To compare the quality of permanent prostate brachytherapy (PPB) implants, dosimetric outcomes and urinary morbidity between patients with large (>50 mL) and those with smaller prostates, treated with a dynamic dose-feedback technique as monotherapy for localized prostate cancer.
PATIENTS AND METHODS
The series included patients with pre-existing bladder outlet obstruction managed with planned transurethral resection or incision of the prostate; 155 consecutive men had PPB implants as monotherapy for localized prostate cancer using a dynamic dose-feedback approach. Dosimetric variables assessed included the implant volume, the minimum dose to 90% of the prostate (D90), and the volumes of prostate receiving 100% and 150% of the prescribed dose as a percentage of the total volume (V100 and V150), during and after implantation. Urinary morbidity was recorded in terms of acute urinary retention (AUR), the need for surgical intervention after implantation and the American Urologic Association (AUA) symptom score at baseline, 1.5, 3, 6, 9, 12 and 18 months.
In all, 38 patients had prostate volumes of ≥50 mL; prostate volume had no influence on any dosimetric variable assessed. Two patients with large prostates (≥50 mL) had AUR and required delayed surgery. Three patients with small prostates (<50 mL) had transient retention; the differences were not statistically significant (Fisher’s exact test). AUA symptom scores peaked at 6 weeks and returned to baseline within a year; there were no statistically significant differences between the groups. Eight patients had planned transurethral surgery at ≥4 months before implantation; they all had D90s of >130 Gy and had no incontinence.
Using the dynamic feedback technique, there was no adverse dosimetric and urinary morbidity in men having PPB and with prostates of >50 mL. Likewise, there were no impediments, e.g. pubic arch interference, which precluded a favourable dosimetric implant in men with a large prostate. Large prostates should not be a contraindication to PPB and require no hormonal cytoreduction. Patients with obstructive lower urinary tract symptoms can be managed with planned transurethral prostatic surgery before implantation, without compromising implant quality or morbidity.
the volume of prostate receiving 100% and 150%, of the prescribed dose, as a % of the total volume
acute urinary retention
For transperineal permanent prostate brachytherapy (PPB) as a treatment for localized prostate cancer, there are robust follow-up data beyond 10 years that show similar biochemical control rates to the traditional methods of radical surgery and external beam radiotherapy [1–3]. The availability of PPB, particularly outside the USA (where it is already well established) is increasing rapidly.
Despite several reports to the contrary, there are widespread conceptions that a large prostate volume is a contraindication to PPB due to technical difficulties in achieving dosimetric coverage and increased urinary morbidity [4–6]; this is reflected in referral patterns and guidelines that recommend hormonal cytoreduction [7–10]. However, androgen suppression might compound urinary morbidity and have an associated side-effect profile . Similarly, patients with pre-existing BOO might be denied PPB; the place of planned surgery before PPB has not been established.
The purpose of the present study was to compare the implant quality (in terms of dosimetric outcomes) and urinary morbidity between patients with large prostates (>50 mL) and those with smaller prostates, all treated with a dynamic dose-feedback technique as monotherapy for localized prostate cancer. This series included patients with pre-existing BOO managed with planned TURP or incision of the prostate.
PATIENTS AND METHODS
Between December 2003 and July 2006, 155 consecutive patients had PPB implants as monotherapy for localized prostate cancer; their clinical characteristics are shown in Table 1. A large prostate was not considered a contraindication to implantation and no patients had hormonal manipulation before implantation. Patients with raised AUA symptom scores (>10) had uroflowmetry and cystoscopy. If symptoms were deemed to be a result of correctable BOO (maximum flow rate <10 mL/s, high residual volume, tight bladder neck or obstructive prostate) then the patient was offered prophylactic surgery (bladder neck incision or TURP). The surgical intervention was minimal, to relieve the obstruction and maintain enough tissue to implant at a later date.
