Curative treatment of patients with localized prostate cancer comprises radical prostatectomy or radiation therapy. In external beam radiation therapy, dose escalation is currently investigated to improve outcomes. Several studies provide strong evidence for a dose–response relation of local tumor control, biochemical progression-free survival, and progression-free survival [3, 10, 18, 20, 25, 28, 35, 37].

Formerly published long-term results from a randomized phase III dose escalation trial conducted at the M.D. Anderson Cancer Center demonstrated a significant benefit after dose escalation to 78 Gy in terms of improved freedom from biochemical and clinical progression [18, 28]. Dose-escalation trials using conformal three-dimensional (3D)-radiation therapy showed that the additional anti-tumor effectiveness is accompanied by an increased treatment-related morbidity, i.e., gastrointestinal and genitourinary toxicity [3, 25, 28].

Intensity-modulated radiation therapy (IMRT) might counteract normal tissue toxicity correlated with conventional dose escalation [11]. For example, a large IMRT-based prostate cancer dose-escalation study initiated at the Memorial Sloan-Kettering Cancer Center although not randomized reported a favorable toxicity profile in patients treated with IMRT as compared to those who had received 3D treatment, despite a further increase in the prescribed total dose [35, 36].

In principle, dose escalation can be achieved either by increasing the number of fractions at 1.8–2 Gy per fraction or by increasing the dose per fraction above 2 Gy (hypofractionation). The rationale for using increased doses per fraction is the assumed relatively low α/β ratio reported for prostate cancer. Due to this rather low α/β ratio of about 1–3 Gy [24, 29] prostate cancer cells are hypothesized to be especially susceptible to cell kill by hypofractionated radiotherapy [1]. As the α/β ratio for prostate cancer is also assumed to be lower than that for the rectal wall, hypofractionated radiation therapy should have the potential to improve the therapeutic gain and has consequently been adopted as a strategy to tackle prostate cancer [1, 7, 19]. This concept has been further extended by the simultaneous integrated boost (SIB) concept, where increased doses per fraction are selectively and simultaneously delivered to subvolumes of the target volume [14, 21, 32].

This report is on the acute toxicity and the dose–volume data of the first 40 patients treated at our department with helical tomotherapy using a moderately hypofractionated simultaneous integrated boost IMRT (SIB-IMRT) to a total dose of 76 Gy in 2.17 Gy per fraction applied to the prostate.

Patients and methods

Patients and treatment planning

Starting in February 2008, patients with intermediate risk, localized prostate cancer (cN0 cM0) were treated with SIB-IMRT at the tomotherapy unit in our department. Patients with intermediate risk prostate cancer were defined as (1) not having low-risk features (cT1, Gleason score < 7, and initial PSA  ≤ 10 ng/ml) and (2) not having a risk of ≥ 20% of lymph node metastasis according to the Roach formula [30]. In selected cases, patients were treated with SIB-IMRT, even if they did not fulfill the above criteria for intermediate risk prostate cancer.

A CT scan of the pelvis from the iliac crest to the ischias tuberosities was performed in 5 mm slice thickness for treatment planning. Furthermore, a MRI scan was carried out and fused with the planning CT to optimize the definition of the prostatic volume [13]. The target volumes and organs at risk (OAR) were contoured in iPlan (BrainLAB AG, Feldkirchen, Germany). The gross tumor volume (GTV) comprised the prostatic gland and base of the seminal vesicles. The margins for the clinical target volume (CTV) accounting for microscopic extracapsular tumor spread were 5 mm in all directions except for the rectal interface with no additional safety margin. The planning target volume (PTV1) encompassed the CTV with a safety margin of 3 mm in all directions except for the craniocaudal direction with margins of 5 mm (Fig. 1). The boost volume (PTV2) encompassed the prostatic gland only, with a safety margin of 3 mm in all directions except for the craniocaudal direction where it was 5 mm. The rectum (outer contour) was delineated from the anal verge to the start of the sigmoid colon. In addition, the following OARs were contoured: urinary bladder, femoral heads, sigmoid colon, and remainder of the bowel within 2–3 cm above the PTV1. A help structure (Rectum-76) containing the overlap of the PTV2 with the rectum and 3 mm anteriorly was created in order to limit the dose to this structure to ≤ 100% of the prescribed dose to PTV2. The prescribed dose was 70 Gy in 2 Gy per fraction to the PTV1 and 76 Gy in 2.17 Gy per fraction to the PTV2. The dose calculation was carried out with the inverse treatment planning system of Tomotherapy (Tomotherapy, Inc., Madison, WI, USA). The objective was to cover at least 95% of the PTV2 with 76 Gy (after the first 5 patients that were calculated to the median of the volume). The maximum dose should not exceed 107% of the prescribed dose. Assuming an α/β ratio of 3 or 1.5 for prostate cancer cells, the biologically 2-Gy equivalent dose for the prescribed dose of 76 Gy is 78.6 Gy3 or 79.7 Gy1.5, respectively [22].

Treatment planning contained no formal constraints for the remaining rectum and bladder doses, but high (volume receiving at least 60 Gy (V60) to volume receiving at least 76 Gy (V76)) and intermediate dose (volume receiving at least 35 Gy (V35) to volume receiving at least 59 Gy (V59)) rectal and bladder volumes were kept as low as possible by an iterative planning process.

Fig. 1
figure 1

a, b, c, d Examples for delineation of 70 Gy PTV1 (orange), 76 Gy PTV2 (red), and rectum with rectal balloon (violet), bladder (yellow) and femoral heads (green) on different slices of the planning CT in the craniocaudal direction

Full size image

Quality assurance

Treatment plans for all patients were checked through a plan quality assurance procedure prior to the first treatment. For that purpose, patient treatment plans were re-calculated for suitable phantoms using the tomotherapy planning software.

Treatment

All patients had neoadjuvant hormonal therapy 2–4 months before radiation therapy. Patients were immobilized for treatment in an individually shaped vacuum cushion. For immobilization of the prostate, an endorectal balloon was used.

After set-up, patients received a MV-CT prior to each treatment fraction. This daily image guidance using a MV-CT caused an additional dose of 1 cGy per CT scan, which was typically carried out from 2 cm above the PTV1 to 2 cm below the PTV1 in 6 mm slice thickness. After acquisition, the MV-CT was fused automatically to the planning CT scan. If necessary, this fusion was corrected manually to align the prostatic gland. After this correction, treatment time was approximately 4–5 min with a jaw of 2.5 cm and a pitch of 0.27.

Toxicity evaluation

Gastrointestinal (GI) and genitourinary (GU) symptoms were prospectively documented before, after 20 fractions, and at the end of radiotherapy. Toxicity was scored according to modified CTCAE version 3 criteria (Tab. 1).

Tab. 1 Toxicity score
Full size table

Results

Patient and treatment characteristics

Patient and treatment characteristics are shown in Tab. 2. All patients received the prescribed treatment, except for neoadjuvant hormonal therapy in 4 patients due to intolerance. Treatment planning data (Tab. 3, Fig. 2) assessed for each patient showed good rectal sparing. High (V60–V76) and intermediate dose (V35–V59) rectal volumes were kept low with a median value of the volume receiving more than 65 Gy (V65) of 13.5%. The respective values for the V70 and V75 were 9% and 3%. At the same time, very good dose coverage of PTV1 and PTV2 was achieved, with median doses of 73.7 Gy and 77.0 Gy respectively.

Tab. 2 Patients’ characteristics
Full size table
Tab. 3 Dose statistics of the rectum, bladder, planning target volume 1 (70 Gy PTV) and planning target volume 2 (76 Gy PTV) concerning the mean dose (D mean ), maximum dose (D max ) and volumes (%) irradiated with at least 40 Gy (V40), 50 Gy (V50), etc.
Full size table
Fig. 2
figure 2

Cumulative dose–volume histogram (DVH) using overall patients mean planning values and error bars for standard error of the means

Full size image

Acute toxicity

Incidence of baseline and maximum acute GI and GU symptoms during treatment are provided in Fig. 3. No grade IV GI or GU toxicity was observed. Grade III GU side effects as seen in Tab. 4 occurred in 20% of patients involving nocturia only and merely two of these eight patients had no baseline symptoms. Grade II GU toxicity was observed in 58% of patients. Regarding GI side effects, 25% patients reported grade II symptoms without any grade III toxicity (Tab. 4). No significant correlation was found between dose–volume parameters of the OARs and maximum acute toxicity of the patients.

Fig. 3
figure 3

Incidence of overall baseline (pretreatment) and maximum (onset during treatment) acute genitourinary (GU) and gastrointestinal (GI) symptoms encompassing the detailed symptoms shown in Tab. 4. Numbers in the bars display the number of patients with corresponding toxicity

Full size image
Tab. 4 Detailed incidence of gastrointestinal (GI) and genitourinary (GU) pretreatment symptoms and maximum acute toxicity during treatment
Full size table

Discussion

We report the feasibility of an institutional protocol for definitive treatment of prostate cancer using SIB-IMRT with helical tomotherapy. Acute toxicity and dose–volume histogram (DVH) data were evaluated prospectively in a well-defined intermediate risk patient sample, whereas the risk stratification was performed with a modified scheme based on the D’Amico risk categories. The assessment of acute side effects showed low therapy related GU and GI toxicity in spite of modest dose escalation up to 76 Gy with 2.17 Gy per fraction and a permitted dose of 76 Gy to the anterior rectal wall. This is in line with other studies on dose escalation using IMRT [2, 9, 16, 36] or hypofractionated treatment of prostate cancer [19, 23, 26, 27, 33] and studies combining hypofractionation and dose escalation applying a SIB [5, 12, 14], although the dose to the anterior rectal wall was limited to lower doses in several of these studies as compared to our protocol, either by excluding the rectal overlap from the boost volume or restricting the allowed doses to the overlap regions at lower dose levels.

Guckenberger et al. [12], for example, reported on 100 prostate cancer patients in various risk groups that were treated with definitive conventional IMRT up to doses of 73.91–76.23 Gy with 2.31 Gy per fraction to the prostate and the base of seminal vesicles with safety margins of 5 mm without rectal overlap (PTV-2). PTV-1 encompassed the prostatic gland and the proximal 2 cm of seminal vesicles with a three-dimensional margin of 10 mm except for the posterior direction with 7 mm. In this volume, the total dose was restricted to about 58–60 Gy in 1.84 Gy fractions. According to the risk group, 25% of patients received treatment to the pelvic lymphatics with 46 Gy. The authors reported lower GU and GI side effects with symptoms ≥ grade II in 36% and 8%, respectively. Grade III GU toxicity was observed in only 1% of patients and no grade III GI side effects were seen.

Furthermore, Di Muzio et al. [4] also assessed SIB-IMRT, treating 60 prostate cancer at any stage to different doses with tomotherapy. A subgroup of 31 low-risk patients in their population was treated similar to our patient sample, but using a stronger hypofractionation in the SIB (71.4 Gy, 2.55 Gy per fraction to the prostate and margins of 8 mm, except in the cranial–caudal direction with a margin of 10 mm) and prescribing a lower total dose to the large PTV (61.6 Gy, 2.2 Gy per fraction to the prostate and the proximal portion of seminal vesicles). Overall, DVH data of that study are comparable to our study regarding the rectal Dmean. V40 and V50 are slightly lower in our data, whereas Dmax and V65 are reported slightly lower by the Italian group. The last aspects can be most probably explained by the lower dose prescribed to the overlap volume between the SIB volume and the rectum in the Italian study (65.5 Gy, 2.34 Gy per fraction). Regarding bladder doses higher Dmean, V40 and V55 mean values can be found in the Italian trial compared our data, with a similar value for V60. Grade II and III GU toxicity, assessed with the RTOG score, was reported for 7/31 (22%) and 1/31 (3%) patients, respectively. No higher than grade I GI toxicity was observed.

