Diagnosis of local recurrence after radical prostatectomy


V. Scattoni, Department of Urology, Università Vita-Salute, Scientific Institute H San Raffaele, Via Olgettina 60, 20145 Milan, Italy.
e-mail: scattoni.vincenzo@hsr.it


In the long-term there is biochemical evidence of recurrent prostate carcinoma in ≈ 40% of patients after radical prostatectomy (RP). Detecting the site of recurrence (local vs distant) is critical for defining the optimum treatment. Pathological and clinical variables, e.g. Gleason score, involvement of seminal vesicles or lymph nodes, margin status at surgery, and especially the timing and pattern of prostate-specific antigen (PSA) recurrence, may help to predict the site of relapse. Transrectal ultrasonography (TRUS) of the prostatic fossa in association with TRUS-guided needle biopsy is considered more sensitive than a digital rectal examination for detecting local recurrence, especially if PSA levels are low. Although it cannot detect minimal tumour mass at very low PSA levels (<1 ng/mL) TRUS biopsy is presently the most sensitive method for detecting local recurrence. Nevertheless, the conclusive role of biopsy of the vesico-urethral anastomosis remains unclear. However, 111In-capromab pendetide scintigraphy and [11C]-choline tomography (which are better than conventional imaging for detecting metastatic tumour), have low detection rates for local disease and are considered complementary to TRUS in this setting. Patients with a high PSA after RP may be managed with external beam salvage radiotherapy. An initial PSA of < 1 ng/mL, Gleason score < 8 and radiation dose of 66–70 Gy seem to be key factors in determining success. Although a positive TRUS anastomotic biopsy may predict a better outcome after radiation therapy, the need to take a biopsy in the event of PSA failure remains under investigation. The value of salvage radiation to the prostatic bed for PSA-only progression after RP remains in question.


radical prostatectomy


vesico-urethral anastomosis


prostate-specific membrane antigen


positron emission tomography




Despite the earlier detection of prostate cancer and recent refinements in the technical aspects of anatomical radical prostatectomy (RP), prostate cancer recurrence after RP remains a growing problem. Currently treatment failure is defined as an increase in the PSA level, which provides the earliest evidence of residual or recurrent disease in nearly all patients [1]. There is biochemical evidence of recurrent prostate carcinoma in the long-term in ≈ 40% of patients who undergo retropubic RP [1], with most of the relapses (95%) in the first 5 years [1]. The location of recurrent disease after RP is a critical issue as it may greatly influence the subsequent therapeutic strategy or patient management.

This review critically evaluates the methods available to detect a local recurrence after RP, to analyse factors predicting the detection of cancer recurrence, and the advantages of a positive prostatic fossa biopsy as a prognostic factor after radiotherapy.


The PSA level after RP should decrease to ‘female’ levels within 2–3 weeks, while any detectable and/or rising PSA after RP should be considered an indication of persistent local or distant malignant disease [2]. Generally, the definition of recurrence after RP has relied on a single elevated PSA level, but the reported level of PSA which indicates failure after RP varies. Various PSA thresholds have been used, including > 0.1, > 0.2, > 0.4 and > 0.5 ng/mL [2–5]. Lange et al.[4] proposed a threshold for biochemical failure of 0.4 ng/mL, as all patients with a PSA level above this had clinical evidence of recurrence within 6–49 months after RP. Amling et al.[5] showed that a PSA threshold of ≥ 0.4 ng/mL is more appropriate than lower values as the 3-year risk of an increasing PSA level for patients with a PSA value of > 0.2 ng/mL was only 49%. Occasionally there may be very low and not rising PSA levels that persist for prolonged periods, possibly caused by a retained benign gland. However, residual benign glands are unlikely to result in PSA increases to > 0.2 ng/mL [3]. Recently, Freedland et al.[3] showed that a PSA of > 0.2 ng/mL is an appropriate threshold to define PSA recurrence as such patients had a 100% (95% CI 87–100%) 3-year risk of PSA progression.