Table 1. The clinical characteristics of the 155 patients
The dynamic dose-feedback technique, as described by Potters et al., was applied, using loose 125I sources to a prescribed dose of 145 Gy. Implants were placed by a team comprising a urological surgeon, clinical oncologist, medical physicists and theatre staff. No pubic arch or planning studies were used before implantation. At one visit, each patient was placed in an extended dorsal lithotomy position under general or spinal anaesthetic. This position minimized the impact of pubic arch interference. Patients who had had previous transurethral surgery had cysto-urethroscopy to better understand the prostatic anatomy. Under TRUS guidance and with the urethra defined by aerosolised gel, needles were placed transperineally into the prostate ≈1 cm apart and with a peripheral : central ratio of 3 : 1. If pubic arch interference was apparent, then the patient’s position was easily adjusted during needle insertion and the needles re-implanted. Each needle position was recorded onto computer software in real time. Once needle coverage was satisfactory, transverse images of the prostate were captured by the software at 5-mm increments from base to apex. The prostate, urethra and rectum were outlined and the computer generated an inverse plan, guided by pre-programmed dose rules (Table 2); these rules aimed to keep urethral and rectal doses to a minimum whilst providing coverage to the entire prostate. Once the plan was approved, with or with no manual optimization, loose sources were inserted under TRUS guidance in the longitudinal plane. Each source’s ‘dropped’ position was entered into the computer software in real time so that a dynamic plan was continually updated throughout the implant; if it became suboptimal then the plan could be reiterated, based on existing seed and needle positions. At the end of the procedure, the dynamic dosimetry was approved before the patient left the operating theatre.
Table 2. Dose rules for the computer-generated inverse plan
VNorgan, the volume of the organ that receives N% of the prescribed dose.
Prostate planning target volume has 2-mm margin
The initial 40 patients had urethral catheters placed at the end of the procedure and removed when they were mobile after recovering from the anaesthetic. The protocol subsequently changed and in the subsequent patients, urethral catheters were not left in situ. Patients were typically discharged on the same day, once voiding satisfactorily.
A CT-based dosimetric assessment of the implant took place 4 weeks after surgery. All patients were prescribed the α-blocker tamsulosin 400 µg once daily for a week before the implant and afterward until the symptom scores returned to baseline.
Dosimetric outcomes recorded during surgery included prostate volume, number of seeds implanted, total activity implanted, the minimum dose to 90% of the prostate (D90), and the volume of prostate receiving 100% and 150%, respectively, of the prescribed dose, expressed as a percentage of the total volume (V100 and V150), both during (dynamic TRUS-based) and after (CT-based) implantation. Dosimetric calculations were based on the recommendations of the American Association of Physicists in Medicine Task Group 43 [13,14]. Variseed 7.1 software (Varian Medical Systems, Charlottesville, VA, USA) was used for all dosimetric studies.
In patients who had had prophylactic surgery the prostate was outlined using the method described by Moran et al., whereby the urethral defect was not included in the prostate volume (Fig. 1). A satisfactory implant was defined as one that had a D90 of >130 Gy .
Urinary morbidity data included rates of acute urinary retention (AUR), the need for catheterization and subjective AUA symptom scores (scale of 0–35, a higher score corresponds with worse symptoms) at baseline, 1.5, 3, 6, 9, 12, and 18 months.
The results were analysed statistically using the Mann–Whitney U-test, Pearson’s correlation coefficient and Fisher’s exact test, with statistical significance indicated at P < 0.05.
Of the 155 patients who presented for PPB, 38 had prostates of >50 mL (Table 3); irrespective of prostate size the D90 during surgery was >134 Gy in all patients and no patient was found to be un-implantable ‘on the table’. The CT-based dosimetric analyses after surgery were available for 125 patients (96 and 29 with small and large prostates, respectively); one (4%) with a large prostate had a D90 of <134 Gy, vs three (4%) with a small prostate, although all four had a D90 of >123 Gy. On detailed review, the suboptimal areas were anterior and at the base in all cases. There was no difference between the groups in D90, V100 or V150 either during or after surgery (Table 3), and no significant correlation between prostate size and these dosimetric values (Table 4).
Table 3. Dosimetric outcomes for all patients and those with large and small prostates
Large (≥50 mL, 38)
P (Mann–Whitney test)
Mean (median, range):
40.8 (36.3, 14.5–91.3)
62.0 (60.7, 50–91.3)
33.9 (33.0, 14.5–49.5)
Seed activity, MBq
15.8 (15.5, 14.8–17.8)
15.9 (15.5, 14.8–17.8)
15.8 (15.5, 14.2–17.1)
Number of seeds
71.2 (68, 35–110)
92.9 (91.5, 79–123)
64.1 (63, 35–88)
Total activity, MBq
1117 (1058, 555–1931)
1476 (1436, 1228-1931)
1006 (995, 555–1369)
Activity/mL prostate, MBq
28.5 (28.9, 19.2–41.8)
24.1 (24.1, 19.2–28.1)
30.3 (30.0, 23.2–41.8)
Table 4. Correlation coefficients of prostate size vs dosimetric variables
There was a strong association between prostate size and the amount of activity required to achieve satisfactory dosimetric coverage. There was also a strong negative correlation between prostate size and the amount of activity implanted per unit volume of the prostates, with the largest requiring about half the amount of activity/mL than the smallest (Tables 3 and 4).
