Abstract
Aim: To quantify the impact of decreased margins for two treatment techniques, three-dimensional conformal radiotherapy (3D-CRT) and volumetric-modulated arc therapy (VMAT), on local control in curative treatment of prostate cancer. Materials and Methods: The planning target volume (PTV) margins were decreased in steps of 1 mm from 10 to 1 mm. Treatment plans using 3D-CRT and VMAT technique were produced for all margin sizes and the dose to the neuro vascular bundles (NVB), that was not included in the PTV, was investigated. Results: Due to the more conformal dose delivery using VMAT, the dose to the NVB decreased more rapidly by VMAT compared to the 3D-CRT plans. The dose difference was significant for margins from 1-7 mm. Conclusion: One should be very cautious before clinical routines are changed, bearing in mind whether the change means more conformal treatment technique, smaller margins or target segmentation in different imaging modalities.
Prostate cancer (PC) is the most common cancer among men in Sweden, EU and North America. In Sweden, 9500 men develop PC each year and approximately 20% of all Swedish men will be diagnosed with PC during their lifetime. Each year, 2500 men die from PC in Sweden (which is 5% of all mortality). Many patients experience side-effects of their PC treatment such as rectal bleeding, incontinence and impotence. In addition to great costs for the society, these side-effects cause impaired quality of life for a long period of time.
With technical advances in radiotherapy, such as three-dimensional conformal radiotherapy (3D-CRT) (1), intensity-modulated radiotherapy (IMRT) (2) and volumetric arc therapy (VMAT) (3), the conformity of the dose distribution has been improved (4). Despite this, in radiotherapy for PC, the proximity of the prostate and the organs at risk (OAR) are limiting factors (5).
The alteration in position and shape of the prostate and the OAR implies uncertainties in set-up and delivery of radiotherapy (6-8), which could lead to treatment fields partly or completely missing the target (9). In the International Commission on Radiation Units and Measurements (ICRU) report 50 an additional margin is applied to the delineated clinical target volume (CTV) to account for geometrical uncertainties, the planning target volume (PTV) (10).
The PTV margin has been reduced from typically 15-20 mm to 3-7 mm with the use of fiducial markers (11) and image-guided radiotherapy (IGRT) (12), where daily pre-treatment imaging was used to take into account the inter-fractional variation of the geometrical position of the prostate (13).
According to studies from Heemsbergen et al. (14) and Witte et al. (15), local control was reduced using IMRT compared to 3D-CRT box technique in patients with high-risk tumours. It has also been shown by Chao et al. that the microscopic spread in the neuro vascular bundles (NVBs) is dependent on the differentiation of the tumour (16).
In this comparative treatment planning study, we studied the difference in tumour control probability (TCP) in the NVBs for different PTV margins (1-10 mm in steps of 1 mm) using the 3D-CRT box technique and VMAT in order to explain the clinical observation in the above-mentioned studies.
Materials and Methods
Patients. Fourteen consecutive patients with biopsy-proven localized adenocarcinoma T1-T3NX/0MX/0 indicated for curative radiotherapy of localized PC were included in the study. The prostate CTV was 21 ml to 69 ml (mean=44 ml).
Patient immobilization and positioning. For treatment planning purposes, all patients underwent computed tomographic (CT) scanning using 2.5 mm slice thickness. The patients were placed on their back on the table, and a knee and foot fixation was used.
Target volumes and OAR. The prostate gland was contoured and defined as the CTV, using magnetic resonance imaging (MRI) as support for target segmentation. The seminal vesicles were excluded from the CTV. The CTV to PTV margin was 1-10 mm in all directions, increased in steps of 1 mm for a total 10 PTVs for each patient (Figure 1). The NVBs were segmented separately and not included in the CTV, thus the NVBs did not affect the PTV. The NVB-CTV was defined as a circle of 7 mm (Figure 1) in order to cover 90% of microscopic spread in tumours with Gleason score 7 (16). The rectum, the bladder and the femoral heads were segmented as OAR. The length of the rectum was 3 cm above and below the CTV in the longitudinal direction and was defined as the whole content inside the outer wall. The bladder was segmented as the whole content inside the outer wall.
Treatment planning. For each patient in the study, two treatment plans for each PTV were made using 3D-CRT and VMAT respectively, which gave a total of 20 treatment plans per patient. The treatment plans were made in RayStation 5 (RaySearch laboratories, Stockholm, Sweden) using an in-house developed ironPython script, which is a built-in functionality in the treatment planning system. The treatment plans were made for an Elekta Synergy instrument (Elekta, Stockholm, Sweden) with an Agility multi-leaf collimator. The 3D-CRT treatment plan used four equally spaced beams (0. 90. 180 and 270 degrees) using 15 MV x-ray beams and the VMAT treatment plan one full arc (182-178 degrees) using 6 MV x-ray beams. All plans were calculated with a voxel size of 1×1×2.5 mm3. The prescribed dose was 78 Gy in 39 fractions.
Evaluation criteria. The dose coverage of the PTV was to meet the criteria that 98% of the volume should receive at least 95% (D98%>95%) and 2% of the volume at most 107% (D107%<2%) of the prescribed dose (10). Due to different PTV margins, there were no dose evaluation criteria for the OAR. Depending on the size of the PTV, two levels of average dose to the rectum and the bladder were used as optimization criteria to reach a dose level in the OAR as low as possible with retained dose conformity to the PTV.
