- To report the oncological and functional outcomes of hemi salvage high-intensity focused ultrasound (HSH) in patients with unilateral radiorecurrent prostate cancer.
Patients who receive radiotherapy (RT) for localised prostate cancer have a risk of disease recurrence . The Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE) study found a two-fold higher likelihood of cancer-specific mortality in patients treated with RT vs radical prostatectomy (RP) . However, for patients with high-risk prostate cancer, no significant difference in long-term cancer control was found when comparing external beam RT (EBRT) combined with androgen-deprivation therapy (ADT) with RP . Using the Phoenix criteria, the GETUG 06 study found the estimated risk of biochemical recurrence (BCR) after EBRT to be 32% in patients treated with 70 Gy and 23.5% when treated with 80 Gy, after 5 years . After brachytherapy (BT), the Seattle group reported a BCR rate close to 20% at 15 years follow-up and a cancer-specific mortality rate of 16% . Most patients (>90%) with BCR receive ADT, that supresses the PSA output.
Every whole-gland salvage therapy approach, e.g. RP, high-intensity focused ultrasound (HIFU) or cryotherapy, carries the potential of serious morbidity, such as development of severe urinary incontinence, urethral stenosis, and urethrorectal fistula [6-9]. To date, the best predictor of successful treatment outcome is pre-salvage therapy serum PSA level, with better outcomes associated with PSA levels of ≤5 ng/mL . MRI-guided targeted prostate biopsy may be valuable in detecting EBRT recurrence [6, 11, 12].
In patients who present small-sized local recurrence after primary RT, hemi salvage HIFU (HSH) therapy has been shown to be an effective treatment with reduced complications and side-effects .
The objective of the present two-centre study was to evaluate the oncological and functional outcomes of HSH in patients with unilateral radiorecurrent prostate cancer.
Between 2009 and 2012, 48 consecutive patients (mean [sd, range] age of 68.8 [6.0, 58–82] years) with radiorecurrent prostate cancer were enrolled for participation at Edouard Herriot Hospital, Lyon, France (27 patients) and Oslo University Hospital, Aker, Norway (21). This study received approval by the local ethics committees in France and in Norway. All patients signed a letter of informed consent before enrolment. No patient was missing during the follow-up.
Inclusion criteria were BCR (defined as Phoenix criteria: nadir ≥2 ng/mL, in 46/48, or three consecutive rises in PSA level in two of 46), unilateral prostate cancer at MRI verified with prostate biopsy and absence of metastases verified at bone scan, pelvic CT or MRI. Previous therapy for prostate cancer included EBRT in 46 (mean [sd, range] 72.5 [3.3, 64–78] Gy) and BT in two, and 11 (23%) also received neoadjuvant hormone therapy.
The Gleason score, D'Amico risk group, and PSA levels before and after RT, are given in Table 1 and the patients' characteristics before HSH are given in Table 2. The mean (sd, range) time from the end of RT to HSH was 5.9 (2.6, 2.7–14.6) years.
|D'Amico risk group|
|Mean (sd, range)|
|Gleason score||6.7 (1.3, 3–9)|
|PSA level, ng/mL||14.2 (12.1, 2.9–70)|
|Mean (sd, range)|
|Nadir PSA, ng/mL||0.71 (0.64, 0–3.3)|
|Years to nadir PSA||1.8 (1.2, 0–5.5)|
|Years to nadir +2||4.8 (2.6, 1.8–12.8)|
|Gleason score, n (%)|
|Mean (sd, range):|
|Total biopsy samples||14.9 (5.1, 6–32)|
|Positive biopsies||3.4 (1.8, 1–8)|
|Prostate volume, mL||21.2 (9.7, 7–60)|
The Lyon group used a 3T MR 750® MRI (General Electric, Waukesha, WI, USA) and a 32-channel pelvic phased-array coil. Sequences were axial T2-weighted (T2w), diffusion-weighted images (DWI) using b0 and b2000 to generate apparent diffusion coefficient (ADC) maps, and dynamic contrast-enhanced (DCE) MRI (temporal resolution, 7 s) after i.v. injection of 0.2 mL/kg gadoterate meglumine (Dotarem®, Guerbet, Roissy, France) at 3 mL/s.
