Intensity-modulated and 3D-conformal radiotherapy in hypofractionated prostate cancer treatment using Elekta Beam Modulator^TM micro-MLC: A dosimetric analysis

To the Editor,

External beam radiotherapy is a therapeutic option for the radical treatment of localized prostate cancer [1]. High-dose conformal radiotherapy with conventional 2 Gy daily fractions to a total dose of 74–80 Gy is the standard of care, with an overall treatment time of 7–8 weeks [2]. However, hypofractionated radiation therapy (HFRT) for localized prostate cancer is becoming very attractive and several phase 3 randomized controlled trials have demonstrated the safety of moderate hypofractionation [3–8]. Many of these trials employed intensity-modulated radiation therapy (IMRT) with image-guidance in order to deliver the hypofractionated treatment [4,6,8]. However, three-dimensional conformal radiotherapy (3DCRT) has been used too [3,5,7]. Respect to 3DCRT, IMRT increases the dose gradient near the structures maximizing the dose to the tumor while sparing normal tissue. However, it requires more time for planning and for daily treatment delivery.

Micro-multi-leaf collimators (micro-MLCs) offer the potential to improve target dose distribution and normal tissues sparing both in 3DCRT and in IMRT [9,10].

The Elekta Beam Modulator^TM is a high definition multileaf collimator (4 mm leaf width at the isocenter) integrated on the Elekta Synergy S linear accelerator (Elekta Oncology Systems, Crawley, UK) [11].

The purpose of this dosimetric study is to evaluate the impact of the Elekta Beam Modulator^TM micro-MLC in the development of IMRT and 3DCRT plans, using two different hypofractionation schedules for prostate cancer radiotherapy.

Material and methods

Patients

For this dosimetric study, planning computed tomographies (CTs) from 10 consecutive patients affected by localized prostate cancer (T1-T2) were selected; all these patients underwent 3D-conformal image-guided radiotherapy (IGRT) between April and September 2008, at our Radiation Oncology Therapy Unit. The radiotherapy treatment was delivered using an Elekta Synergy S linear accelerator equipped with the Elekta Beam Modulator^TM micro-MLC (Elekta Oncology Systems), with a conventional fractionation schedule: 66 Gy (2 Gy per fraction) to prostate and seminal vesicles, 76 Gy (2 Gy per fraction) to the prostate only.

Volume definition

All 10 patients underwent CT and MRI scan under radiotherapy planning conditions, as described elsewhere [12]. For each patient clinical target volume (CTV) and organs at risk (OARs) were outlined. In particular, for this retrospective analysis the target volume consisted of prostate + seminal vesicles (CTV). Planning target volume (PTV) was generated by an asymmetric expansion of CTV (6 mm in all directions except at the posterior margin, where a 4 mm expansion was used).

The rectum was contoured as solid organ from the 8th slice (2 cm) above the anal verge to the recto-sigmoid junction; the bladder was contoured in its entirety.

Elekta Beam Modulator^TM micro-MLC features

The Elekta Beam Modulator^TM consists of 40 opposed pairs of 75-mm thick tungsten leaves, which project a width of 4 mm at the isocenter. The micro-MLC features are described in details elsewhere [9,11].

Dose prescription

For this study, we selected two different hypofractionation schedules corresponding to the same biological equivalent dose (BED) of 77 Gy (dose fraction 2 Gy, α/β = 1.5 Gy) [13]:

Dose prescription 1 (DP1): total dose 60 Gy in 20 daily fractions.

Dose prescription 2 (DP2): total dose 45 Gy in 10 daily fractions.

3DCRT and IMRT planning

A total of 40 plans were developed for the 10 patients. Twenty 3DCRT plans were processed, 10 with DP1 and 10 with DP2, respectively, with the same six fields arrangement (gantry angles: 45°, 90°, 135°, 225°, 270°, 315°).

Twenty step-and-shoot IMRT plans were developed, 10 with the DP1 and 10 with DP2, respectively, with the same field arrangement (gantry angles: 0°, 45°, 100°, 260°, 315°) and the same optimization parameters. Treatment plans were processed on Pinnacle Treatment System (Philips Medical System, Andover, MA, USA) and optimized using the Direct Machine Parameter Optimization Algorithm (DMPO). Table I shows the DMPO optimization parameters and the IMRT objective functions.

