Correlation of prostate-specific antigen nadir and biochemical failure after high-intensity focused ultrasound of localized prostate cancer based on the Stuttgart failure criteria – analysis from the @-Registry
Department of Urology, University of Regensburg, Regensburg
Roman Ganzer, University of Regensburg, Department of Urology, Krankenhaus St Josef, Landshuter Straße 65, D-93053 Regensburg, Germany. e-mail: firstname.lastname@example.org
Study Type – Prognosis (retrospective cohort)
Level of Evidence 2b
What’s known on the subject? and What does the study add?
The prostate-specific antigen (PSA) nadir after high-intensity focused ultrasound (HIFU) therapy for clinically localized prostate cancer has been shown to correlate with treatment failure in single centre studies. Purpose was to determine if the PSA nadir after HIFU can be used as a predictor of the biochemical disease-free survival in a multicentre study.
This multicentre analysis confirms that PSA nadir after HIFU predicts biochemical DFSR in a statistically significant manner.
• To determine if the prostate-specific antigen (PSA) nadir after high-intensity focused ultrasound (HIFU) can be used as a predictor of the biochemical disease-free survival rate (DFSR).
PATIENTS AND METHODS
• Patient data were derived from the multicentre-based @-Registry, the largest registry to report outcomes in patients with localized prostate cancer after Ablatherm® HIFU.
• PSA level was measured at 3-month intervals. Patients were stratified into four PSA nadir groups: group 1, ≤0.2 ng/mL; group 2, 0.21–0.5 ng/mL; group 3, 0.51–1 ng/mL; and group 4, >1 ng/mL.
• Biochemical treatment failure was defined according to the Stuttgart definition (PSA nadir + 1.2 ng/mL) and the Phoenix definition (PSA nadir + 2 ng/mL).
• Biopsy was performed at 3–6 months post-HIFU or if a PSA level was recorded that was considered clinically relevant.
• The present study included 804 patients. Biochemical treatment success rates at 5 years according to the Stuttgart definition for the four PSA nadir sub-groups were as follows: 84, 64, 40 and 30% for groups 1–4, respectively.
• The equivalent 5-year biochemical success rates using the Phoenix definition were 94, 74, 66 and 47%, respectively.
• Significantly more patients had a negative biopsy in the lowest PSA nadir group than in the other sub-groups (91.6 vs 73.1%; P < 0.001).
• The present study is limited by its retrospective nature and variations in clinical practice across participating centres.
• This multicentre analysis confirms that PSA nadir after HIFU predicts biochemical DFSR in a statistically significant manner.
American Society of Therapeutic Radiology and Oncology
disease-free survival rate
external beam radiotherapy
high-intensity focused ultrasound
The PSA nadir after high-intensity focused ultrasound (HIFU) therapy for clinically localized prostate cancer has been shown to correlate with treatment failure . Uchida et al. reported on a series of 115 patients from a single centre treated with HIFU and showed that those patients with a PSA nadir of 0–0.2 ng/mL had a lower clinical failure rate than patients with a PSA nadir >0.21 ng/mL. Similar results with longer follow-up were reported by Ganzer et al., who evaluated 103 patients for a median (range) follow-up of 4.9 (3.0–8.6) years and reported that a PSA nadir ≤0.2 ng/mL was strongly correlated with freedom from disease recurrence. To date, there is no universal consensus on the definition of biochemical failure in patients treated with HIFU for prostate cancer. Therapeutic success after HIFU has been previously described using a variety of definitions. The earliest studies defined a complete response as a negative postoperative biopsy and a PSA level <4.0 ng/mL [3,4]. Stricter response criteria of treatment failure were applied by Gelet et al. as any positive biopsy or three successive elevations of PSA with a rate of 0.75 ng/mL per year or greater. The American Society for Therapeutic Radiology and Oncology (ASTRO) definition of disease-free status based on biochemical outcome has more recently been applied to HIFU . Owing to the difficult application in clinical practice, the ASTRO criteria were revised in 2005 with the Phoenix ASTRO criteria, defined as the time at which the PSA nadir + 2.0 ng/mL is reached . Although these criteria were restricted to radiotherapy series, they have also been used in HIFU series to define biochemical failure.The Stuttgart definition was then developed using information available from patients treated with HIFU. This definition proposed that the best positive predictive value for eventual clinical recurrence was achieved when biochemical failure was defined as the time from treatment to the rise in PSA of 1.2 ng/mL over nadir value . Of a number of PSA scenarios tested, this new definition was shown to correlate most closely with clinical failure after HIFU, although it has not yet been validated by other groups.
The present study examines biochemical failure defined according to the recently published Stuttgart criteria as well as the Phoenix definition and their correlation with four different PSA nadir values as predictors of clinical failure. Patient data were derived from the multicentre-based @-Registry, the largest registry to report outcomes in patients after HIFU for localized prostate cancer.
PATIENTS AND METHODS
The @-Registry includes data from consecutive patients with prostate cancer who have been treated with Ablatherm® HIFU (EDAP TMS, Vaulx-en-Velin, France) during the period February 1993 to July 2009. Data from 3236 patients were reviewed and further analyses conducted on patients based on the following inclusion criteria: minimum follow-up 3 years; localized disease (T1–2); no previous hormone therapy; Gleason score ≤7; and PSA level ≤20 ng/mL. Patients who underwent TURP at the time of HIFU or before HIFU were also included in the analysis. The patients fitting these criteria were identified as low, intermediate and high risk according to D’Amico’s 2003 risk group categories .
After HIFU, patients were followed with PSA measurement at 3-month intervals. Biopsy was usually recommended at 3–6 months post-HIFU or if a PSA level was recorded that was considered clinically relevant by the treating physician. Secondary treatment was instituted according to individual physician clinical judgment based on PSA levels and biopsy results. Patients were stratified into four PSA nadir groups: group 1, ≤0.2 ng/mL; group 2, 0.21–0.5 ng/mL; group 3, 0.51–1 ng/mL; and group 4, >1 ng/mL. Biochemical treatment failure was defined according to the Stuttgart definition (PSA nadir + 1.2 ng/mL) and the Phoenix definition (PSA nadir + 2 ng/mL).
Statistical analyses were carried out with SPSS statistical software, version 17 (SPSS, Chicago, IL, USA). Depending on distributions, parametric and non-parametric tests were applied. Survival curves were based on Kaplan–Meier models and the log-rank test was used for univariate comparisons. Disease progression was defined by biochemical failure (Stuttgart or Phoenix definitions). P < 0.05 was considered to indicate statistical significance.
A total of 804 patients were included in this retrospective analysis. The number of patients included from the individual centres was as follows: Lyon (France), 434; Regensburg (Germany), 165; Como (Italy), 109; and Montpellier (France), 96.
As predetermined for this analysis, patients excluded were those who had previously undergone hormonal therapy, with less than 3-years follow-up or with high-risk disease. The baseline characteristics are shown in Table 1. The median (range) follow–up was 5.0 (3–15) years, the mean (SD) follow-up was 5.9 (2.2) years and the mean (SD) time to reach PSA nadir was 14.9 (12.1) weeks. PSA nadirs of ≤0.2, 0.21–0.5, 0.51–1 and >1 ng/mL were reached by 428 (53.2%), 130 (16.2%), 90 (11.2%) and 156 (19.4%) patients, respectively. Prostate volume was determined intraoperatively immediately before HIFU in most (80%) patients and from transrectal ultrasonography reported at the time of prostate biopsy for the remainder. PSA level before TURP and baseline prostate volume were both significantly higher in patients who achieved a PSA nadir >1 ng/mL (group 4; P < 0.001). There was a significant difference with respect to the percentage of patients undergoing TURP before to HIFU treatment across the groups (P < 0.001). The distribution of patients according to D’Amico risk group categories (low, intermediate and high risk) was balanced across all four PSA nadir sub-groups. Similarly, no significant differences were shown for distribution of disease stage (T1 and T2) across the sub-groups.
