VMAT technique enables concomitant radiotherapy of prostate cancer and pelvic bone metastases

Abstract

Background. Prostate cancer (PCa) patients with metastatic disease often suffer from skeletal pain and urinary retention impairing their quality of life. Prophylactic radiotherapy to bone metastases planned concomitantly with primary PCa radiotherapy could enable more precise control of combined dose in healthy tissues when compared to sequential palliative treatment.

Materials and methods. Volumetric modulated arc therapy (VMAT) was planned for 14 PCa patients with primary bone metastases. The bone planning target volume (PTV_bone) was contoured together with the PTVs of prostate (pr), pelvic lymph nodes (ln) and seminal vesicles (sv). Another virtual plan was calculated excluding PTV_bone for dose volume histogram (DVH) comparison. DVHs were additionally compared to a set of actual VMAT treatment plans of a control cohort of 13 high risk PCa patients treated with PTV_pr, PTV_sv and PTV_ln. The prescribed doses varied between 42 and 76 Gy for PTV_bone.

Results. Recommended healthy tissue tolerances (Quantec) were not exceeded except for one patient's rectum V_50Gy value. Rectum doses did not increase significantly due to the inclusion of PTV_bone. For bladder, there was a slight increase for V_65Gy and V_50Gy (2.7% and 7.4%). The DVHs of metastatic and non-metastatic patients were comparable. There were no differences in the PTV_pr DVH parameters, while mean PTV_ln dose increased by 3.7 Gy–4.4 Gy due to the increased treatment volume related to PTV_bone. All side effects were < grade 3 during the mean follow-up duration of 25 months.

Conclusions. VMAT offers a good optimization tool for adding extra PTVs to the radiotherapy plan. Radiotherapy of bone metastases concomitantly with irradiation of the primary prostate tumor is a safe and well-tolerated approach and deserves to be studied in a randomized setting.

For patients with metastatic prostate cancer (PCa) the median survival has conventionally been around 30 months when treated with hormonal therapy [1]. The increasingly potent new systemic therapy options for metastatic PCa are now prolonging survival not by months but by years [2,3]. However, use of radiation therapy (RT) in the metastatic setting has been preserved almost exclusively for palliation of pain. Obstructive uropathy is the key urological complication of progressive PCa and associated with poor prognosis [4]. RT is known to improve obstructive symptoms in up to 80% of patients [5]. Besides the many achievements in systemic therapies also RT techniques have improved by the introduction of intensity-modulated radiation therapy (IMRT) and most recently volumetric modulated arc therapy (VMAT) [6]. VMAT has been shown to improve RT of PCa by facilitating delivery of more conformal doses to planning target volume (PTV) and by reducing the dose of organs at risk [6–8]. Moreover, VMAT has been proven to be an efficient tool when treating multiple metastases and complex disease [9–11], and even superior to Cyberknife in hypofractionated PCa treatments [12].

Maximal local control with RT appears to prolong survival of PCa patients with lymph node metastases [13]. If one takes advantage of the treatment optimization and image guidance possibilities enabled by the present VMAT techniques in combination with positron emission tomography (PET) based delineation and gold seed implants [14], it is possible to irradiate distant metastases simultaneously with the primary locoregional PCa. Giving a radical dose to the prostate will most probably also give the best palliation, if urinary obstruction can be prevented.

The aim of the present analysis was to evaluate the feasibility of concomitant irradiation of locoregional prostate cancer and bone metastases, which was utilized for 14 patients, using VMAT technique. To achieve this we retrospectively compared the RT dose-volume data of treated PCa patients with PTV_bone to retrospectively recalculated plans of the same patients, where bone targets had been subtracted. We also compared their DVH data and side effects to a separate conventionally treated control cohort with locoregional disease.

Material and methods

Patients

Fourteen consecutive patients treated according to normal clinical practice at the Docrates Cancer Center (DCC, Helsinki, Finland) with prostate cancer and pelvic bone metastases were included in this analysis. (Patient demographics are presented in Supplementary Table I, to be found at online http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.962665). Patients went through multiparametric imaging at DCC. MR imaging was done using endorectal coil to ensure optimal image quality with 1.5 T scanner (Siemens Magnetom Espree, Erlangen, Germany). Magnetic resonance (MR) spectroscopy to evaluate choline and citrate peaks, dynamic contrast-enhanced subtraction imaging, diffusion-weighted imaging and multiple anatomical sequences were used for diagnosis, and to evaluate local extension of the disease. Whole body computed tomography (CT) imaging (Siemens Sensation Open) or PET/CT with ¹⁸F-choline or Na-¹⁸F was done (Siemens Biograph 6) for screening of distant metastases.