Table I. Settings used in DMPO optimization and IMRT planning objectives for DP1 and DP2. For PTV V₉₅ is the percent volume receiving at least 95% of the prescription dose, D₅ is the percent of the prescription dose covering 5% of the target volume. OAR dose constrains are expressed in terms of percent of volume receiving the x absolute dose.

DMPO parameters
Maximum iterations	35
Maximum number of segments	60
Minimum segment area	2 cm^2
Convolution dose iteration	7
Minimum segment MUs	5
Planning objectives
	DP1	DP2
PTV	V95 > 95%	V95 > 95%
PTV	D5 < 110%	D5 < 110%
Rectum	V58 < 20%	–
Rectum	V54 < 25%	V43 < 25%
Rectum	V50 < 35%	V40 < 35%
Rectum	V42 < 50%	V33 < 50%
Rectum	V33 < 60%	V27 < 60%
Bladder	V58 < 35%	V43 < 50%
Bladder	V50 < 40%	V40 < 40%

MU, monitor units.

A criterion of 95% of the target volume receiving the 95% of prescribed dose was satisfied for all plans.

Dosimetric analysis

Dose-volume histograms (DVHs) were used to provide the quantitative comparison between plans; in particular, we analyzed differences between PTV conformity and coverage, differences between rectum and bladder dose sparing.

The PTV dose coverage and conformity, expressed in terms of coverage and conformity index (CVF and CN), were previously defined [9]. For OARs, a set of appropriate V_x (percent of OAR volume receiving the x dose) was evaluated. We chose Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC) constraints for bladder and rectum [14,15]. In particular, V₄₀ < 60%, V₅₀ < 50%, V₆₀ < 35%, V₆₅ < 25% and V₇₀ < 20%, for the rectum and V₆₀ < 40%, V₆₅ < 50% and V₇₀ < 35% for bladder were converted for the two different hypofractionated schedules, using α/β = 3 Gy. The following constraints were obtained:

DP1: V₃₃ < 60%, V₄₂ < 50%, V₅₀ < 35%, V₅₄ < 25% and V₅₈ < 20% for rectum and V₅₀ < 40%, V₅₄ < 50% and V₅₈ < 35%, for bladder.

DP2: V₂₇ < 60%, V₃₃ < 50%, V₄₀ < 35%, V₄₃ < 25% for rectum and V₄₀ < 40%, V₄₃ < 50% for bladder.

Statistical analysis

Statistical analysis was carried out using SPSS 9.0 (SPSS Inc, Chicago, IL, USA); data were tested for normality with the Kolmogorov-Smirnov test, and different datasets were compared with paired Student's t-tests (two-tailed). All pairwise comparisons were conducted using the Bonferroni-Holm correction [16].

Results

IMRT shows a better conformity respect to 3DCRT (CN_(3DCRT) = 0.59 ± 0.05, CN_(IMRT) = 0.77 ± 0.05, p = 0.00; CNs are expressed in terms of mean value, n = 10). A good PTV coverage of the 95% isodose was obtained both in 3DCRT and IMRT, without statistically significant differences (CVF_(3DCRT) = 0.97 ± 0.01, CVF_(IMRT) = 0.98 ± 0.01, p = 0.22; CVFs are expressed in terms of mean value, n = 10). CVF and CN for DP2 prescription confirm DP1 results (data not shown).

Figure 1 shows, for every single patient, rectum and bladder V_x values obtained using DP1 3DCRT and IMRT plans, respectively. Figure 2 shows the values obtained in DP2 3DCRT and IMRT plans, respectively. In particular, in DP1 3DCRT plans rectum dose constraints are not respected in three patients (ID 1: V₅₀ = 37.3%, V₅₄ = 30.8%; ID 7: V₅₀ = 35.5%, V₅₄ = 29.5%; ID 8: V₅₀ = 37.2%, V₅₄ = 29.9%) and bladder dose constraints are not respected in two patients (ID 3: V₅₀ = 46.1%, ID 4: V₅₀ = 46.2%). Using IMRT, all QUANTEC constraints were always satisfied. Respect to 3DCRT, IMRT plans allow a better OARs sparing in both DP1 (rectum: V_33(3DCRT)/V_33(IMRT) = 1.19, p = 0.0010; V_42(3DCRT)/V_42(IMRT) = 1.35 p = 0.0001; V_50(3DCRT)/V_50(IMRT) = 1.54 p = 0.0000, V_54(3DCRT)/V_54(IMRT) = 1.61 p = 0.0001 and V_{58(3DCRT) /}V_58(IMRT) = 1.56 p = 0.0170; bladder: V_50(3DCRT)/V_50(IMRT) = 1.36, p = 0.009, V_54(3DCRT)/V_54(IMRT) = 1.27 p = 0.003 and V_58(3DCRT)/V_58(IMRT) = 1.61 p = 0.002) and DP2 (rectum: V_27(3DCRT)/V_27(IMRT) = 1.13, p = 0.0120, V_33(3DCRT)/V_33(IMRT) = 1.27 p = 0.0003, V_40(3DCRT)/V_40(IMRT) = 1.44 p = 0.0002, V_43(3DCRT)/V_43(IMRT) = 1.33 p = 0.0230 for rectum; bladder: V_40(3DCRT)/V_40(IMRT) = 1.26 p = 0.010, V_43(3DCRT)/V_43(IMRT) = 1.66 p = 0.004), as shown in Table II.

Figure 1. DVH-based parameters (V_x) for rectum and bladder in 3DCRT (gray) and IMRT (black) for DP1 prescription. Dashed lines represent QUANTEC constrains. Gray solid and black dotted lines represent the V_x mean value for 3DCRT and IMRT respectively (n = 10).

Figure 2. DVH-based parameters (V_x) for rectum and bladder in 3DCRT (gray) and IMRT (black) for DP2 prescription. Dashed lines represent QUANTEC constrains. Gray solid and black dotted lines represent the V_x mean value for 3DCRT and IMRT respectively (n = 10).

Table II. Comparison of DVH-based parameters for rectum and bladder in 3DCRT and IMRT for DP1 and DP2 prescriptions respectively (n = 10); V_x are expressed in terms of mean value ± SD. P_{significant levels} were calculated following Bonferroni-Holm correction.

DP1
	3DCRT	IMRT
	%	%	Vx(3DCRT)/Vx(IMRT)	psignificant level	p-Value
Rectum
V33	44.5 ± 11.3	37.5 ± 8.3	1.19	0.025	0.0010
V42	33.9 ± 10.1	25.3 ± 7.3	1.35	0.017	0.0001
V50	25.9 ± 8.6	16.9 ± 6.1	1.54	0.010	0.0000
V54	20.5 ± 7.6	12.7 ± 5.1	1.61	0.013	0.0001
V58	10.2 ± 5.4	6.5 ± 3.5	1.56	0.050	0.0170
Bladder
V50	27.6 ± 15.0	24.3 ± 12.6	1.36	0.050	0.009
V54	23.3 ± 13.0	18.5 ± 9.6	1.27	0.025	0.003
V58	17.5 ± 10.5	10.9 ± 5.8	1.61	0.017	0.002
DP2
	%	%	Vx(3DCRT)/Vx(IMRT)	psignificant level	p-Value
Rectum
V27	39.8 ± 10.3	35.1 ± 11.5	1.13	0.025	0.0120
V33	31.5 ± 9.2	24.9 ± 9.3	1.27	0.017	0.0003
V40	21.1 ± 7.3	14.7 ± 6.9	1.44	0.013	0.0002
V43	12.1 ± 5.5	9.1 ± 4.9	1.33	0.050	0.0230
Bladder
V40	23.7 ± 13.0	18.8 ± 9.9	1.26	0.050	0.010
V43	18.4 ± 10.8	11.1 ± 5.6	1.66	0.025	0.004

SD, standard deviation.

Table III shows an evaluation of total monitor units (MUs) and time for daily treatment delivery in 3DCRT and IMRT for both hypofractionated schedules (MUs and time are expressed in terms of mean value for n = 10 patients).

Table III. Monitor units (MUs) and time for daily treatment delivery of 3DCRT and IMRT for DP1 and DP2. MUs and Times are expressed in terms of mean value ± SD.

	3DCRT	IMRT
MUs			psignificant level	p-Value
DP1	536 ± 40	697 ± 77	0.05	0.0003
DP2	781 ± 86	1038 ± 136	0.03	0.0003
Times for daily treatment delivery (min)
DP1	4.0 ± 0.3	8.0 ± 0.6	0.05	0.0003
DP2	5.0 ± 0.5	10.0 ± 0.7	0.03	0.0002

SD, standard deviation.