Table 1. Baseline characteristics for patients who underwent HIFU. Patient numbers in parenthesis: number for whom data are available if not the entire patient group/sub-group cohort
Group 1 (PSA ≤ 0.2 ng/mL)
Group 2 (PSA 0.21–0.5 ng/mL)
Group 3 (PSA 0.51–1 ng/mL)
Group 4 (PSA >1 ng/mL)
NS, not significant.
No. of patients
Mean (sd) PSA, ng/mL
7.7 (3.9) (n= 804)
7.0 (3.8) (n= 427)
7.8 (3.7) (n= 130)
8.9 (4.1) (n= 90)
9.0 (3.9) (n= 156)
Median (range) Gleason score
6 (2–7) (n= 800)
6 (2–7) (n= 425)
6 (2–7) (n= 130)
6 (3–7) (n= 90)
6 (2–7) (n= 155)
Mean (sd) prostate volume, mL
23.9 (10.4) (n= 740)
21.6 (9.1) (n= 399)
23.9 (9.5) (n= 121)
26.0 (10.1) (n= 77)
28.9 (12.5) (n= 143)
No. (%) of patients with TURP before HIFU
431 (56.0) (n= 769)
241 (58.1) (n= 415)
72 (58.1) (n= 124)
54 (63.5) (n= 85)
64 (44.1) (n= 145)
No. (%) of patients at low risk
No. (%) of patients at intermediate risk
No. (%) of patients at high risk
No. (%) patients with disease stage s
Treatment variables for the total patient cohort and PSA nadir sub-groups are shown in Table 2. With regard to the mean number of lesions per HIFU treatment and mean prostate treatment volume, no significant differences were identified between any of the groups. There was a significant difference reported for mean number of HIFU sessions between groups (P < 0.036) as well as for mean treated volume ratio (i.e. treated volume to measured prostate volume; P < 0.001).
Table 2. Treatment parameters for total patient set and according to PSA nadir sub-group. Data are means (sd)
Group 1 (PSA ≤ 0.2 ng/mL)
Group 2 (PSA 0.21–0.5 ng/mL)
Group 3 (PSA 0.51–1 ng/mL)
Group 4 (PSA >1 ng/mL)
NS, not significant.
Lesions per treatment
Treated volume (mL)
Treated volume ratio (treated volume/prostate volume)
Biochemical treatment success rates at 5 years according to the Stuttgart definition for the four PSA nadir subgroups were 84, 64, 40 and 30%, respectively (P < 0.001). The equivalent 5-year biochemical success rates using the Phoenix definition were 94, 74, 66 and 47%, respectively (P < 0.001). Kaplan–Meier estimates up to 7 years according to the Stuttgart and Phoenix definitions are shown in Fig. 1 for the entire populations and in Fig. 2 according to the PSA nadir sub-groups. Biopsy results were available for 649 (80.7%) patients during follow-up (Table 3). Results indicate that significantly more patients had a negative biopsy in group 1, the lowest PSA nadir group, than in the other sub-groups (91.6 vs 73.1%; P < 0.001).
Table 3. Negative biopsy rates in the total patient cohort and according to PSA nadir sub-group. Data are n (%)
High-intensity focused ultrasound is an evolving treatment option for localized prostate cancer. Five-year progression-free survival rates for HIFU are of the order of 69–78%[10,11] and are similar to other treatment modalities, such as 76–84%[12–14] reported for radical prostatectomy, and 55–65%[15,16] reported for external beam radiotherapy (EBRT) in the patients with equivalent clinical stage (T1–T2), although the studies on the latter two therapies included a minority of patients with Gleason score >8. The value of PSA nadir in predicting outcome of HIFU has been examined in a previous publication . This single-centre study involved a small cohort of 103 patients and showed that PSA nadir after HIFU correlated highly significantly with 5-year DFSR based on the Phoenix criteria, with rates of 95, 55 and 0% for PSA nadir sub-groups of ≤0.2, 0.21–1 and >1 ng/mL, respectively. The present study has gone further to sub-divide the second PSA nadir sub-group into nadirs of 0.21–0.5 and 0.51–1 ng/mL. Additionally, the present multicentre study included 804 patients, a much larger cohort than previously evaluated, specifically from the Ablatherm treatment registry (@-Registry). This secure online database consists of case report forms specifically engineered to collect appropriate pre- and post-treatment information for patients who have undergone prostate HIFU utilizing the Ablatherm device from multiple treatment centres. It represents the single largest multi-institutional repository of prospectively maintained information regarding the efficacy of prostate cancer therapy with HIFU.
In the current analysis, 5-year biochemical DFSR after HIFU has been determined using two newer definitions of failure: the Phoenix and the Stuttgart. The latter definition of PSA nadir + 1.2 ng/mL has been shown to have a 77.5% sensitivity and a 79.0% specificity in detecting biochemical failure in a previous analysis of patients in the @Registry . The present analysis of the @-Registry shows that, independent of the definition of failure applied, PSA nadir is an early predictor of treatment success. Excellent 5-year biochemical DFSRs were obtained for 84–94% of patients if a PSA nadir ≤0.2 ng/mL was reached, and for 64–74% of patients if a nadir of 0.21–0.5 ng/mL was reached.
This report confirms the findings of the previous comparative analysis of the Phoenix and Stuttgart definitions, where high PSA nadir values (Phoenix: PSA nadir + 2 ng/mL) were associated with a higher specificity and apparently better outcomes than a definition such as Stuttgart, which, by employing a lower PSA nadir threshold (Stuttgart: PSA nadir + 1.2 ng/mL), is both specific and sensitive . It is noteworthy that PSA often increases initially above nadir with therapies such as HIFU, brachytherapy and EBRT. It is important, therefore, that the decision as to whether an increase in PSA level is clinically meaningful is based on the managing physician’s discretion. Nevertheless, an increase above PSA nadir by 1.2 ng/mL should give rise to concern.
The present analysis of the @-Registry also evaluated biopsy outcome according to PSA nadir sub-group. A significantly higher negative biopsy rate was reported for the lowest PSA nadir group (≤0.2 ng/mL) than for PSA nadir groups of 0.21–0.5, 0.51–1.0 and >1 ng/mL (P < 0.001). Poissonier et al.  reported on a study of 227 patients treated with the Ablatherm HIFU device, with treatment failure defined as any positive biopsy or a PSA level >1 ng/mL with three consecutive rises . Patients with significantly rising PSA levels received adjuvant treatment; however, of the 28 patients in this category, nine (32%) had negative biopsies. The authors would like to suggest that biopsy findings alone should not be used as a criterion for treatment failure.
The patient baseline data in the present study show similar variables among the four different PSA nadir groups, except for baseline PSA, prostate volume and rate of TURP before HIFU. Patient selection is of paramount importance to achieve a post-treatment PSA nadir <0.2 ng/mL and excellent oncological outcome. Of the variables, the only one that can be manipulated is the prostate size and anterior-posterior distance through the use of pre-HIFU TURP for downsizing of the gland. The shrinkage effect can be even more extended by separating TURP and HIFU by an interval of some weeks . The concept of TURP before HIFU was established by Vallancien in 2004 . It could be shown that it furthermore reduces catheterization time and improves urinary status without increased morbidity. The findings in the present study support the recommendation to include TURP into the HIFU treatment concept.
Registry data aim to assess the real-life situation of a therapy. Although this is valid in the present study, there are limitations to using such data, as they are highly dependent on the way the registry has been set up . At best the data might be an accurate record of outcome at one or more sites, but they can be subject to variations in practice and are limited in value when making comparisons with other studies.
In conclusion, the present multicentre analysis confirms that PSA nadir after HIFU predicted biochemical DFSR in a statistically significant manner. Excellent oncological outcome was obtained for 84–94% of patients if a PSA nadir ≤0.2 ng/mL was reached, and for 64–66% of patients if a PSA nadir ≤0.5 ng/mL was reached. The @-Registry represents that largest study with the longest follow-up in this field and this analysis of patients included in the registry confirms that PSA nadir is a reliable indicator of success of HIFU that can easily be used in daily clinical practice.
CONFLICT OF INTEREST
None declared. Source of Funding: supported by an unrestricted educational grant from EDAP.