TCP was calculated for the prostate-CTV and the NVB-CTV, and normal tissue control probability (NTCP) was calculated for the rectum and the bladder. TCP and NTCP parameters are presented in Tables I and II, respectively. For TCP and NTCP calculations, the RayBiology module in RayStation was used.
Statistics. A paired, double sided t-test was used for both TCP and NTCP comparisons (Matlab 2016b, MathWorks Inc., Natick, MA, USA). A significance level of 5% was used. Only one p-value is mentioned for the cases where significance was shown in the results, due to the great number of p-values in this work. Instead, confidence intervals are used to show if the difference in TCP and NTCP was significant or not for each PTV margin. The difference was considered significant if the confidence interval did not contain zero.
Results
In Figure 2, the TCP values for 3D-CRT and VMAT are plotted for the left and right NVB and for the prostate.
The TCP values for the left and right NVB were calculated with four different α/β-values: 1.5, 3.0. 6.0 and 10 Gy (see Figure 2). The TCP values for both NVBs were significantly higher when the 3D-CRT technique was used for margins of 1-7 mm (p<0.05). The parameters used for the TCP model are shown in Table I.
In Figure 3, the NTCP values and their confidence intervals for the rectum and bladder are shown. The parameters used for the NTCP model are shown in Table II. There was a significant difference between the two treatment techniques for all margins used, both for the rectum and the bladder, where the probability for normal tissue complications were significantly lower when VMAT was used.
All treatment plans fulfilled the dose coverage criteria D98% >95% for the corresponding PTV. The mean TCP values for the prostate are shown in Figure 2.
There were no differences in TCP values for the prostate-CTV, regardless of PTV margin and α/β-value used (see Figure 2).
Discussion
There are at least two studies describing an increase in clinical local failure during radiotherapy of prostate cancer when the treatment techniques have become more conformal (14, 15). It has also been shown that the microscopic spread of tumour cells in the NVB is dependent on the differentiation of the tumour (16).
In this study, we investigated the impact on the TCP in the NVBs during radiotherapy of PC for different treatment techniques and margins. We showed that there was a significant difference in TCP for NVB-CTV using 3D-CRT and VMAT for PTV margins up to 7 mm (p<0.05), where the TCP was higher when using 3D-CRT. For larger margins, the NVBs were included in the PTV and the TCP was similar for both treatment techniques. But the NTCP for the rectum and the bladder was significantly lower for VMAT, irrespective of margin, due to the more conformal dose distribution around the PTV.
This study confirms that in regard to the TCP in the NVB, more conformal techniques influence local control. The decreased local control is due to a more conformal dose distribution and the reduction of the PTV margins, which reduces side-effects. The margins can be safely reduced due to on-line verification of the prostate position (IGRT and fiducials). Therefore, this study shows the importance of better outlining of CTV in PC, taking into account information concerning microscopic spread in the NVBs and information of capsular infiltration for other investigations such as MRI. The disadvantage of a larger CTV is an increased dose to the rectum. By separating the rectum from the prostate using rectum spacers (26-31), the dose to the rectum will be lowered. This will be even more important when using small margins and very conformal techniques such as VMAT and proton therapy.
Even the modality in which the CTV is outlined affects the risk of small PTV margins. It has been shown by Gunnlaugsson and co-workers that the CTV is 23% larger when it is defined on computed tomography compared to defining it on MRI (32). They also showed that during extreme hypo-fractionated radiotherapy, the prostate increases in volume during the treatment by up to 15%. In this case, the risk of local relapse is larger when outlining the CTV on MRI if the same PTV margin is used for both computed tomography and MRI. Therefore, it is of great importance to understand the meaning of the PTV margin which has to be adapted to the modality used.
Since this is a treatment planning study, we only studied a static scenario, which explains the high TCP values for prostate-CTV regardless of the margin size since prostate-CTV is always inside the PTV. In this case, the TCP values are used as a measure of acceptable dose coverage of the CTV. For a dynamic case, the TCP values would decrease with decreasing PTV margin. In addition, the same parameters in the TCP model were used for both prostate and NVBs. However, in this case, the TCP value was used for relative comparisons and absolute values are of less importance.
There are clinical observations in which there were fewer local relapses after radiotherapy of PC when the treatment technique was less conformal (14, 15). This study shows that the dose coverage of the NVBs decreases when narrower PTV margins are used in combination with changing treatment technique from 3D-CRT to VMAT. Therefore, one should be very cautious before clinical routines are changed, as to whether the change means more conformal treatment technique, smaller margins or target segmentation in different imaging modalities.
Acknowledgements
The Authors acknowledge support from “Stiftelsen Onkologiska klinikens i Uppsala forskningsfond” (local foundation).
Footnotes
This article is freely accessible online.
Conflicts of Interest
The Authors have no conflicts of interest to report. The Authors alone are responsible for the content and writing of the article.
- Received February 17, 2017.
- Revision received March 10, 2017.
- Accepted March 13, 2017.
- Copyright© 2017, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved