What's known on the subject? and What does the study add?
Previous studies have identified the most important prognostic factors of the likely outcomes of salvage prostate whole-gland ablation, including initial clinical stage, biopsy Gleason score, and PSA (total and doubling time). There is potential for further optimization of candidate selection for salvage cryoablation with curative intent and nadir PSA achieved after whole-gland cryotherapy may provide additional prognostic value.
The study shows that the most important prognostic factors of biochemical progression-free survival for patients who have undergone whole-gland salvage prostate cryotherapy are nadir PSA achieved after therapy and pre-therapy biopsy Gleason score. Based on these two prognostic variables, we have identified risk stratification groups (low, intermediate and high) which help predict the expected outcomes of salvage whole-gland prostate cryotherapy in a given patient. This risk stratification constitutes a useful clinical tool in defining which patients maybe best suited for this local salvage treatment method.
To assess the prognostic variables predicting the risk of biochemical progression-free survival (bPFS) after salvage prostate whole-gland cryotherapy using the Phoenix definition of bPFS.
Patients and Methods
A total of 132 patients underwent prostate whole-gland salvage cryotherapy with curative intent. No patient underwent neoadjuvant/adjuvant hormonal ablative therapy, and all had extended post-salvage prostate-specific antigen (PSA) follow-up data.
Cox univariate and multivariate logistic regression analyses of potential predictors of bPFS were conducted.
Kaplan–Meier analyses of bPFS was also performed.
At a mean (range) follow-up of 4.3 (0.9–12.7) years, the median (range) post-cryotherapy nadir PSA achieved was 0.17 (0–33.9) ng/mL.
On multivariate analysis, predictors of bPFS were nadir PSA post-cryotherapy and pre-salvage biopsy Gleason score (P < 0.001 and 0.009, respectively).
Risk stratification groups (low, intermediate and high) were developed based on the presence of zero, one or two adverse risk factors, the risk factors being either a nadir PSA >2.5 ng/mL or biopsy Gleason score ≥7, with the Kaplan–Meier bPFS curves of these risk groups being significantly different (P = 0.02 and <0.001, respectively).
Post-salvage nadir PSA and pre-salvage biopsy Gleason score are important predictors of outcome in this patient cohort.
Low-, intermediate- and high-risk groups can be determined based on these variables and can define patients best suited for prostate cryotherapy.
Locally recurrent prostate cancer after primary radiotherapy affects 10–70% of treated patients [1, 2]. Traditionally, these patients were either treated with salvage hormonal ablative therapy with no curative potential or observed until clinical progression was found. This approach probably resulted from the widely held belief that local salvage therapies had unacceptably high morbidity; however, this belief has changed in recent years with the emergence of local salvage therapies, particularly salvage cryoablation, using third-generation cryotherapy devices. For local prostate salvage cryotherapy biochemical disease-free survival rates of 50–70% at 5 years have been reported in appropriately selected patients . Spiess et al.  reported a pre-treatment nomogram predicting the likelihood of biochemical failure after prostate salvage cryoablation using readily available clinical data. Their model was a useful predictive tool for clinicians and patients when considering treatment options.
Nevertheless, it is clear that despite the increased use of prostate cryoablation and referral to tertiary care offering this treatment, it remains under-used, particularly in the salvage setting. The reasons for this are probably multi-factorial, but one of these could be the inability of referring physicians (urologists, radiation oncologists and allied healthcare professionals) to identify suitable candidates for this potentially curative salvage treatment. In recent years, important prognostic factors of the likely outcomes of prostate whole-gland salvage cryoablation have been identified, including initial clinical stage, biopsy Gleason score, and PSA measurements (total PSA and PSA doubling time) [5-8]; however, further optimization of the selection of suitable candidates for salvage cryoablation with curative intent is possible and it is possible that nadir PSA achieved after whole-gland cryotherapy may provide additional prognostic value. A previous study by Greene et al. . assessed the role of the nadir PSA achieved after salvage prostate cryoablation and showed that a nadir PSA of ≤0.5 ng/mL predicted the likelihood of subsequent biochemical failure.
The aim of the present study was to define the predictors of outcome in patients undergoing prostate whole-gland salvage cryotherapy in a contemporary patient cohort. In addition, we aimed to develop a risk stratification model using these prognostic variables which could be easily integrated into clinical practice.