Four patients had a planned bladder neck incision and four had TUR(P) before implantation, with a mean (range) resection weight of 4 (1.5–10) g. At least 4 months elapsed between surgery and implantation. A further two patients had had TUR(P) several years before implantation; these had been carefully evaluated for their suitability. During these implants, care was taken to clearly visualize the urethral defects and to minimize radiation levels to these areas, both during the planning of the implant and whilst placing sources (guided by real-time TRUS and urethroscopy before implantation). All of these patients had satisfactory dosimetric outcomes.
Five patients (3%) required catheterization within 72 h of the implantation: of these, two passed voiding trials within a week with no further need for intervention (prostates of 34.1 and 44.1 mL); one passed a voiding trial but required intermittent self-catheterization for symptoms over a year (prostate 32 mL); one failed to void spontaneously and was managed with a suprapubic catheter for 9 months before undergoing TURP (prostate 68.5 mL); finally, one passed a voiding trial within a week but had a bladder neck incision for symptoms at 1 year after implantation (prostate 62.5 mL). Neither of the patients who required surgery had subsequent incontinence. There was no significant difference between those with a large (>50 mL) or small prostate in the numbers who had AUR or required surgery (Fisher’s exact test P = 0.60 and 0.06, respectively). None of the patients who had previous surgery required catheterization.
There was no significant difference in AUA symptom scores between the groups at any time assessed. In both groups, symptom scores were at their highest 6 weeks after implantation, and had returned to baseline by a year (Fig. 2).
In this series of prostate glands treated with a dynamic dose-feedback method of PPB, prostate size had no influence on implant quality or urinary morbidity. The D90 is the only dosimetric variables that has been shown to predict for biochemical failure-free survival after PPB implants [17,18]. We chose a threshold of 130 Gy (equivalent to 90% of the prescribed dose of 145 Gy) to determine implant quality, based on the study by Potters et al., that showed an 80% 4-year biochemical failure-free survival for those with a D90 of <130 Gy, compared to 92% for those implanted with a D90 of >130 Gy. Whilst the need for CT-based analysis of implants after placement is questioned in the face of the intraoperative evaluation , this presents the opportunity to make a detailed assessment of the implants, particularly important for the novice and after implanting the more technically challenging prostates. At present, this is our routine practice, although patient non-attendance, hip prostheses and delays in processing diminished the availability of postoperative data in the present study. The finding of under-dosed areas at the base and anterior is reassuring, as these are rarely areas of cancer in low-risk patients  and the D90 per se might be misleading in these cases . Nevertheless, in the present study, prostate size did not influence implant quality, which was satisfactory in the vast majority of large (and small) prostates implanted.
The implantation of large prostates using the pre-plan Seattle method was described in 33 patients [4,22]; 80% of the target volume was covered in all cases, although five had incomplete coverage of the anterior and lateral margins (the D90 values were not reported) and the AUR rate was high, at 36%.
The degree of pubic arch interference and patient suitability for PPB can be assessed by pubic arch studies before implantation . The extended dorsal lithotomy position opens the transperineal passage and is further improved by increased flexion at the hips [4,5]. An advantage of intraoperative techniques is that if the pubic arch is found to be obstructive to needles, the patient’s position can be altered without having to stop the procedure. By adjusting the position in this way, no implant was prevented from proceeding in the present series, and studies beforehand were not required. Similarly, using an intraoperative nomogram technique, Stock and Stone  showed satisfactory coverage in large prostates; the D90 was >140 Gy in 65 of 66 patients. In a more recent study of procedures carried out in 33 community hospitals using the same nomogram technique, larger prostates were associated with higher doses (a median D90 of 175 Gy for ≥40 mL, vs 149.5 Gy for <25 mL) . This relationship was not found in the present study, and is probably due to the dynamic dosimetric assessments and feedback during the procedure maintaining a check on high doses.