Ultrasensitive PSA assays that detect serum PSA levels of < 0.01 ng/mL may reveal a relapse of prostate cancer even 1 or 2 years earlier than the conventional assay, but the clinical utility is limited by higher rates of false-positive results [6].


Pathological variables, e.g. Gleason score, involvement of seminal vesicles or lymph nodes and margin status at surgery can be used to predict eventual local or distant failure (Table 1). Extracapsular extension, positive surgical margins and a Gleason score of ≤ 7 seem to correlate with local recurrence, while seminal vesicle invasion, positive lymph nodes and Gleason score > 7 seem to correlate with systemic relapse [7–9].

Table 1.  Clinical and pathological variables used to predict eventual local recurrence or distant metastases in patients with a PSA failure after RP
Local recurrenceDistant metastases
Gleason score <7Gleason score ≥7
No seminal vesicle invasionSeminal vesicle invasion
Negative pelvic lymph nodesPositive pelvic lymph nodes
PSA detectable >1 year after RPPSA detectable <1 year after RP
PSA velocity <0.75 ng/mL/yearPSA velocity >0.75 ng/mL/year
Doubling time >6 monthsDoubling time ≥6 months

Even though the pathological characteristics of the surgical specimen, together with the preoperative PSA level, may greatly influence the risk of biochemical recurrence, the timing and pattern of PSA recurrence are the most important variables when trying to distinguish between a local and distant recurrence [7–9]. Many studies have reported that a short interval from surgery to detectable PSA indicates occult metastases, while with an interval of >1–2 years after RP the relapse is more likely to be local [7–9]. Partin et al.[10] showed that local cancer relapse was more probable when the PSA velocity was <0.75 ng/mL/year. Patel et al.[11] studied the PSA kinetics and showed that a PSA doubling time (or equivalent log slope PSA) of >12 months correlates with a local relapse. On the contrary, Saleem et al.[12] showed no statistically significant differences between the PSA velocity of patients with a positive or negative anastomotic biopsy. As a general rule patients with a high-grade tumour, seminal vesical involvement, a PSA doubling time of <6 months and PSA detectable at <1 year after RP have a higher risk of having systemic disease, while patients with a low to moderate grade tumour with no seminal vesical invasion, positive surgical margins, a PSA doubling time of > 6 months and PSA detectable 1 year after RP have a high likelihood of having a local recurrence. However, these variables, either alone or combined, have not been reliable in predicting the growth rate and pattern of recurrence. These variables may thus be used to tailor imaging and biopsies that may further help in establishing the site of recurrence [7].


The DRE has been used with variable success in assessing local recurrence after RP [8]. Pound et al.[13] showed that none of the 1916 men followed for a mean of > 5 years after RP had local recurrence or distant metastasis with an undetectable serum PSA level. These authors concluded that a DRE or imaging methods are unnecessary in patients with an undetectable PSA after RP. Öbek et al.[14] confirmed that an abnormal DRE after RP is always associated with a detectable PSA. Thus, a DRE is considered a routine examination only in patients with a rising PSA level. Even if an abnormal DRE after surgery strongly suggests local recurrence, the inaccuracy of DRE after RP has been established. Moreover, a DRE cannot be used to differentiate benign scars from early recurrent cancer, and local recurrence at biopsy of the vesico-urethral anastomosis (VUA) with a normal DRE has been reported [8,13,14].