IMRT requires more MUs (DP1: MUs_(3DCRT) = 536 ± 40, MUs_(IMRT) = 697 ± 77; DP2: MUs_(3DCRT) = 781 ± 86, MUs_(IMRT) = 1038 ± 136). The mean time for an IMRT daily treatment delivery is twice the time measured for a 3DCRT one: (DP1: Time_(3DCRT): (4.0 ± 0.3) min, Time_(IMRT): (8.0 ± 0.6) min; DP2: Time_(3DCRT): (5.0 ± 0.5) min , Time_(IMRT): (10.0 ± 0.7) min.

Discussion

Prostate cancer HFRT is becoming of increasing interest with the recognition of a potential improvement in the therapeutic ratio, when treatments are delivered in larger size fractions. In fact, from a radiobiologic point of view slowly proliferating cells, such as prostate cancer cells, might be more sensible to high doses per fraction due to the high number of DNA double strand breaks caused by each fraction.

The safely use of hypofractionated regimens is possible thanks to radiation oncology technological progress (3DCRT, IMRT and image-guidance). In particular, daily IGRT seems to be determinant, allowing precise targeting and organs at risk sparing [17]. The use of IMRT or 3DCRT appears to have no impact on outcomes or toxicity for HFRT. In fact, Wong et al. reported that IMRT and 3DCRT are both employed in prostate cancer hypofractionated regimens; the main difference between the two techniques is that 3DCRT requires less planning and execution time [18].

A recent systematic review of hypofractionation for localized prostate cancer confirms that there is no uniformly used radiation therapy technique or a single hypofractionated regimen [19]. For instance, Arcangeli et al. carried out a hypofractionated treatment (62 Gy/3.1 Gy per fraction) using 3DCRT, whereas Martin et al. used IMRT (60 Gy/3 Gy per fraction) [3,20].

We decided to investigate moderate hypofractionation (DP1: 60 Gy/3 Gy per fraction and DP2: 45 Gy/4.5 Gy per fraction) because this type of fractionation is emerging as a standard treatment. However, data on extreme hypofractionation (5–10 Gy per fraction) are lacking because randomized controlled studies are not available.

Several studies compared 3DCRT versus IMRT in prostate cancer radiation therapy from a dosimetric point of view. Reddy et al. compared 3DCRT and IMRT radiotherapy treatment of prostate+ seminal vesicles in terms of dose to rectum, bladder and femurs [21]. They showed that both in 3DCRT and in IMRT, the prostate+ seminal vesicles volumes play a significant role in the dose delivered to rectum and bladder. They concluded that from a dosimetric point of view 3DCRT and IMRT plans were comparable for patients with small to medium size prostate+ seminal vesicles volumes (< 85 cm³), while IMRT was better suited for treatment of patients with large prostate. Buckey et al. compared 3DCRT and IMRT prostate radiation therapy plans to ascertain if there were dosimetric advantages in performing IMRT. They concluded that both 3DCRT and IMRT plans were able to meet the dosimetric constraints from RTOG study 0415 [22].

In this scenario, the choice of IMRT respect to 3DCRT should take into account all available resources in terms of equipment and personnel.

Our study compares hypofractionated 3DCRT and IMRT plans in terms of dosimetric parameters, using the Elekta Beam Modulator^TM micro-MLC as beam shaper.

We demonstrate that in 3DCRT plans the averaged values of all analyzed V_x respect QUANTEC constraints. This result is due to the 4 mm m-MLC, which better protects OARs from radiation and better adapts every beam to the shape of the PTV [11]. Only few cases in DP1 had respectively rectum V₅₀, V₅₄ (in three patients) and bladder V₅₀ (in two patients) outside the requested constraints. We registered in these patients a clinical target volume larger than average value (> 60 cm³; n = 10). Compared to 3DCRT, IMRT plans have better target dose conformity and better normal tissue sparing for both dose prescriptions. However, in IMRT monitor units and time needed for daily treatment delivery are quite doubled respect to 3DCRT.

In conclusion, IMRT is better than 3DCRT in terms of target conformity and OARs dose sparing in hypofractionated prostate cancer radiotherapy. However, also 3DCRT plans processed with Elekta micro-MLC respect QUANTEC constraints except for patients with large prostate volumes.

IMRT seems to be the best technique from the dosimetric point of view. IMRT is more costly in terms of equipment, personnel and time. We can conclude that as IMRT also 3DCRT plans processed with a micro-MLC provide good radiotherapy treatment plans, while IMRT should be mandatory in particular patient settings (i.e. in patients with large prostate).

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