The Norwegian group used a 1.5 T Tesla Avanto® MRI (Siemens, Erlangen, Germany) and a six-channel Body MATRIX® coil. Sequences were axial T2w, DWI using b50 and b1000 to generate ADC maps, and an additional b2000. Axial and coronal T1w images were used to identify skeletal and lymph node metastases. Nordic ICE® (NordicNeuroLab, Bergen, Norway) software was used for post-processing images as previously described by Rud et al. .
Local recurrence of prostate cancer was defined as a focal area with high signal intensity on b1000 or b2000 native images compared with muscle signal, an ADC of <120 × 10-5 mm2/s in the peripheral zone, and an ADC of <100 × 10-5 mm2/s in the transitional zone.
BCR was defined using Phoenix criteria (a rise by ≥2 ng/mL above the nadir PSA level. D'Amico risk categories were low/intermediate and high .
The following schemes and criteria for functional outcomes were used: IPSS, Ingelman-Sundberg score  using the ICS questionnaire (A and B) , International Index of Erectile Function-5 (IIEF-5 points) , and the European Organisation for the Research and Treatment of Cancer (EORTC) quality of life questionnaires (QLQ-C-30) .
TRUS-guided prostate biopsy consisted of 12 randomised cores that included the proximal area of the seminal vesicle, and 1–3 biopsies from each MRI target using MRI/TRUS image fusion technology (Koelis®, La Tronche, France) .
HSH was performed under spinal or general anaesthesia using the Ablatherm® Integrated Imaging device (EDAP TMS, Vaulx-en-Velin, France). During the fire phase, the software automatically controlled the rectal position, and a cooling system maintained rectal mucosa temperature at 14 °C. Between each fire phase, the focal point position inside the prostate gland was controlled by the operator, in real time .
In patients with negative apical biopsy, a 4-mm security distance was held to the sphincter, while in positive apical biopsy, the HIFU shots were performed closer to the sphincter. In patients with tumour invasion in the prostate base, the HIFU shots included the proximal region of the seminal vesicles.
Patients with obstructive voiding symptoms at the time of treatment underwent an endoscopic bladder neck resection or incision during the same procedure to prevent postoperative obstruction. In these patients, treatment area was extended medially with at least two HIFU shots that included urethra into the treatment zone as shown in Fig. 1.
Specific radiorecurrence parameters for salvage HIFU treatment were introduced in 2002 and incorporate the unique characteristics of irradiated tissue, e.g. compromised vascularisation and radiation-induced fibrosis .
In the Lyon group, patients were discharged after 72 h after Foley catheter removal and post-voiding residual measurement by ultrasound.
In the Oslo group, patients were discharged either on the operative day or after 24 h depending on home distance and comorbidity. Foley catheter removal and post-voiding residual measurement by ultrasound were performed in an out-patient clinic 5 days after HSH.
Follow-up during the first year occurred every 3 months and included oncological and functional outcome parameters as described before HSH.
Patients with rising PSA levels underwent control MRI prostate and subsequent prostate biopsy. Patients with biopsy confirmed local recurrence were encouraged to undergo a second HSH treatment, particularly when malignancy was found in the untreated contralateral lobe. ADT was suggested to patients who refused a second HSH treatment after BCR or histological recurrence, or metastases.
In cases of BCR without local recurrence on MRI the Lyon group underwent a metastatic evaluation, consisting of an 11C-choline positron emission tomography (PET) scan, and the Oslo group, bone scan and pelvic MRI. Control prostate biopsy was not taken in patients with confirmed metastases.
Statistical analyses were performed with IBM SPSS Statistics software v.21 (IBM Corp., New-York, NY, USA). Depending on distribution, paired sample t-test or Wilcoxon signed-ranks test were used to compare oncological and functional outcomes before and after HSH.
Survival curves were based on Kaplan–Meier models, and the log-rank test was used for univariate comparisons. Actuarial survival rates were based on life table methods. Progression-free survival (PFS) rates were calculated using the combined criteria of BCR (Phoenix criteria) and/or the need for ADT, whichever occurred first. Patients were censored on the date of their last PSA evaluation. A P < 0.05 was considered to indicate statistical significance.