This analysis was performed according to the principles of the Declaration of Helsinki. All 14 patients were treated between August 2010 and November 2012. For every treated patient two RT plans were analyzed: the actual plan used for treatment (Plan 1) included pelvic bone metastases (PTV_bone), prostate (PTV_pr), seminal vesicles (PTV_sv) and regional pelvic lymph nodes (PTV_ln), while in the second retrospectively calculated virtual plan bone metastases were omitted (Plan 2). Our criteria for acceptance of the final actual radiotherapy plan included compliance with the Quantec or Radiation Therapy Oncology Group (RTOG) Concensus recommendations [15–18].

In order to assess whether the PTV_ln,sv,prost differed significantly from those of routine prostate cancer treatment cohort, and to better evaluate the influence of adding PTV_bone on healthy tissue doses, we analyzed the radiotherapy plans of matched 13 high risk prostate cancer patients treated with curative intent at DCC (Plan 3) without metastases. In our analysis, dose volume histogram (DVH) data of three different radiotherapy treatment plans were compared. (Dose prescriptions for each group are presented in Supplementary Table II, available online at http://www.informahealthcare.com).

Contouring

For DVH analysis, rectum was segmented from above the anal verge to the turn of the sigmoid colon, including the rectal contents as recommended by Quantec [15]. Bladder was contoured based on a single CT-simulation image. Patients were instructed to empty the bladder 1/2 h before imaging and drink 4 dl water while waiting. Similar instructions were complied with every treatment fraction aiming to minimize bladder volume variation during the radiation course. Both bladder and rectal walls were extracted as 3 mm interior contour for DVH comparisons. Femur head contours were based on automatic segmentation by Eclipse v.8 (Varian, Palo Alto, CA, USA). However, when bone metastases overlapped with femoral heads, only the contralateral contour was used. Two patients with bilateral acetabulum metastases overlapping femoral heads were excluded from the femoral head-summary.

All patients had an additional MR-imaging (without endorectal coil) in succession with CT-simulation (within 1 h). Images were coregistered based on the fiducial markers implanted in local anesthesia about one week prior to simulation. PTV_pr was contoured based on the MR (T2 weighted)-images and clinical target volume (CTV) with 5 mm margin. PTV_sv was in most cases contoured according to the CT images (CTV) with 5 mm margin. PTV_ln was based on the CT images and was contoured according to the RTOG consensus [18] with 5 mm margin. Bone metastases were mostly contoured based on the PET images, which were coregistered with dose planning CT and MRI. PTV_bone had 5 mm margin over GTV_bone. Number of pelvic bone metastases ranged from 1 to 8 (median 2).

Overlapping contours were subtracted for reporting, and for dose optimization purposes, with 5 mm margins. For example PTV_pr and PTV_bone were subtracted from PTV_ln. Thus, PTV_ln of Plan 1 differs from Plan 2 since group 2 was recalculated without PTV_bone.

Radiation therapy

Before RT 12 patients received both LHRH agonist (luteinizing hormone-releasing hormone) and antiandrogen therapy, one patient received LHRH agonist alone and one patient received antiandrogen alone. In addition, seven patients were treated with zoledronic acid or denosumab and one patient had been treated with docetaxel and Samarium therapy. Eight patients had received RT to mammary glands as prophylaxis for the side effects of bicalutamid and one patient had been given radiotherapy to a metastasis in cervical vertebra.

All patients were treated with RapidArc™ (Varian) VMAT technique. For treatment planning, templates were used for equivalent optimization starting point, but all plans were finalized individually. Two full arcs were typically used, but for some patients with multiple bone metastases a complementary arc was introduced using the same isocenter. Treatment planning was accomplished with Varian Eclipse vs. 8.6 and 10.0, with AAA dose calculation algorithm.