Baseline patient characteristics including total serum PSA level, pre-salvage biopsy Gleason score, and clinical stage were collected, as were the clinical data pertaining to the treatment outcomes of prostate salvage cryotherapy, including serial total serum PSA values, nadir PSA achieved and post-cryotherapy prostatic biopsy results if obtained because this was deemed clinically indicated by the the treating physician (i.e. if there was a rising PSA level after salvage cryotherapy or clinically suspected locally recurrent or residual disease).
Figure 1 shows the study inclusion and exclusion criteria. A total of 614 patients underwent prostate whole-gland salvage cryotherapy within the COLD registry. Of these patients, 463 patients had received neoadjuvant, concomitant or adjuvant hormonal ablative therapy and were excluded. Of the remaining 151 patients, fewer than two consecutive post-salvage cryotherapy PSA measurements were available for 19 patients, who were thus excluded, which left 132 patients who constituted our study population.
The primary endpoint was to determine the prognostic factors predicting the subsequent risk of biochemical failure using the validated Phoenix definition of biochemical failure (nadir + 2 ng/mL) [10, 13].
Descriptive statistics were reported using frequency distributions and percentages for categorical variables as well as mean and range of values for continuous variables. We used Kaplan–Meier analysis for biochemical disease-free survival (bPFS) and a P value <0.05 was considered to indicate statistical significance . Cox proportional hazard univariate and multivariate models were used to assess the potential prognostic factors of bPFS . All potential prognostic factors with a P value <0.25 in the univariate model were included in the multivariate model. Two patients were not included in our Kaplan–Meier analysis based on the risk stratification groups owing to missing data on pre-salvage biopsy Gleason score in one patient and missing data on bPFS using the Phoenix definition in another.
The characteristics of the study population (N = 132) are shown in Table 1. The mean (range) age of our patients at time of diagnosis was 70.1 (46.0–87.0) years. Most patients had an initial biopsy Gleason score <6 or 7 (59 and 25%, respectively). The clinical stage at time of diagnosis was ≤ T2b in 112 patients (85%) and the mean total serum PSA at presentation was 6.2 ng/mL (4.9–34.2%). The D'Amico risk group of patients at time of diagnosis was low in 18 (14%), intermediate in 82 (62%), and high in 32 (24%). The median (range) nadir PSA achieved after cryotherapy within our patient population was 0.17 (0–33.9) ng/mL. Within the study population, a nadir PSA <0.1 ng/mL was achieved in 47 (35.6%), 0.1–0.6 ng/mL in 39 (29.8%), 0.7 –2.5 ng/mL in 27 (20.5%), and ≥2.5 ng/mL in 19 patients (14.4%). The mean (range) follow-up after salvage therapy was 4.0 (0.9–12.7) years.
Table 1. Patient demographics (N = 132).
Mean (range) age at diagnosis, years
Initial biopsy Gleason score, n (%)
Clinical stage at diagnosis, n (%)
Mean (range) PSA at presentation
D'Amico risk group, n (%)
Median (range) nadir post-cryotherapy PSA achieved
Mean (range) follow-up, years
As shown in Fig. 2, a Kaplan–Meier prediction analysis of bPFS using the Phoenix definition was conducted for our entire study population, with the 1-, 2-, and 5-year rates being 87.8, 72.4, and 45.5%; respectively. Kaplan–Meier analysis of bPFS stratified by the nadir PSA achieved after cryotherapy is shown in Fig. 3. There was a significant difference in bPFS between all nadir PSA subgroups.
Univariate Cox regression analysis of potential predictors of bPFS is shown in Table 2. Total serum PSA level at diagnosis, nadir PSA after cryotherapy, and the pre-salvage biopsy Gleason score were significant (P = 0.004, P < 0.001, and P = 0.007, respectively), but only nadir PSA after cryotherapy and pre-salvage biopsy Gleason score remained significant on multivariate analysis (P < 0.001 and P = 0.009, respectively). Kaplan–Meier analysis of bPFS stratified by pre-salvage biopsy Gleason score was subsequently conducted (Fig. 4) and this confirmed that patients with a Gleason score ≤6 had an improved bPFS compared with patients with a higher pre-salvage Gleason score (P = 0.02).