AUR is a recognized complication after PPB, with published rates of 5–36%[22,25]. Significant risk factors include: a large prostate volume, high symptom scores before implantation, a low peak urinary flow rate on uroflowmetry, androgen suppression, number of seeds and physician’s experience [11,25–28]. The current study showed a higher rate of AUR amongst those with a larger prostate (5%) than with a smaller prostate (3%), although this was not statistically significant. Similarly, surgical intervention after implantation was only required in two patients with larger prostates. It is possible that these were type 2 errors resulting from a small sample. Nevertheless, the rates of AUR in the present series compare favourably with those from established centres. Reasons for this might include the avoidance of androgen suppression, patient selection, prophylactic surgery for those at risk, and a judicious level of total activity implanted.
There is little evidence to suggest that the use of neoadjuvant hormonal therapy influences the biochemical outcome after PPB, particularly in low-risk patients [29,30]; its use is generally to reduce glands and thus make an implant technically more feasible . Crook et al., amongst others, reported that androgen suppression is a predictive factor for AUR independent of prostate size. Therefore, if pubic arch interference can be overcome by other means, the use of androgen suppression is best avoided. There might be a financial incentive to the consumer to reduce the prostate and thus reduce the number of sources required, but this is counterbalanced by the cost and morbidities of hormone manipulation. A disadvantage of this brachytherapy technique is that the precise number of sources required is unknown until the actual time of implant. Although an estimate can be gained from preoperative TRUS (carried out at the time of biopsies in our institution) there is an inevitable waste. Direct comparisons are not possible, but it appears that the amount of activity implanted using this approach appears to be considerably less than in others [5,31]; the corollary of this might be less urinary morbidity [11,32].
The larger glands required less radioactivity and therefore fewer seeds/mL to achieve satisfactory doses. In any given implant some activity is ‘lost’ to the surrounding tissues; the amount lost depends on the surface area of the prostate. As the volume increases, the ratio of surface area to volume decreases, hence proportionally less activity is ‘lost’ to the surrounding tissue for a large prostate.
Patient selection is an important factor in reducing morbidity after PPB . As described, patients had an assessment that involved urinary symptom scoring and uroflowmetry as a minimum. If these were unfavourable and the patient was either unwilling or unsuitable for radical prostatectomy, then planned prophylactic surgical intervention was undertaken. Historically, previous TUR(P) was a contraindication for PPB due to high rates of subsequent urinary morbidity, including incontinence . However, with modern peripheral loading techniques, implants have been satisfactory in terms of dosimetric coverage and morbidity, providing the TUR cavity is small (e.g. <25% of the total volume) and there is a reasonable margin of prostate tissue to implant (e.g. 1 cm laterally and posteriorly) [15,34]. When planning transurethral surgery before PPB it is important to maintain these values and to leave a sufficient interval to allow re-epithelialization of the urethra. Also, whilst outlining the prostate margin for purposes of dosimetry, it is important not to include the defect as part of the target volume, as this will give the impression of under-dosing the target (Fig. 1). In the present series of planned transurethral surgery (albeit small), the dosimetric and urinary outcomes were satisfactory and there were no cases of incontinence.
Although there were few patients at the later follow-up times, the urinary morbidity data showing a peak of symptoms at 6 weeks after implantation and then returning gradually to baseline by 1 year, is consistent with other reports, as is the finding that prostate volume was not a significant factor [6,26].
In conclusion, using a dynamic dose-feedback technique for PPB allowed the implantation of large prostates with similar dosimetric and urinary morbidity outcomes to those of smaller glands, with no need for neoadjuvant hormonal therapy. Planned and judicious transurethral surgery before PPB can result in satisfactory outcomes. Large prostates and previous transurethral surgery should not be considered contraindications to PPB.
The prostate brachytherapy programme at Guy’s & St Thomas’ NHS Foundation Trust was funded by the Guy’s & St Thomas’ Charity. We thank the following medical physicists and dosimetrists for their valuable input into this work: Dominic Withers, Angela Daynes, Charles Deehan, Keri Owen (Guy’s & St Thomas’ NHS Foundation Trust), Irena Blasiak-Wal and Kathy Cooke (Cromwell Hospital).
CONFLICT OF INTEREST
None declared. Source of funding: Guy’s & St Thomas’ Charity.