Three RT plans that were created in the beginning of the RT course were typically used for the treatment. RT began with large fields including PTV_bone, PTV_pr, PTV_sv and PTV_ln. The second plan typically included PTV_bone and PTV_pr, while the third plan included PTV_pr alone. As presented in Supplementary Table II to be found at online http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.962665, the total dose to prostate ranged between 76 and 80 Gy with 2 Gy/day fractionation. PTV_sv was also fractionated 2 Gy/day, while PTV_ln simultaneously received 1.8 Gy/day. The fractionation for bone metastases varied. Typically the metastases were fractionated equally to the lymph nodes during the first five weeks (25*1.8 Gy) and then boosted to higher doses with variable fractionation. Daily dose ranged between 1.8 Gy and 3.5 Gy, while EQD_2Gy (α/β 1.5 Gy) for bone metastases ranged from 42 Gy to 76 Gy depending on the number and site of metastases. The prescribed PTV_bone doses increased with time as more experience on the concurrent treatment of PCa metastases was obtained. Treatment plans were normalized according to the PTV_pr prescription.

All patients were treated with Varian IX 2300 linear accelerator equipped with on-board imaging (OBI). Orthogonal x-ray imaging was used daily to match prostate fiducial markers with the digitally reconstructed radiography (DRR) image. Cone beam CT (CBCT) imaging was used in first three fractions to check bladder and rectum filling, and its influence to the PTVs. If there was difference between the bones and the markers, the PTV_bone margin was increased accordingly based on the daily prostate movement.

Data analysis

A sum dose plan which included all treated volumes was created for every patient. For virtual plans (Plan 2) PTV_ln was re-contoured excluding the overlapping bone metastasis target, and the plan was recalculated (without PTV_bone). DVH data from sum plans were exported and re-read with in-house Matlab program to provide dose-volume parameters separately for each patient and to calculate mean DVH curves for separate groups. To estimate possible statistical difference between plans in this analysis the two-tailed Student's t-test was used. Paired test was used between the Plans 1 and 2, whereas two-sample unequal variance test was used for Plan 3.

Side effect follow-up

Common toxicity criteria (CTCAE, v. 4.03, 2010) were used to evaluate the possible side effects of RT. Side effects were registered before and during radiotherapy, and annually at the oncologist's follow-up appointments. The mean follow-up time at data cut-off in March 2014 was 25 months for the metastatic group and 31 months for the matched control cohort.

Results

The mean DVH comparison between the plans including bone metastases (Plan 1) and the recalculated plans (without PTV_bone, Plan 2) are presented in Figure 1. The PTV_pr plot was equal for both plans. PTV_bone is only presented for Plan 1. The DVH for PTV_bone was non-smooth due to the variable fractionation and due to the dose distribution from prostate booster plans. The PTV_ln plot was steeper and the mean dose lower for Plan 2. All healthy tissue DVHs were non-significantly elevated for Plan 1.

Figure 1. DVH comparison between plan 1 with the bone metastasis group (solid line) and the recalculated plan 2 group without bone metastasis (dashed line). For clarity, bladder wall, rectal wall and PTV_sv were excluded. PTV_bone DVH (purple line) is only presented for plan 1.

In Plan 1 with bone metastasis, D_mean corresponded exactly to the prescribed dose of prostate and the plan was homogenous (V₉₅ = 99.7%, V_{1cc, max} = 103%) as presented in Table I. There was no significant difference in DVH parameters of PTV_pr between Plan 1 and Plan 2, whereas mean PTV_ln dose was higher in Plan 1 due to irradiation of bone targets. V₉₅ was slightly worse in recalculated Plan 2 compared to the actual treatment Plan 1.

Table I. The DVH data (mean, range) for prostate and lymph node treatment planning volumes (PTV_prost, PTV_ln) for the different plans (described in Material and methods).

	Plan 1	Plan 2	Plan 3
PTVpr
Volume (ml)	100 (61–142)	100 (61–142)	85 (53–224)
Dprescr (Gy)	78.6 (78–80)	78.6 (78–80)	77.8 (76–78)
Dmean (Gy)	78.6 (78.0–80.4)	78.5 (78.0–80.3)	77.8* (75.9–78.2)
V95% (%)	99.7 (99.1–100.0)	99.5 (98.7–99.9)	99.2 (96.8–99.9)
V1cc,max (%)	103 (102–106)	103 (102–104)	103 (102–105)
PTVln
Volume (ml)	760 (623–985)	849* (658–1024)	744 (505–1101)
Dmean (Gy)	53.4 (47.6–61.1)	49.7* (47.4–53.9)	49.0** (46.0–53.4)
V95% (%)	99.5 (95.9–100.0)	99.0* (99.5–99.7)	99.2 (96.5–100.0)

*Difference statistically relevant when compared to group 1 (0.01  p  0.05); **Difference statistically relevant when compared to group 1 (p  0.01).

