” biopsies in the face of a rising PSA in the first 2 years after brachytherapy.35 Eight men with an early rising PSA after brachytherapy were inappropriately offered salvage radical prostatectomy, but declined. With further follow up, the PSA decreased spontaneously and all remain disease free.

Our approach is to reserve repeat prostate biopsies after brachytherapy for cases where the PSA has not begun to decrease by 30–36 months.1 Before this, results might be misleading and result in unnecessary “salvage”. If the PSA were to increase to >10 ng/mL at any time in this period, then repeat systemic investigation, such as abdominal/pelvic computed tomography and bone scan, are certainly warranted. Newer techniques as choline positron emission tomography (PET) and/or endorectal magnetic resonance imaging with spectroscopy might also be useful. This is especially true when brachytherapy is being used for intermediate risk disease where there is a higher risk of systemic failure.

Biochemical recurrence

Unfortunately, there is no reliable way to distinguish a benign bounce from a biochemical failure. At the time of any rise in PSA, the implant dosimetry should be reviewed to assess dosimetric coverage and margins, especially at the known locations of the cancer as indicated by the initial digital rectal exam, biopsies and transrectal ultrasound. An excellent implant is more likely to be associated with a bounce than a recurrence, although this should not diminish the vigilance of PSA follow up, as cancers can recur in the adjacent seminal vesicles outside of the implanted area.36

The time of onset of the PSA increase after LDR brachytherapy might help to discriminate between a bounce and a biochemical recurrence, as bounces generally occur sooner than biochemical failure. Caloglu et al. reported a marked difference in the timing of the first PSA rise for a bounce (15–18 months) versus the timing for biochemical failure (34 months).26 These results are consistent with those published by Crook et al., who reported a median time to bounce of 15 months compared with 30 months for failures.1


Brachytherapy is a unique radiation modality distinct from EBRT. Whether LDR or HDR, brachytherapy is an ablative treatment that frequently eliminates all PSA-production in the prostate. It is important for those involved in the care of brachytherapy patients to appreciate the unique PSA kinetics after brachytherapy. An early rise in the PSA should be considered a bounce until proven otherwise. Postbrachytherapy biopsies should not be carried out until at least 30 months postimplant. It is safe to conclude that a rising PSA before this is either a benign bounce or a systemic failure, not a local failure. Restaging with computed tomography imaging and bone scan should be considered for any patient with a steep rise approaching 10 ng/mL, especially in cases with unfavorable pathological features.

Conflict of interest

None declared.