Several previous prospective randomized trials and their meta-analysis proved that in the treatment of breast tumours, radiation therapy is an important part of breast conserving therapy. Breast conserving surgery and the subsequent irradiation of the remaining breast tissue ensure the same survival rate as radical breast surgery.1, 2 Nowadays, accelerated partial breast irradiation (APBI) is becoming more and more an accepted treatment for early-stage invasive breast tumours.3, 4, 5, 6, 7, 8 Two main advantages are short overall treatment time (5 days) and smaller irradiated volume, which helps for the organs at risk (OAR) to be spared.
In the first external beam APBI trials the patients were treated with 3 dimensional conformal (3D-CRT) technique.9 Studies with long follow up times can be also found with brachytherapy.10, 11 These trials have the largest number of patients treated and the longest follow-up time. In our previous paper the dosimetrical differences of the external beam and interstitial brachytherapy for APBI were evaluated.12 It was found that the target volume can be appropriately irradiated by both techniques, but brachytherapy generally spares normal tissues and organs at risk better than IMRT. In 2006 a phase II sequential trial was launched at our institution.13 In the first part of the trial between 2006 and 2011 forty-four patients were treated with 3D-CRT technique, while in the second part between 2011 and 2014 sixty patients were treated with image-guided “step & shoot” intensity-modulated (IMRT) technique, using 5 coplanar fields. Currently at our institution APBI treatments are carried out with “sliding window” IMRT technique according to our preliminary study.14
In current study, we conducted a dosimetric comparison of the traditional 3D-CRT, the intensity-modulated “step & shoot” (SS), “sliding window” (SW) and “RapidArc” (RA) techniques for APBI.
For this study 40 patients were selected, who were previously treated with early-stage invasive breast cancer at our institution between 2006 and 2014 in a phase II APBI trial.13,15 The study protocol was evaluated and accepted by the institutional and national ethics committees, and all patients provided written informed consent before enrolment. The trial was registered at
During planning CT scan, patients were positioned supine in a wingboard fixation device (Civco, USA). Around both breasts and on the surgical scar metal wires were placed to increase the precision of contouring. CT scans started from the mandible and included the complete volume of the lungs and the volumes of both breasts. The most crucial part of partial breast irradiation is the exact localization of the excision cavity. Using titanium clips and the 3 mm thickness of the planning CT slices, the borders of surgical cavity could be defined in accepTable precision.
The contouring and 3D-CRT plans were made with Pinnacle 8.0m (Philips, The Netherlands) treatment planning system, while for the plans of SS, SW and RA the Eclipse 11 (Varian, USA) planning software was used. In the first step of target volume definition, the tumour bed was contoured by the radiation oncologist based on the seroma and the surrounding clips. In the second step, on the basis of the intact surgical margins and pathological results, it was expanded in 6 directions with margins of different size to define CTV. The minimal and maximal value of the extension between the tumour bed and the CTV was 5 and 18 mm, respectively.16 The CTV was limited by the border of ipsilateral breast tissue. The safety margin between the CTV and PTV was 5 mm. As the skin is an organ at risk during partial breast irradiation, a 5 mm thick volume of the skin was cropped from PTV to create PTVeval. Heart, ipsilateral and contralateral breasts and lungs, non-target breast were contoured as organs at risk. Non-target breast was created with the extraction of PTV from ipsilateral breast.
36.9 Gy in 9 fractions was delivered on 5 consecutive days, twice a day with at least 6 hours interval between fractions. 100% of the PTVeval had to be covered with at least 95% of the prescribed dose. The allowed maximum dose was 110%. To make the techniques comparable, the same target coverage was achieved for all plans. Based on our protocol, the dose limits of the organs at risk were as follows: V100% < 35% and V50% < 60% for the ipsilateral breast, V30% < 20% for the ipsilateral lung, V15% < 10% for the heart in case of right sided tumours, while in case of left sided tumours, the value of heart V5% had to be smaller than for conventional whole breast irradiation. Vxx% is the percentage of a region of interest (ROI) receiving at least the XX percent of the prescribed dose. These requirements were fulfilled in all of the plans. For contouring and for planning aims the NSABP B-39/RTOG 041317 guideline was used.
The 3D-CRT plans were created with 4 or 5 wedged conformal non-coplanar fields from tangential directions (Figure 1). The IMRT plans were created with 5 or 6 coplanar fields covering 190°, four of these fields positioned in the medial “first” and lateral “last” 30°. The RA plans consisted of two coplanar arcs created between the most medial field of the IMRT plans and gantry angle of 180° (Figure 1). Avoidance sectors were not applied, only strict constrains were given for ipsilateral lung, heart and spine. All plans were created for linear accelerator equipped with 5 mm wide MLC and 6 MV photon energy was used. CCC algorithm in Pinnacle and AAA algorithm in Eclipse were used for dose calculations.
For all techniques and all contoured organs volumes, minimum -, maximum and mean doses were recorded. Also Vxx% and Dxx% parameters were evaluated for all ROIs.
For the PTVeval homogeneity index and conformation number were calculated. Homogeneity index was calculated according to the ICRU83:
And the conformation number is:
where PTVref is the volume of PTVeval which is covered by the reference isodose, VPTV is the volume of PTVeval and Vref is the tissue volume encompassed by the reference isodose curve.18 The plan quality index (PQI), defined by Leung
Statistical analysis was performed for all parameters with repeated measures ANOVA and Tukey post hoc test or Friedman test and Dunn post hoc test in GraphPad Instat (GraphPad Software, USA).