For rectum, one patient exceeded slightly the Quantec recommendation for V_50Gy, while all other values were well below the limits (Table II). The small (< 3.4%) violation for single DVH parameter was considered acceptable when the final treatment plan was chosen. There was no statistically significant increase in rectum doses except for the single value of V_75Gy (0.4% difference). All bladder parameters were well below the Quantec limits (Table II). However, there was slight increase in V_65Gy and V_50Gy values (2.7% and 7.4% differences) in Plan 1 compared to Plan 2. Even though there are no recommended dose limits for rectal and bladder walls (organ excluding contents) by Quantec, their DVH parameters were also calculated to find out possible differences. However, the results were quite similar in comparison to full organs; all reported doses were slightly higher for Plan 1 when compared to Plan 2. Only one dose-volume point was statistically significantly higher for the rectal wall (V₇₅) whereas three points were significantly increased for the bladder wall (V₇₀, V₆₅, V₅₀). For femoral heads, there was no significant difference between the plans and all mean values were below recommendations (Supplementary Table III, to be found at online http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.962665).

Table II. The DVH parameters (mean, range) for rectum and bladder compared to the Quantec recommendations [15–17].

	Plan 1	Plan 2	Plan 3	Quantec
Rectum
Volume (ml)	74 (48–109)	74 (48–109)	87 (48–183)	NA
V75Gy (%)	4.1 (1.6–12.5)	3.7* (1.3–11.8)	2.8 (0.0–6.4)	< 15
V70Gy (%)	7.4 (3.8–18.3)	7.1 (3.3–17.7)	5.8 (0.2–10.8)	< 20
V65Gy (%)	10.8 (6.3–23.4)	10.3 (5.6–22.5)	9.0 (1.3–15.5)	< 25
V60Gy (%)	15.6 (9.1–28.7)	13.9 (8.2–27.1)	12.9 (3.2–21.4)	< 35
V50Gy (%)	27.3 (16.1–53.4)***	23.1 (14.5–37.0)	23.3 (10.6–38.9)	< 50
Rectum wall
Volume (ml)	31 (24–40)	31 (24–40)	41 (26–62)	NA
V75Gy (%)	6.5 (2.9–17.6)	6.0* (2.3–17.3)	4.8 (1.4–10.3)	NA
V70Gy (%)	10.4 (6.9–22.0)	10.0 (6.1–21.7)	9.4 (3.9–15.6)	NA
V65Gy (%)	14.1 (10.4–25.5)	13.2 (9.5–25.0)	13.0 (6.3–20.4)	NA
V60Gy (%)	18.9 (13.6–38.0)	16.5 (12.6–28.1)	17.0 (9.5–27.5)	NA
V50Gy (%)	29.6 (19.9–61.3)	24.1 (18.5–35.9)	26.1 (16.1–40.7)	NA
Bladder
Volume (ml)	297 (153–595)	297 (153–595)	198 (81–691)	NA
V75Gy (%)	0.3 (0.0–1.4)	0.2 (0.0–1.5)	0.1 (0.0–0.7)	< 15
V70Gy (%)	3.5 (1.6–6.3)	2.8 (0.0–5.5)	4.0 (0.7–21.2)	< 25
V65Gy (%)	5.8 (2.7–11.1)	4.3 (0.0–8.2)	6.4 (1.7–28.7)	< 35
V60Gy (%)	8.7 (3.8–17.7)	6.0* (0.2–11.1)	8.9 (2.6–35.9)	< 50
V50Gy (%)	23.3 (9.2–43.8)	15.9** (7.8–26.9)	21.9 (8.9–71.8)	NA
Bladder wall
Volume (ml)	63 (41–103)	63 (41–103)	62 (36–166)	NA
V75Gy (%)	1.1 (0.0–6.4)	0.7 (0.0–6.7)	0.3 (0.0–1.9)	NA
V70Gy (%)	9.5 (5.8–15.9)	8.0 (0.0–14.4)	7.6 (1.5–21.2)	NA
V65Gy (%)	13.8 (9.0–21.6)	10.7* (0.0–19.9)	11.1 (3.5–27.1)	NA
V60Gy (%)	18.5 (11.9–31.5)	13.5* (1.0–25.3)	14.6 (6.0–32.6)	NA
V50Gy (%)	37.5 (20.9–62.5)	27.7** (19.1–42.0)	30.4 (17.1–68.5)	NA

*Difference statistically relevant when compared to group 1 (0.01  p  0.05); **Difference statistically relevant when compared to group 1 (p  0.01); ***Exceeds Quantec recommendations.