As already mentioned, the lower GI toxicity observed in the two above presented studies compared to our results might be explained to some extent by providing a stronger dose limitation to the rectal overlap or even sparing the rectal overlap from the boost volume compared to our treatment protocol.

On the other hand, Kassim et al. [15] showed that excluding the rectal overlap from the boost volume might result in a marked decrease of tumor control due to underdosages, as they reported on a planning study that assessed in each case two plans of 36 prostate cancer patients to a total dose of 78 Gy in 2 Gy per fraction to the boost volume once including and once excluding the rectal overlap, respectively.

Comparing the different GU toxicities of the above discussed studies including our own, aside from the different pretreatment symptoms, the use of different toxicity scores must also be considered and might explain the discrepancies to some extent. As can be seen in Tab. 1, nocturnal urinary frequency higher than 6 for example is classified as grade III toxicity in our modified score. In contrast to that, only GU symptoms requiring medical intervention are defined as grade III toxicity using the CTCAE score. Furthermore, most of the GU side effects were assessed more sensitively with our adapted score compared to the standard CTCAE or RTOG scores. Regarding this, the reported overall GU toxicity of our study is in the range of already published studies, since the only observed grade III symptom in the present trial was nocturia.

DVH data of the herein analyzed patient group compare favorably to the data of a patient sample that was treated previously with 3D conformal radiotherapy to doses of 74 Gy at our institution using 10 mm margins without daily image guidance [8]. Rectal V35, V50, and V65 in that sample were 47%, 35%, and 22% as compared to 48%, 27%, and 14%, respectively, in the present study.

The reduction of the safety margins in our study though was well considered, as these margins compensate for the extent of extracapsular spread, intrafractional motion during radiotherapy and uncertainties in contouring. Possible interfractional set-up errors are minimized in our protocol by daily MV-CT scans prior to radiation and are, therefore, not incorporated into the safety margins [38]. Regarding the extent of extracapsular spread, Schwartz et al. [31] found a range of extraprostatic tumor spread from 0–5.9 mm by analyzing 404 whole mounted prostatectomy specimens and stated a GTV to CTV margin of 5 mm sufficient to account for microscopic spread. For intrafractional motion, Kotte et al. [17], analyzing 427 patients with 11,426 prostate position verifications based on fiducial gold markers, calculated that a lower limit for margins of 2 mm would be sufficient to account for intrafractional prostate position shift with slightly larger margins in the craniocaudal direction. In contrast, Fiorino et al. [6], analyzing 410 MV-CTs of 17 prostate cancer patients treated with tomotherapy, reported margins of at least 5–6 mm being appropriate to compensate for intrafractional motion, IGRT intrinsic uncertainties, and interobserver variability with an estimated standard deviation of 1 mm for the latter two. In a recently published study, Wang et al. [34] assessed the intrafractional prostate motion of 59 patients with or without an endorectal balloon for prostate immobilization and showed that using an endorectal balloon 3 mm margins are sufficient to compensate for the prostate motion in 95% of treatment time compared to 5 mm in the non-endorectal group.

Considering these results our margins of at least 8 mm (5 mm GTV to CTV expansion regarding microscopic spread and 3 mm CTV to PTV extension including margins for intrafractional motion and uncertainties in contouring) in every direction except to the rectum as an anatomical barrier with 3 mm (no margin for microscopic spread) seem to be appropriate to minimize the risk of geographical miss, though not considering IGRT intrinsic uncertainty and interobserver variability with explicit margins.

Conclusion

These preliminary results regarding acute tolerability of this institutional treatment protocol for slightly hypofractionated prostate SIB-IMRT and IGRT with tomotherapy are promising. Assessing late toxicity, local control, and overall survival are issues of an ongoing study.