The mean (sd) PSA nadir after HSH was 0.69 (0.83) ng/mL, and the time to attaint PSA nadir was 21 (20.7) weeks. The median (interquartile range) follow-up was 16.3 (10.5–24.5) months. At the end of the follow-up period 32/48 (67%) were free of BCR. Disease progression was identified in 16/48 (33%), in which six (13%) were confirmed metastases.
Local recurrence was identified at MRI and control prostate biopsy in eight of the 48 patients (17%). Prostate biopsy revealed cancer foci in both lobes in four patients, and in the untreated lobe in another four. BCR without proven local recurrence or metastases occurred in two patients. Local recurrences in four of 16 (25%) patients were treated with a second HSH session and 12/16 (75%) received ADT.
The overall PFS rates at 12, 18, and 24 months, with respect to D'Amico risk group, Gleason score before HSH and PSA level before and after HSH are presented in Table 3. The D'Amico risk group before EBRT did not influence the 18-month PFS rate. In contrast, the 18-month PFS rate was significantly associated with Gleason score during post-EBRT recurrence (≤7, 82%; ≥8, 34%; P = 0.047), and by the PSA level before HSH (≤4 ng/mL, 80%; >4 ng/mL, 49%, P = 0.002).
|PFS rate||% (n at risk)||P|
|Overall PFS rate at:|
|12 months||83 (37)|
|18 months||64 (23)|
|24 months||52 (11)|
|PFS rate according to:|
|Initial D'Amico risk group:||0.599|
|Low / intermediate at:|
|12 months||88 (17)|
|18 months||66 (8)|
|24 months||33 (2)|
|12 months||74 (17)|
|18 months||54 (12)|
|24 months||45 (6)|
|Gleason score before HSH||0.047|
|12 months||88 (23)|
|18 months||82 (15)|
|24 months||63 (9)|
|12 months||34 (8)|
|18 months||34 (2)|
|PSA level before HSH:||0.002|
|≤4 ng/mL at:|
|12 months||94 (18)|
|18 months||80 (14)|
|24 months||70 (8)|
|>4 ng/mL at:|
|12 months||73 (19)|
|18 months||49 (9)|
|24 months||32 (3)|
|PSA nadir after HSH||0.273|
|≤0.5 ng/mL at:|
|12 months||87 (19)|
|18 months||72 (12)|
|24 months||72 (5)|
|>0.5 ng/mL at:|
|12 months||77 (18)|
|18 months||56 (11)|
|24 months||37 (6)|
The 18-month PFS rate did not differ significantly between patients with a PSA nadir after HSH of ≤0.5 ng/mL and those with a PSA nadir of >0.5 ng/mL after HSH (72% vs 56%, P = 0.3).
Spontaneous voiding re-established after a mean (sd) catheterisation time of 3.9 (1.2) days. The pad-free, leak-free urinary continence status after HSH was 36/48 (75%) patients, one pad a day in eight (17%), and severe incontinence in four (8.3%). All four of these patients had a local recurrence involving the apex, and the HSH therefore was not performed using a sphincter safety margin after patient consent. However, three of these four patients were disease free at last follow-up. The functional outcomes of the ICS (A and B), IPSS, IIEF-5 and EORTC QLC-30 are given in Table 4.
|Questionnaire||Mean (sd) score||P|
|before HIFU||after HIFU|
|ICS (A)||0.7 (2.0)||2.3 (4.5)||0.045|
|ICS (B)||0.6 (1.4)||1.6 (3.0)||0.114|
|IPSS||7.1 (5.6)||8.6 (5.1)||0.129|
|IIEF-5||11.2 (8.0)||7.0 (5.8)||<0.001|
|EORTC QLC-30||35.7 (8.7)||36.8 (8.6)||0.220|
Two patients had delayed pubic bone osteitis and one of these developed pubovesical fistula appearing 23 months after HSH. No rectal fistula occurred.
The present study is the first prospective two-centre study on salvage focal treatment with total cohort follow-up. However, there are several limitations of the present trial.