As presented in Supplementary Table IV (to be found at online http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.962665), 57% of Plan 1 patients had grade 1–2 GU disorders prior to start of RT. One patient had grade 2 urinary retention before RT, but at annual follow-up visits it had become better (grade 1). One patient suffered from prolonged grade 2 incontinence after the RT. None of the patients had ≥ grade 3 GU disorders during or after the treatments. Two years after the RT 50% of the patients had grade 1–2 GU disorders, which was 7% less than before the RT. None of the patients reported > grade 1 GI disorders during or after the RT. None of the patients reported any femoral side effects. The side effect profiles were comparable with the matched control cohort though the state of the Plan 3 patients prior to RT was better (23% Gr 1–2, 8% Gr 3 before RT). None of the Plan 3 patients had Gr 2–3 toxicities due to the RT.

Discussion

To the best of our knowledge, this is the first study where the feasibility of concomitant VMAT of PCa and pelvic bone metastases has been analyzed. According to the analysis, the treatment was well tolerated and the healthy tissue tolerances were not exceeded while there was no deterioration in the PTV_pr DVH parameters. Yoo et al. compared IMRT and RapidArc™ planning for high risk PCa RT including prostate, seminal vesicles and regional lymph nodes [19]. Though we included an additional bone PTV, our bladder (V_65Gy) and rectum doses (V_70Gy) were 50% and 30% lower than the corresponding doses in their study while our maximum PTV_pr point dose was 5% lower indicating that we had superior PTV_pr homogeneity. Both Aznar and Pesce [7,8] reported DVH data for prostate Rapidarc™ RT including seminal vesicles. Our rectum and bladder doses were 14–31% lower and PTV_pr coverage superior (V_95% 1% higher) when compared to Pesce et al. Our bladder doses were also lower (5–50%), but rectum doses higher (33–43%) when compared to Aznar et al., while our PTV coverage was better (V_95% 5% higher).

Two different plans (1 and 2) for the same patient, same body anatomy, same contours and dose prescriptions enabled straightforward analysis about the influence of additional bone PTV to other target volumes and healthy tissues. Slight increase on rectum (only V₇₅) and bladder (only V₅₀ and V₆₅) DVH parameters was noticed when the bone targets were included. However, all the mean values were much lower than the Quantec [15–18] recommendations indicating low probability for side effects. Rectum-V₅₀ recommendations were exceeded in one patient, but metastasis PTVs (bone, lymph node) were large and alongside with the organ, making the small deviation acceptable. Moreover, patient had only temporary grade 1 proctitis.

Comparison in a non-randomized setting between actual treatment plans of metastatic (Plan 1) and non-metastatic (Plan 3) patients treated at DCC was also done, but it must be acknowledged that the number of treated patients was small and individual variations (volume, dose prescription) without strict protocol might have had an effect on results. However, the only statistical difference in the size of the contours was noticed for femoral heads (p < 0.05), which was due to the extraction of the overlapping PTV_bone for Plan 1. When actual treatment plans (Plan 1) were compared to the actual treatment plans of matched control cohort plans (Plan 3), no statistical difference was noticed between the doses to organs at risk, which indicates that patient related differences are larger than differences caused by treatment planning methods. For one control cohort patient, bladder doses were high (still below Quantec limits), which was due to the unusually small bladder in the CT simulation image. However, that patient had no genitourinary side effects during follow-up.