The predictive value and the specificity of the DRE was studied by Saleem et al.[12], who showed that 71% of patients with an abnormal DRE had a positive biopsy. By combining DRE with PSA values the specificity may be greatly increased; while 28% of patients with a normal DRE had a positive TRUS-guided biopsy, only 20% of patients with a negative DRE and a PSA of < 1.0 ng/mL, and none with a PSA of < 0.5 ng/mL, had a positive biopsy [12]. Conversely, Leventis et al.[15] showed that the DRE was less sensitive but more specific than TRUS. They noted that in cases of a peri-anastomotic recurrence diagnosed with TRUS, about half also had a palpable nodule; however, when the recurrence site was in the area of the bladder neck, < 20% of the lesions were found on a DRE. Thus, in cases of cranial recurrence, the DRE is not so sensitive because it is difficult to assess lesions that are in an anterior position or that extend along the bladder wall. However, patients with a positive DRE and TRUS had a 100% positive biopsy rate, compared with 49% of those who only had a positive TRUS. These data confirm the high specificity of the DRE, which would appear to be less sensitive than TRUS [15].


TRUS of the prostatic fossa may be indicated to complement the DRE in patients who have a biochemical relapse after RP. TRUS of the VUA is generally considered to be an accurate diagnostic procedure in detecting prostate cancer recurrence, because it provides a precise evaluation of the normal (Fig. 1) and pathological prostatic fossa anatomy [16–20]. Particularly in cases of low PSA levels or cranial disease close to the bladder neck, TRUS, despite being less specific, is considered more sensitive than a DRE [15–22]. However, the definitive role of TRUS-biopsy of the VUA before treatment, in the event of a PSA failure, is still under investigation and some authors have shown that routine TRUS anastomotic biopsy before treatment is not advisable [8,9,16]. A negative biopsy does not preclude local recurrence and a positive biopsy does not exclude systemic disease [16].

Figure 1.

Longitudinal TRUS image showing a normally appearing prostatic fossa.


In general, recurrence of cancer is detected in 40–50% of men, with nearly a third requiring two or more TRUS-guided biopsy sessions to obtain a final diagnosis [16]. One major concern is the relative inability of TRUS and a TRUS-biopsy to detect local recurrence with a PSA of < 1 ng/mL [23], and the need to wait for a subsequent PSA increase to obtain a biopsy-confirmed local recurrence before proceeding with radiotherapy [24,25].

Several studies have described a statistically significant correlation between TRUS-suspected areas and positive biopsies [15,16,23]. Generally speaking, TRUS findings are considered pathological if any lesion, usually hypoechoic, is identified next to the peri-anastomotic site, bladder neck or retro-trigone [16,23] (Fig. 2). Other criteria suggesting local recurrence on TRUS include asymmetric thickening or fullness of the VUA or loss of the integrity of the retro-anastomotic plane [16,23]. Several authors reported that the incidence of abnormal TRUS findings may be 49–95%[15–23].

Figure 2.

TRUS images of the prostatic fossa showing a hypoechoic lesion adjacent and posterior to the VUA. Callipers mark the two diameters of the lesion. The lesion was not palpable on a DRE.

Leventis et al.[15] reported that the most common TRUS-detectable lesion site was the area of the VUA (with a 61% positive biopsy rate). The other sites were the anterior or posterior neck (positive biopsy rate 54%) and, less frequently, the retrovesical space, posterior and cephalic to the bladder neck. Finally, 14% of their patients had more than one site showing a suspected lesion, with a 71% positive biopsy rate. In cases of normal TRUS findings of the prostatic fossa, only 20% of patients had a positive biopsy, while 62% of those with a suspected lesion had proven local recurrence [11]. Saleem et al.[12] also reported a statistically significantly higher risk of cancer detection in cases of positive TRUS. Our results of 119 patients biopsied at the VUA confirmed the high sensitivity of TRUS and TRUS-guided biopsy of a detectable lesion. Indeed, while the overall positive biopsy rate was 50%, in cases of TRUS-suspected lesions [26] it increased to 67%. On the contrary, patients with high PSA levels (> 2 ng/mL) and negative TRUS do not appear to benefit from biopsy of the VUA, as they do not have a local recurrence but rather a distant metastasis. We therefore concluded that the sensitivity of TRUS of the VUA is higher than TRUS of the prostate as, in the absence of the prostate, any irregular hypoechoic area of the anastomosis is more likely to be a prostate cancer.