Firstly, MRI protocols used in detection of prostate cancer were not standardised across institutions. The Lyon group performed MRI combining T2w, DW MRI and DCE MRI while the Oslo group performed MRI using T2w and DW MRI. To our knowledge, only one study  (n = 16) has compared T2w, DW MRI and DCE MRI after BT. They reported a 77% sensitivity when combining all three sequences, contrary to 68%, when using only T2w and DW MRI (P > 0.05) . Diagnosis of radiorecurrent prostate cancer using T2w and DW MRI sequences is documented by several studies reporting sensitivity ranging from 62% to 93% [21, 22]. In addition, a recent study did not show a significant difference in cancer detection comparing 1.5 T and 3 T devices .
Secondly, we did not take a control prostate biopsy in patients with BCR and distant metastases occurring during the follow-up, as the biopsy result would not influence the treatment and the biopsy procedure might lead to unnecessary complications.
Thirdly, 11C-choline PET scan, which seems to be the most effective method for metastatic evaluation, was not used in all patients with BCR. However, the diagnosis of metastatic diseases was not an endpoint in the present study.
Fourthly, the present study was limited by the relatively brief follow-up period and limited patient cohort, which hinders any conclusion about oncological effectiveness. A follow-up of 5 years is probably necessary for more accurate biochemical evaluation, cancer-specific and metastases-free survival outcomes after HSH.
Finally, the validity of using Phoenix criteria for failure could be criticised. Nevertheless, to date no common agreement exists on the definition of BCR after ablative therapies. Furthermore, in the largest published reports on HIFU and cryotherapy, oncological outcomes are based on Phoenix criteria as applied to post-RT recurrence [6, 24, 25].
Accurate patient selection is an essential pre-condition for achieving optimal cancer control with a focal salvage-therapy approach. In theory, the selection process should first exclude patients with subclinical metastases, and then further evaluate only those with small unilateral local recurrences. Detection of occult lymph node and bone metastases is hampered by limitations in current imaging technology, and by characteristics of the malignancy that make visualisation difficult. Metastases were detected during follow-up in two of 39 patients (5%) in the Ahmed et al.  study, and in six of 48 patients (12.5%) in the present study. The high-rate of metastases in the present study may be explained by the elevated percentage of patients with high-risk disease, who probably could have had micrometastases before HSH.
Broad awareness exists of the shortcomings of standard pretreatment metastases evaluation, typically involving bone scintigraphy and abdominal pelvic CT/MRI . Improvement in detecting metastases has recently been reported using whole-body MRI and 11C-choline PET/CT [27, 28], and may offer better pretreatment evaluation in the future.
The best biopsy strategy for patient selection for focal therapy is still controversial. A transperineal template-guided mapping biopsy of the prostate (TTMB) appears to provide more detailed information about cancer grade and localisation compared with a standard 12-core biopsy schema . The principal drawback of TTMB remains in its invasive approach, high cost, and complexity, which is directly divergence to the mini-invasive concept of focal therapy. In addition, due to prostate movement, organ swelling and biopsy needle deflection during the procedure, biopsy needle placement does not necessary correspond to the localisation suggested by a grid.
In recent years, numerous studies have shown an increasing agreement between MRI and RP specimen results for determining anatomical location and extent of disease, especially if the tumour volume is >0.5 mL or Gleason grade ≥4. However, MRI still has limitations in the detection of small and multiple cancer foci [30, 31].
Recent advances in the three-dimensional (3D) registration of the biopsy based on organ tracking using MRI/TRUS real-time elastic-fusion techniques may provide more objective tumour mapping than the traditional biopsy technique . We think that the future progress in imaging and computerised biopsy procedures may further improve patient selection for focal therapy and compensating for the disadvantages of TTBM.
Moreover, prostate MRI has yielded encouraging results in detecting and localising recurrence after EBRT and is systematically used in both patient selection and treatment guidance at our institutions (Fig. 2) [11, 12].