Table 2. Univariate and multivariate analysis of predictors of bDFS.
Bold indicates statistically significant variables on univariate and multivariate analyses.
Serum PSA level at diagnosis
Nadir PSA post-cryotherapy
Biopsy Gleason score (pre-salvage treatment)
Clinical stage at diagnosis
Risk stratification groups (low, intermediate and high) were developed based on the presence of zero, one or two adverse risk factors. Risk factors were defined as either a post-cryotherapy nadir PSA >2.5 ng/mL or a pre-salvage biopsy Gleason score ≥7. The Kaplan–Meier bPFS curves of the low-, intermediate- and high-risk groups were significantly different (Fig. 5).
The present study shows that the nadir PSA achieved after salvage cryotherapy and the pre-salvage biopsy Gleason score are the strongest predictors of bPFS within this present patient cohort. Although previous studies have alluded to the clinical importance of the nadir PSA achieved after prostate cryotherapy [2, 12], the present study is unique in that we were able to confirm that the nadir PSA and pre-salvage biopsy Gleason score were the only prognostic factors predicting the subsequent risk of PSA progression when assessed in a multivariate prognostic model of patients treated in a homogeneous fashion (i.e. salvage cryoablation without neoadjuvant, concomitant or adjuvant hormonal ablative therapy). Furthermore, based on these two prognostic variables, we have developed low-, intermediate-, and high-risk groups which help define which patients are best suited for prostate whole-gland salvage cryotherapy and provide patients with realistic expectations of their treatment outcomes.
In a previous study from our group, we were able to show that a nadir PSA <0.6 ng/mL was an important endpoint of bPFS in the primary prostate cryotherapy setting . In the present study, we repeated the univariate and multivariate Cox regression analysis of bPFS using a PSA <0.6 ng/mL as our study endpoint as opposed to the Phoenix definition, with the same two variables (nadir PSA and biopsy Gleason score) predictive of bPFS on multivariate analysis (P = 0.001 and 0.007; respectively). This validates not only PSA <0.6 ng/mL as an important endpoint of bPFS in the salvage cryotherapy setting, but shows the robustness of the nadir PSA and pre-salvage biopsy Gleason score as predictors of outcome in this clinical setting. From a pathophysiological standpoint, patients with a nadir PSA >2.5 ng/mL are suspected of harbouring micrometastatic disease as patients with a complete prostate whole-gland ablation would be expected to have reached a low to undetectable nadir PSA value (typically <0.6 ng/mL). In addition, our Kaplan–Meier analysis of bPFS stratified by the various nadir PSA subgroups would support the hypothesis that the risk of biochemical failure increases as the nadir PSA increases. The important clinical question arises of how can we use these findings in the care of patients undergoing prostate salvage cryotherapy with curative intent. We propose that patients failing to achieve a nadir PSA <2.5 ng/mL should be carefully observed, and that salvage hormonal ablative therapy should be considered when clinically indicated, such as in the presence of a short PSA doubling time or where there is clinical evidence of metastatic progression. The present study emphasizes that the nadir PSA achieved after cryotherapy is an important therapeutic endpoint to document and consider in our care of patients.
In addition, the present study indicates that pre-salvage biopsy Gleason score is a strong predictor of bPFS, as shown in our multivariate Cox regression analysis and Kaplan–Meier survival plots. Although the concept that pre-salvage biopsy Gleason score can help predict the treatment outcomes of salvage prostate cryotherapy has been previously published [8, 10], the present study emphasizes its strong prognostic role, with it constituting one of only two statistically significant clinical variables in a multivariate model. In our Kaplan–Meier analysis, we report 1-, 2-, and 3-year bPFS rates for patients with a pre-salvage biopsy Gleason score ≤6 of 90.9, 80.4 and 78.2%, respectively, as compared with rates of 83.0, 59.7 and 50.5% at similar timepoints for the subset of patients with a pre-salvage biopsy Gleason score ≥7. The present study is consistent with the previous study, which used a pre-salvage nomogram predicting the likelihood of biochemical failure with prostate salvage cryotherapy and used the pre-salvage biopsy Gleason score as one of its prognostic variables .