While the characteristics of the appearance of normal and pathological VUA on TRUS have been accurately described [17,23], no optimal VUA biopsy strategy has been defined, either for sampling location or number of cores. Shekarriz et al.[27] proposed that biopsies be taken in the axial projection from all four vesico-urethral quadrants, regardless of TRUS findings, with the biopsy directed toward the most suspicious-looking region within each quadrant. They showed that this technique optimises cranio-caudal localization and eliminates the possibility of inadvertent sampling from the contralateral side. Saleem et al.[12] took six systematic biopsies on longitudinal scans in patients with a negative TRUS and DRE. They took one core from each side of the bladder neck, one from each side of the anastomosis and two cores from the retrovesical region. Patients with a positive TRUS or abnormal DRE had systematic biopsies plus TRUS-guided biopsies of the abnormal findings.

Leventis et al.[15] reported a technique of longitudinal random guided biopsy of the prostatic fossa, including two cores from each side of the anastomosis, one toward the bladder neck and one toward the external urethral sphincter, plus additional lesion-guided biopsies. Our data suggested that the biopsy scheme with six systematic anastomotic biopsies (four needle biopsies at VUA level and two of the prostatic fossa) plus additional biopsies directed to TRUS-detectable lesions, is capable of diagnosing local recurrence in half of cases at the initial biopsy [26]. All these studies obtained similar results, ranging from a 30% positive rate on initial biopsy to an overall detection rate of ≈ 60% on repeated biopsy [15–23,26,27] (Table 2).

Table 2.  Results of VUA biopsy based on TRUS findings, according to different studies
RefNo. of patientsTRUSNN (%) of patients with +ve biopsyMean (sd or range) PSA with +ve biopsy, ng/mL
  • *

    Positive predictive value;

  • †negative predictive value.

[12] 91Positive6433 (52) 
Negative6717 (25)7.8 (13)
[27] 45Positive3422 (65) 
Negative 11 2 (18)5.2 (5.4)
[15] 99Positive5031 (62) 
Negative4910 (20)2.4
[23] 114PositiveNR66.9*5.7 (0.2–35)
[26] 119Positive6443 (67)1.6 (3.1)
Negative5519 (34) 


In the absence of evident distant metastases, a rising PSA after RP caused by a locally confined tumour is thought to indicate growing residual or recurrent cancer at the prostatic fossa. Many studies show that the higher the serum PSA level the higher the positive biopsy rate [15,16,23]. Shekarritz et al.[27] showed that the PSA levels of patients with a positive biopsy were significantly higher than in those with a negative biopsy. In particular, using a PSA threshold of 1.0 ng/mL there was a strong correlation with biopsy findings; while only 25% of patients with a PSA of ≤ 1.0 ng/mL have a positive biopsy, 71% of those with a PSA of > 1.0 ng/mL have a biopsy-confirmed local recurrence [27]. Leventis et al.[15] also reported a statistically significant increase in the number of positive biopsy findings associated with increasing serum PSA levels. Our study [26] did not show a higher positive biopsy rate when the PSA level was > 1.0 ng/mL in patients having 6–8 biopsies taken from the prostatic fossa and VUA. We also reported an overall higher detection rate (50%) at first biopsy, with a lower mean PSA (1.8 ± 3.7 ng/mL) than those reported by others, most of whom reported an overall detection rate of > 50% only in repeated biopsies [21–23]. Moreover, while several studies reported a very low positive biopsy rate at PSA levels of ≤ 0.5 ng/mL [28] we found a positive biopsy rate of > 40%[26]. Our higher detection rate at the initial biopsy stage could be either because more biopsies were taken or it may reflect the underlying differences within the population studied. Saleem et al.[12] reported that all patients with a negative DRE and a PSA of < 0.5 ng/mL had a negative biopsy; as a result, the authors decided to use TRUS and a biopsy only if the PSA was > 0.5 ng/mL. On the contrary, we consider that the sensitivity of TRUS extends even to those patients with very low serum PSA levels, as > 60% of those with positive TRUS and a PSA of ≤ 0.5 ng/mL had a biopsy-confirmed local recurrence. We suggest taking a biopsy as soon as the PSA level is > 0.2 ng/mL because the advantage of early diagnosis has been confirmed and many studies report an increase in PSA-free survival rate in patients undergoing radiation therapy with a PSA recurrence of ≤ 1.0 ng/mL [29,30].

To improve PSA sensitivity Shekarritz et al.[27] also evaluated the correlation between the biopsy results and pathological findings; patients with extraprostatic disease are at higher risk of developing local cancer recurrence, as > 70% of patients with extraprostatic cancer extension had biopsy-confirmed relapse. The authors suggested criteria for biopsy of the VUA, by combining PSA and the pathological stage of disease. They divided patients into three categories; low risk if the PSA is ≤ 1 ng/mL and the disease organ-confined; intermediate risk if the disease is organ-confined and PSA levels of > 1 ng/mL, or with extraprostatic extension and PSA levels of < 1 ng/mL; and high risk if the disease is extraprostatic and PSA levels >1 ng/mL.

In conclusion, detecting a cancer recurrence seems to be poorly correlated with the increased number of biopsy cores, but rather with TRUS and DRE findings, PSA level or pathological stage at RP.


The FDA-approved 111In-capromab pendetide scan (ProstaScint®) uses an IgG murine monoclonal antibody which directly binds to the glycoprotein prostate-specific membrane antigen (PSMA) on prostatic epithelial cells, but not with secretory glycoproteins such as PSA or prostatic acid phosphatase [31]. PSMA is a transmembrane glycoprotein which is expressed more abundantly in malignant than in normal or hypertrophic prostate tissue. Immunoscintigraphy depends on the degree of PSMA expression rather than the actual size of a metastatic lesion or degree of PSA expression. The clinical use of the ProstaScint scan has focused on detecting extraprostatic sites of metastasis and/or selecting men for salvage radiotherapy after RP [31–35]. Initial recommendations were that immunoscintigraphy may be best used for patients with a moderate-to-high probability of extraprostatic disease, as determined by an elevated serum PSA level, high Gleason score or advanced clinical stage, to detect metastatic disease. The general tendency to identify recurrent prostate carcinoma at an earlier stage after RP has recently lead to the use of ProstaScint in differentiating local from distant recurrence after RP, to better select patients for possible salvage radiotherapy to the prostatic bed and avoid unnecessary radiation in patients with metastatic disease. The relation between prostatic fossa biopsies and scintigraphy results was best described by Raj et al.[31] in a recent study of 255 patients with early biochemical evidence of failure (PSA ≤ 4.0 ng/mL) (Table 3) [31–35]. Assuming that the clinical or biochemical response to therapy could be equated with the site of disease recurrence, the results for regional uptake of the antibody were: sensitivity 76%; specificity 54%; positive predictive value 90%. The authors showed that after stratifying the results by different PSA thresholds there was no significant change in the proportion of patients with positive scans. The authors concluded that capromab pendetide scintigraphy is more sensitive than either CT or bone scintigraphy for identifying recurrent disease and can locate early PSA recurrences, even if the lack of histological confirmation of the signal limited the validity of the study.

Table 3.  Detection of disease recurrence using immunoscintigraphy
RefPSA threshold ng/mLNo. of patientsPositive scans, n (%)Local uptake, n (%)
[33]3.6  8  7 4
[34]1.0 38 30 (79)14 (37)
[35]4.0 34 23
[31]4.0255185 (72)78 (31)
[32]183108 (60)62 (34)

Moreover, scans could be falsely positive as a result of inflammation and vascular perturbations from surgery, which could contribute to the low 50% positive predictive value, while the low negative predictive value of 70% suggests that TRUS with guided biopsies may be more useful for detecting local disease [32–35]. With no reference standard and no pathological confirmation of the signal in most studies, the role of 111In-capromab pendetide scintigraphy in detecting local recurrence is incompletely defined and awaits further confirmation.


PET imaging is based on the combination of positron-emitting isotopes with pharmaceuticals that accumulate in normal and abnormal cells. PET with 18F-fluorodeoxyglucose (FDG-PET) is showing increasing usefulness in detecting, staging and monitoring the therapeutic response of a wide variety of neoplasms [36]. However, in most primary prostate tumours there is low FDG uptake, probably because prostate adenocarcinoma is a slow-growing neoplasm. In addition, the physiological urinary excretion of FDG can interfere with imaging in the pelvis.

Recently, MR spectroscopy has shown an elevated level of phosphatidylcholine in neoplastic tissue; all cells use choline as a precursor for the biosynthesis of phospholipids, which are the essential components of cell membrane. These observations have been the rationale for introducing a new positron-emission radiotracer for oncological PET studies, [methyl-11C]-choline (11C-choline). Previous PET studies have shown that 11C-choline successfully visualizes various tumours with a high signal-to-background ratio, particularly those in the pelvis, where the background radioactivity, related to the tracer, is very low. Although 11C-choline accumulation is present in normal prostate its excretion via the urinary tract is negligible, so that its accumulation in the bladder is very low [37].

In a recent study in 100 patients, 11C-choline-PET was useful for re-staging patients after RP who had increasing serum PSA levels. It was better than FDG-PET and complementary to conventional imaging techniques, but with the advantage of staging the disease in one step. 11C-choline-PET could detect local recurrences or distant metastases in about half the patients, and in all eight cases in which a negative 11C-choline-PET was not confirmed by conventional imaging but diagnosed with local recurrence on TRUS and biopsy [38].

In addition, a new imaging technique combining PET and CT (PET/CT) has been introduced into clinical use. By combining morphological and functional information, 11C-choline-PET/CT could be the first step in re-staging patients with evidence of rising PSA levels after RP. Those patients with 11C-choline-PET/CT positivity can start the most appropriate therapy, and if negative, pelvic morphological imaging only should complete the diagnostic phase to exclude local recurrences (Fig. 3).

Figure 3.

Images from a 51-year-old man previously treated with RP (1999) with increasing PSA (2.9 ng/mL at the time of PET) assessed using an integrated system of PET/CT with 11C-choline for clinical re-staging. The CT transaxial image (a), PET transaxial image (b) and PET/CT fused image (c) clearly show a local recurrence (arrows).


Salvage radiotherapy to the prostate bed has been used to treat patients with a rising PSA after RP who have confirmed or presumed local recurrence. In general, salvage radiotherapy seems to decrease PSA to undetectable levels in about half the patients (20–83%) after a follow-up of 2–5 years [29,30,39–57] (Table 4). However, the utility of salvage radiation therapy is controversial [45,46], with optimistic, e.g. [42,30], and pessimistic viewpoints [44,45]. The 50% response rate after salvage radiotherapy confirms that a high proportion of patients with a rising PSA level may have systemic disease. Thus it remains necessary to identify the prognostic factors that will aid in selecting patients who will benefit from this therapy.

Table 4.  Prognostic factors and results of salvage radiotherapy for rising PSA levels after RP
RefNo. of patientsRadiotherapy, dose, GyFollow-up, monthsPSA-free survival rate, % at x yearsPrognostic factors
  • *

    before radiotherapy; SV, seminal vesicle; RT, radiotherapy.

[24] 4664 3750 at 3PSA* < 1.1 ng/mL; dose > 64 Gy
[29] 3564 6069 at 5PSA*, pathological stage, interval recurrence to RT
[41] 4160 5524 at 5PSA* < 2.7 ng/mL
[44] 826410832 at 2Interval to PSA recurrence
10 at 5 
[42] 4766 3664 at 3PSA* < 2 ng/mL; SV or node involvement
[30] 7866 2568/78 at 2PSA* < 2 ng/mL
[54] 3470 3847 at 3Gleason score, PSA* < 4.0 ng/mL
[55] 6868 3555 at 3Preoperative PSA < 20 ng/mL
[45] 3858.7 3530 at 5Time to postoperative failure > 1 year
[40] 2361 4616 at 5PSA ≤ 2.0 ng/mL
[56] 8266–70 4864 at 5PSA* < 2 ng/mL and pathological stage
[57]16664 5246 at 5Gleason score, SV involvement
[53] 9566 2943 at 3PSA*
24 at 5 
[43] 5366 3346 at 3Gleason grade, PSA*
[50] 7364.8 5445 at 10Gleason score and SV invasion
[51] 5464 4535 at 5SV invasion, Gleason score > 6, PSA* < 1.2 ng/mL
[39] 11566.6 4246 at 4+ve margins, extracapsular extension, no SV involvement
[52] 6166.6 3639 at 4PSA* < 1 ng/mL, Gleason score
[48] 7170.2 5440 at 5PSA* < 1 ng/mL, Gleason ≥ 7, +ve biopsy, RT dose ≥ 70 Gy
[47] 5165.7 4556 at 3PSA*, SV invasion, lymphovascular invasion
16 at 5 

Some investigators report better results when using external beam irradiation in patients with local prostate cancer recurrence documented on biopsy, and suggest using radiation therapy only if the biopsy is positive [29,54]. Rogers et al.[54] reported a 48% 3-year second PSA relapse-free rate after radiotherapy in patients with biopsy-confirmed local recurrence. Conversely, other studies have shown no survival benefit in patients after radiotherapy with documented local recurrence, as opposed to patients affected by PSA relapse alone [24,25]. For instance, Koppie et al.[25] evaluated two different groups of patients, one with PSA failure only and the other with biopsy-confirmed cancer recurrence. There were no differences in results between these groups and a positive histology gave no chance of improving the outcome. The delay caused a longer time to radiation therapy even though a worse disease-free outcome in patients undergoing prostatic fossa biopsy has not been reported. The authors concluded that a longer time to salvage treatment might affect patients if PSA increases progressively until there is a biopsy-confirmed cancer recurrence.

However, the PSA level before radiotherapy appears to be the most important factor for predicting the outcome of salvage treatment, regardless of a positive biopsy. Patients with a PSA relapse of > 2.0 ng/mL are at greater risk of developing a recurrence after external beam radiation and tend to have a worse outcome than patients with a PSA of <2.0 ng/mL [43]. Koppie et al.[25] reported a significant improvement in progression-free survival in patients with a PSA of ≤ 1.0 ng/mL, and Van der Kooy et al.[29] reported an 8-year relapse-free survival of 67%, 39% and 42% in patients with PSA levels before radiation therapy of ≤ 1.0, 1.1–4 and > 4 ng/mL, respectively. Vanuytsel et al.[43] analysed the prognostic value of PSA both as a categorical and a continuous variable, and suggested that no specific optimal PSA threshold can be defined, while PSA is strongly correlated with better disease-free survival when considered as a continuous variable. Recently, Song et al.[52] reported that patients with a higher PSA level (>1 ng/mL) and Gleason score of ≥ 8 are less likely to benefit from salvage radiotherapy because of the likelihood of microscopic systemic involvement. In another recent analysis of 71 patients having salvage radiotherapy, with a median (range) radiation therapy dose of 70.2 (58.4–73.8) Gy, we reported an actuarial 5-year biochemical disease-free survival rate that was significantly higher in patients with biopsy-confirmed recurrence (64.3% vs 28.1%, P = 0.042, log-rank) and in patients with Gleason score ≤ 7 (61.5% vs 26.8%, P = 0.028) [48]. Moreover, multivariate analysis (Cox model) showed a positive prognostic significance for a PSA of ≤ 1.0 ng/mL (P = 0.001, hazard ratio 1.8), biopsy-confirmed recurrence (P = 0.03, ratio 2.4), Gleason score ≤ 7 (P = 0.04, ratio 4.2) and radiotherapy dose ≥ 70.2 Gy (P = 0.006, ratio 0.8) [48]. We concluded that salvage radiotherapy with a dose of ≥ 70.2 Gy for PSA recurrence of prostate cancer seems to be more effective in patients with biopsy-confirmed cancer recurrence, PSA ≤ 1.0 ng/mL before radiotherapy and a Gleason score of ≤ 7, irrespective of pathological stage at surgery and time to PSA failure. In our experience of the evidence for biochemical recurrence, biopsy of the prostatic fossa remains a mandatory procedure because it is prognostic [48].