Evaluation of 50 salvage prostatectomies by Leibovici et al. , found that 66% of patients had a solitary cancer focus, and that 74% had tumour extension beyond the urethra. Therefore, it appears important to perform an ‘extended’ hemi-ablation involving the contralateral lobe without the intention of urethral preservation. Furthermore, early detection of radiorecurrent cancer, while the cancer is still amenable to focal therapy, is mandatory.
In the present study, nadir PSA values achieved after HSH were somewhat higher than those reported with whole-gland HIFU treatment. The largest series published to date on salvage HIFU therapy is by Crouzet et al.  with 290 patients and a mean 48-month follow-up. They reported a 0.14 ng/mL median nadir PSA after whole-gland treatment, and actuarial cancer-specific and metastases-free survivals of 80% and 79%, respectively. These rates correspond with those reported by Chade et al.  in a study of 404 patients treated with salvage RP. Their actuarial rates of cancer-specific and metastases-free survival were 83% and 77%.
After focal salvage HIFU, Ahmed et al.  reported an actuarial PFS rate of 49% at 2 years (Phoenix criteria), with 16 patients (41%) undergoing palliative hormone therapy by final follow-up. The biochemical survival outcomes reported in the Ahmed et al. study were comparable to the PFS rates in the present study, achieved with hemi-ablation, of 52% at 2-year follow-up.
The morbidity prevalence in the present study was lower than those reported for standard salvage treatments. Nguyen et al.  evaluated the outcome of salvage RP in 531 patients and found an incontinence rate of 41%, an anastomotic stenosis rate of 24% and a 4.7% rate of urethrorectal fistula. In a more recent series of salvage RP, a 20% rate of severe incontinence and a 1% rate of urethrorectal fistula were reported . Salvage cryotherapy was evaluated in 510 patients by Nguyen et al. , who found severe incontinence in 36%, bladder neck stricture in 17%, and urethrorectal fistula in 2.6%. In a series of 84 patients treated with salvage HIFU, Ahmed et al.  reported incontinence in 38%, strictures requiring endoscopic intervention in 20%, and urethrorectal fistula in 2.4%. In a series published by Crouzet et al. , HIFU morbidity for radiorecurrent cancer was significantly reduced using dedicated acoustic parameters, with severe incontinence in 19.5% and urethrorectal fistula in 0.4%. Similarly, Berge et al.  used the same acoustic parameters for whole-gland salvage therapy, and reported severe urinary incontinence in 17.2%.
The preservation of both urinary function and health-related quality of life at pre-treatment levels (Table 4) in the present study is in contrast to the results of whole-gland salvage HIFU reported in a study by Berge et al.  conducted with a comparable patient cohort, clinical characteristics and follow-up.
Periprostatic irradiation may result in tissue alteration that intensifies with time and increases the risk of late morbidity . In the present study, one patient had delayed pubic osteitis, possibly attributable in an endoscopic procedure performed 6 months after HSH. The bone infection healed after prolonged bladder drainage and antibiotics. Another patient with diabetes, had >2 months untreated urinary infection and urinary obstruction resulting in suprapubic pain. MRI revealed a pubovesical fistula occurring 23 months after HSH. The patient initially had been treated with transurethral bladder drainage and antibiotics. At 3 months MRI showed fistula healing. However, the subsequent MRI control after 6 months verified pubovesical fistula recurrence and the patient underwent cysto-prostatectomy and urinary deviation. Histopathological findings on whole-mount sectioned prostate gland were negative for cancer both in the HIFU-treated and untreated prostate lobe. Recently published results of pubovesical fistula in patients previously treated by RT show the low success rate after conservative treatment .
Only a few published studies have reported on focal salvage therapy outcomes. In a series of 39 patients treated with focal salvage HIFU with a median 17-month follow-up, Ahmed et al.  reported morbidity outcomes that were similar to the present study, with perfect continence in 64%, but a significant reduction in the IIEF-5 score. A small series of 19 patients treated with focal salvage cryoablation reported complications that included incontinence, urethral strictures or ulcer .
Encouraging functional outcomes after focal salvage cryotherapy have recently been described in a retrospective study conducted by Abreu at al ., where there was no observed urinary incontinence after treatment. However, contrary to the present patient cohort, the authors did not clarify if patients with extreme apex cancer localisation have been included in the study or if the peri-sphincteric part of the prostate was treated with radical intent.