Of our study cohort, only 19 patients (14.4%) had a nadir PSA >2.5 ng/mL, which supports the premise that our collaborating physicians within the COLD registry have performed well in selecting patients best suited for prostate salvage cryotherapy. Furthermore, most of the study cohort had a nadir PSA <0.1 ng/mL or between 0.1 and 0.6 ng/mL (35.6 and 29.8%, respectively).
Within our patient cohort treated with prostate salvage cryotherapy in the absence of neoadjuvant, concomitant or adjuvant hormonal ablative therapy, we report encouraging outcomes with 1-, 2-, and 5-year bPFS rates of 87.8, 72.4, and 45.5%, respectively. The present study population was unique in that the therapeutic outcomes of salvage prostate cryotherapy in the absence of any form of hormonal manipulation were assessed. Our results suggest that there was appropriate selection of patients for prostate salvage cryotherapy in addition to improvements in the technology of cryotherapy devices and their respective imaging techniques.
We acknowledge several limitations to the present study, including the fact the COLD registry constitutes a registry database collected across multiple treatment centres (academic and community). Data entry was completed by a number of different healthcare professionals which led to some degree of data heterogeneity and potential data entry inconsistencies. It is also noteworthy that not all data were entered for all patients within the COLD registry, including details of the type of primary radiation therapy (type, dose, neoadjuvant/adjuvant hormonal therapy) received in the vast majority of patients. The role of industry funding is a possible limitation, but the separation of the sponsor from the data through the intermediary (Watermark, Inc) is intended to limit this potential. Furthermore, there was no central pathological review, and many pathologists now suggest that Gleason scoring after radiation therapy may not be meaningful or reproducible. Finally, it is possible that the outcomes could be different with longer-term surveillance.
The COLD registry whole-gland salvage cryotherapy patient population (N = 614) upon which the present data analysis was based was similar to that of our previous publications, but the inclusion and exclusion criteria used in selecting the study population of interest differed. For example, in the present study, we excluded patients who had received neoadjuvant/concomitant/adjuvant hormonal ablative therapy as well as patients who had fewer than two post-salvage cryotherapy PSA measurements, which resulted in a study population of 132 patients. In a previously published salvage prostate cryotherapy peer-reviewed publication , we used the same COLD registry salvage whole-gland cryotherapy patient population but excluded patients receiving neoadjuvant hormonal ablative therapy as well as patients missing data pertaining to their post-salvage cryotherapy continence status, as this was pertinent to the clinical question of interest, which resulted in a study population of 183 patients.
It is difficult to determine the exact proportion of patients treated with prostate cryotherapy captured within the COLD registry, in this case in the salvage whole-gland setting, but it is estimated that ≤10% of the total number of patients treated in North America with salvage cryotherapy using third-generation cryotherapy devices are captured within the COLD registry. The COLD registry is unique as it captures both the experience and outcomes of community and academic centres with prostate cryotherapy.
Unfortunately, salvage cryotherapy outcomes have not improved over the past 10 years [5, 6]. There are multiple possible reasons for this, including poor patient selection and delayed referrals to tertiary care prostate cryotherapy centres. We believe imaging methods such as endorectal MRI and positron emission tomography/CT may play a role in the evaluation of patients before salvage prostate cryotherapy with curative intent, although this needs to be further evaluated in future prospective trials .
Previous studies have characterized other important prognostic factors within this patient population, including the pre-radiotherapy clinical stage, pre-salvage total serum PSA level and doubling time, and biopsy Gleason score [5, 7, 8]. Despite these efforts, we continually strive to better define patients best suited for salvage prostate cryotherapy and the present study provides impetus for further investigation of salvage cryoablation.
In conclusion, the long-term oncological results of salvage cryoablation are encouraging and support the curative potential of this salvage treatment method in appropriately selected patients. The most important prognostic factors of bPFS within this patient population consist of the nadir PSA achieved post-salvage therapy and the pre-salvage prostatic biopsy Gleason score. Based on these two prognostic variables, we have developed risk stratification groups (low, intermediate, and high) which help predict the expected outcomes of salvage whole-gland prostate cryotherapy in a given patient. This risk stratification grouping constitutes a useful clinical tool in defining which patients may be best suited for this local salvage treatment method.