According to these indications, patients with a low PSA level and a positive biopsy seem to be more likely to have a better outcome with radiotherapy. Unfortunately, detecting recurrence is more difficult in those with a PSA of ≤ 1.0 ng/mL, because the presence of a very small local recurrence may be undetected by anastomotic sampling [16]. Indeed, 10–40% of patients with a negative biopsy and a PSA of < 1.0 ng/mL show a PSA decrease after radiotherapy, suggesting the presence of undetected local recurrences [16]. Conversely, 30–50% of patients with a PSA level of > 2.0 ng/mL have no biopsy-confirmed recurrence, and some authors consider radiotherapy unnecessary, if not damaging, with no positive histology [49]. Moreover, salvage radiotherapy, as with any other treatment, is not without morbidity; most patients have acute gastrointestinal and/or genitourinary side-effects, while late toxicity can occur in ≈ 20% of patients [47,50]. Urinary incontinence may deteriorate in 5–10% of patients, but < 8% of patients with good continence develop any type of incontinence after salvage radiotherapy [7]. Considering that only half of the patients have a PSA response, some with no clinical benefit may also have side-effects that may influence their quality of life.

The American Society of Therapeutic Radiology and Oncology recently published a consensus panel report recommending that patients receive salvage radiotherapy before the PSA is > 1.5 ng/mL, and that at least 64 Gy be delivered to the prostatic bed. Higher PSA levels probably indicate an increased risk of micrometastases, limiting the efficacy of local treatment [58]. The use of hormonal therapy with radiotherapy is still under investigation and at present there is no standard role for concomitant hormonal therapy in patients who undergo salvage radiotherapy for a rising PSA after RP.

In conclusion, histological confirmation of a cancer recurrence does not appear to be necessary for radiotherapy, but might predict a better biochemical disease-free survival after salvage treatment. A PSA of < 1 ng/mL before radiotherapy, a Gleason score of < 7 and radiotherapy dose (> 64 Gy) appear to be the most important factors in predicting the outcome of salvage treatment [58].


For biochemical relapse after RP detecting the site of the recurrence (local vs distant) is critical for defining the optimum treatment. Pathological and clinical variables, e.g. Gleason score, involvement of seminal vesicles or lymph nodes, margin status at surgery or PSA kinetics, may help to predict the growth rate and pattern of recurrence. TRUS biopsy of the prostatic fossa seems to be an accurate diagnostic method to detect prostate cancer recurrence. Especially in those with a low PSA level, TRUS, despite being less specific, is more sensitive than DRE or other new imaging methods, e.g. immunoscintigraphy or 11C-choline-PET. Unfortunately, the inability of TRUS to detect minimal tumour mass when the PSA is very low (< 1 ng/mL) has been confirmed and the conclusive role of biopsy of the VUA remains unclear.

Salvage radiotherapy to the prostate bed is currently used to treat patients with a rising PSA level after RP who have confirmed or presumed local recurrence. Published data show that the lower the PSA level, the longer the PSA-free survival after salvage radiation therapy. In this setting the need for a TRUS biopsy of the VUA before salvage radiotherapy has been questioned, but it has been suggested that a positive biopsy may predict a better outcome after radiation therapy.