The trend towards better outcomes after focal treatment, as reported in largest published studies on salvage HIFU and cryotherapy, is shown in Table 5 [6, 8, 10, 24, 25, 36, 42-44]. The major limitations of all published studies are their retrospective character and limited control about follow-up rate.
|Reference||Salvage treatment||Patients, n||Study type||Centres||Follow-up, months median (range) or mean (sd)||Follow-up rate, %||Incontinence requiring pad at last follow-up, %||Incontinence grade III, %||IPSS score||IIEF score||HRQL score||Rectal fistula %||Pubic bone osteitis, n (%)|
|Pisters et al. (2008) ||Wg cryo||279||R||multi||21.69 (24.9)||–||4.4||-||–||–||–||–||–||–||1.2||–|
|Murat et al. (2009) ||Wg HIFU||167||R||1||18.1 (3–121)||–||49.5||9.5||–||–||–||–||–||–||3||–|
|Williams et al. (2011) ||Wg cryo||176||R||1||90 (12–168)||94||40||3||–||–||–||–||–||–||2||–|
|Crouzet et al. (2012) ||Wg HIFU||290||P||1||48 (–)||–||50||9||–||–||–||–||–||–||2||8/290 (2.7)|
|Ahmed et al. (2012) ||Wg HIFU||84||R||2||19.8 (3–35.1)||–||38||–||8.3||11.6||8.6§||6.2§||100.2*||100.4*||2.4||1/84 (1.2)|
|Eisenberg and Shinoaha (2008) ||Focal cryo||19||R||1||18 (6–33)||79||5||0||–||–||–||–||–||–||0||–|
|Wenske et al. (2013) ||Focal cryo||44||R||1||37 (2–173)||–||0||0||–||–||–||–||–||–||5.5||–|
|Ahmed et al. (2012) ||Focal HIFU||39||R||1||17 (10–29)**||–||12.8||–||10||13||18‡||13‡||–||–||2.5||–|
|de Castro Abreu et al. (2013) ||Focal cryo||25||R||1||31 (4–90)||100||0||0||–||–||–||–||–||–||0||–|
|Present study||HSH||48||P||2||16.3 (10.5–23.5)**||100||25||8.3||7.1||8.6||11.2§||7.0§||35.7||36.8†||0||2/48 (4)|
It is known that the rate of post-treatment continence conservation depends on quantity of ablative energy applied to the apex region during treatment. Patients with apical tumour recurrence should be informed carefully of the risk of urinary leakage if radical salvage treatment is planned. Precise biopsy mapping of the apex using a 3D real-time navigation system and 3D-TRUS biopsy registration allow accurate apex staging. In patients with negative apex biopsy, apex-sparing salvage HIFU preserves urinary continence (Abstract ♯ 55 presented by Baco et al. at the 4th International Symposium on Focal Therapy and Imaging in Prostate and Kidney Cancer, 2011, Noordwijk, Amsterdam, The Netherlands). Hopefully, promising results of MRI in detection and localisation of prostate cancer will be confirmed in upcoming studies and that safety margins can be tailored to the tumour position . It is expected that the technological advance in therapeutic ultrasound, especially precise identification of the apex, adaptable height of HIFU lesions and perioperative temperature monitoring in treated and surrounding tissue will optimise the oncological and functional outcomes.
In conclusion, HSH in patients with unilateral radiorecurrent prostate cancer results in fewer and less severe morbidity than whole-gland salvage therapies, and may preserve pre-treatment health-related quality of life. Accurate imaging and biopsy are essential to identify malignancy suitable for focal therapy and to exclude metastatic disease. Based on the present results, prospective multicentre clinical trials with long-term follow-up are warranted.
Albert Gelet is investigator and consultant for Edap-TMS.
apparent diffusion coefficient
European Organisation for the Research and Treatment of Cancer
high-intensity focused ultrasound
International Index of Erectile Function-5
positron emission tomography
(external beam) radiotherapy
template-guided mapping biopsy of the prostate