PTV_pr was the primary target volume in patient dose optimization and the objectives were met. Thus PTV parameters were very homogenous and corresponded exactly to the prescribed dose in all plans. However, there was 12% difference (p < 0.05) in PTV_ln contours in actual and virtual plans, because slightly overlapping PTV_bone was extracted with 5 mm margin from PTV_ln for recalculation. The maximum dose parameter for PTV_ln was not found applicable due to the dose gradient from PTV_pr -booster plans. The mean doses for PTV_ln were always higher than the prescribed doses because of the prostate booster plans’ low dose tail. Treating the bone metastases concomitantly (Plan 1) increased significantly (p < 0.05) the mean lymph node doses (4 Gy). The increase was due to the four patients (PTV_ln mean 58–61 Gy) with multiple (7–8) or large bone metastases with high prescribed dose.

PTV parameters for bone metastases were not shown because treatments were personalized and consequently there was a large variation between the prescriptions. However, the mean dose for bone metastases corresponded accurately to the prescribed doses and the dose was homogenous according to the International Commission on Radiation Units & Measurements’ recommendations (ICRU 62) when the dose gradient from prostate booster plans was not accounted for. We started with a low PTV_bone dose prescription (EQD_2Gy = 42–50 Gy) for five patients, but then increased the dose as we did not notice remarkable deterioration on the healthy tissue DVHs. For an experienced physicist or dosimetrist involved in RapidArc™ treatment planning, introducing an extra bone PTV with 70–76 Gy dose prescription should not be too troublesome – we are now prescribing 76 Gy for the PTV_bone in most of the cases. A dose distribution for concomitant bone metastasis and prostate cancer treatment is presented in Figure 2a–f.

Figure 2. The axial (a, c, e) and coronal (b, d, f) dose distributions of 63-year-old prostate cancer patient with CT+ MR detected metastases in left ilium and acetabulum, and in right parailiac lymph nodes. Pre-treatment PSA was 250 μg/l and Gleason score 9. Patient dose prescription was 45 Gy in 25 fractions (fr) to regional lymph nodes, 50 Gy/25 fr to seminal vesicles and 78 Gy/39 fr to prostate. Bone metastases were treated to 73Gy (45 Gy/25 fr + 28 Gy/14 fr). Dose map ranged from 45 Gy to 78 Gy (a, b), from 69.3 Gy to 78 Gy (c, d) and from 70.1 Gy to 78 Gy (e, f). 1.5 years after the radiation therapy, the PSA was < 0.1 μg/l, the bone metastases were sclerotic, and the lymph node metastasis had shrunk.

None of the Plan 1 patients reported > grade 1 gastrointestinal disorders during or after the RT. Before RT 57% of the patients reported weak (≤ grade 2) genitourinary disorders. The percentage increased to 79% at the end of the RT but decreased again to 71% and 50% at the annual follow-ups. The level of toxicity was comparable between Plans 1 and 3 patients though much larger patient cohorts would be needed to investigate possible differences with statistical relevance. The side effects seemed to be in agreement with previously reported series [20].

Two patients had grade 2 GU disorders (nocturia, pollacisuria, presence of urinary urgency) during the treatment, which normalized 7–10 months after the RT course. One patient suffered from prolonged grade 2 incontinence. When patient's plan was studied, the reported bladder DVH parameters were well below average. However, the low dose volume was unusually high for the patient; V_30Gy was 96% whereas the average V_30Gy of the analyzed group was 63%, which might have caused the problems. Also, the patient was elderly (85 years old), which might have an effect though he was otherwise in good performance status (WHO 0–1). This patient was also using diuretics. One patient had urinary retention already before and during radiotherapy caused by cancer growth in the prostate, but the suprapubic catheter could be removed 10 months after the end of radiotherapy as a sign of excellent local response to treatment.

This is a single institute retrospective non-randomized analysis of the feasibility of adding extra PTVs to traditional RT volumes of prostate, seminal vesicles and pelvic lymph nodes specifically in prostate cancer which has been diagnosed with primary bone metastases. Our preliminary experiences suggest that choline PET could be a clinically useful tool when the aim is to delineate active targets most precisely [21]. Based on our calculations, it appears that VMAT RapidArc™ technique offers a good optimization tool in order to add extra volumes to the radiotherapy plan. Although the number of treated patients is quite limited so far, our experience indicates that radiotherapy of bone metastases concomitantly with irradiation of the prostate is a well-tolerated approach and deserves to be studied in randomized setting.

Supplementary material available online

Supplementary Tables I–IV, to be found at online http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.962665.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

T. Joensuu is a founder of DCC. A. Kangasmäki and T. Joensuu are Shareholders of DCC.