The mean volume of the tumour bed, CTV and PTV were 15.4 cm3, 80.8 cm3 and 155.5 cm3, respectively. The ratios of CTV and PTV to the ipsilateral breast volume was 8.6% (range: 4–19.1%) and 16.8% (range: 8.3–40.1%), respectively.
For each technique the V95% value of PTVeval was at least 99.5%, except for 5 cases with 3D-CRT technique. There were no significant differences between the techniques in the values of V95% and V90% (Table 1), neither for PTVeval nor for CTV. Figure 2 shows dose distributions for the four irradiation techniques in a representative case. The homogeneity index was significantly better with SW technique than with 3D-CRT, SS or RA techniques. The difference between 3D-CRT and RA technique was also significant. The average conformation numbers of 3D-CRT, SS, SW and RA were 0.60, 0.80, 0.82 and 0.89, respectively, being all differences significant. The average values of homogeneity index, conformal number and PQI and the significance levels are summarized in Table 1.
Quality indices and total number of monitor units of the four irradiation techniques
Mean values (%) |
ANOVA p value | Post hoc tests results |
|||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
3D-CRT | SS | SW | RA | 3D-CRT vs. SS | 3D-CRT vs. SW | 3D-CRT vs. RA | SS vs. SW | SS vs. RA | SW vs. RA | ||||
99.5 | 99.8 | 99.8 | 99.8 | 0.96 | n.s | n.s | n.s | n.s | n.s | n.s | |||
100 | 100 | 100 | 100 | 0.875 | n.s | n.s | n.s | n.s | n.s | n.s | |||
99.9 | 100 | 100 | 100 | 0.001 | n.s | n.s | n.s | n.s | n.s | n.s | |||
100 | 100 | 100 | 100 | 0.123 | n.s | n.s | n.s | n.s | n.s | n.s | |||
0.068 | 0.074 | 0.058 | 0.081 | <0.001 | n.s | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | |||
0.6 | 0.8 | 0.82 | 0.89 | <0.001 | p<0.05 | p<0.05 | p<0.05 | p<0.05 | p<0.05 | p<0.05 | |||
0.5 | 0.36 | 0.34 | 0.43 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | p<0.05 | |||
677 | 814 | 1077 | 1136 | <0.001 | n.s | p<0.05 | p<0.05 | p<0.05 | p<0.05 | n.s |
3D-CRT = three dimensional conformal radiotherapy; n.s. = not significant; RA = intensity-modulated arc therapy; SS and SW = „step and shoot” and „sliding window” intensity-modulated radiotherapy
Table 2 contains all the dosimetric values of the critical organs and their comparisons between different techniques.
Dose to the organs at risk
Mean values (%) |
ANOVA p value | Post hoc tests results |
||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
3D-CRT | SS | SW | RA | 3D-CRT vs. SS | 3D-CRT vs. SW | 3D-CRT vs. RA | SS vs. SW | SS vs. RA | SW vs. RA | |||
14.3 | 13.9 | 12.1 | 12.7 | <0.001 | n.s | p<0.05 | p<0.05 | p<0.05 | p<0.05 | n.s | ||
31.5 | 26.1 | 25.8 | 21 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | p<0.05 | ||
47.5 | 40.2 | 39.9 | 31.6 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | p<0.05 | ||
105.3 | 108.2 | 106.3 | 110.3 | <0.001 | p<0.05 | n.s | p<0.05 | p<0.05 | p<0.05 | p<0.05 | ||
13.1 | 28.1 | 28 | 36 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | p<0.05 | ||
3.9 | 3.6 | 3.6 | 6 | <0.001 | n.s | n.s | p<0.05 | n.s | p<0.05 | p<0.05 | ||
2.6 | 1.9 | 1.9 | 3 | <0.001 | p<0.05 | p<0.05 | n.s | n.s | p<0.05 | p<0.05 | ||
5.6 | 7.4 | 7.4 | 9.9 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | p<0.05 | ||
2.4 | 2.9 | 2.8 | 5.4 | <0.001 | n.s | n.s | p<0.05 | n.s | p<0.05 | n.s | ||
8.7 | 19.3 | 17.5 | 38.9 | <0.001 | n.s | n.s | p<0.05 | n.s | n.s | p<0.05 | ||
3 | 1.5 | 1.2 | 7.8 | 0.043 | n.s | n.s | n.s | n.s | n.s | p<0.05 | ||
0.8 | 1.6 | 1.6 | 3 | <0.001 | n.s | n.s | p<0.05 | n.s | p<0.05 | p<0.05 | ||
1.3 | 8.1 | 7.8 | 22.2 | <0.001 | p<0.05 | n.s | p<0.05 | n.s | n.s | p<0.05 | ||
0.6 | 0.1 | 0.2 | 0.5 | 0.711 | n.s | n.s | n.s | n.s | n.s | n.s | ||
0 | 3.4 | 3.2 | 9.6 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | p<0.05 | ||
0 | 0.1 | 0.1 | 0.1 | 0.001 | n.s | n.s | n.s | n.s | n.s | n.s | ||
0.5 | 4 | 4 | 5.4 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | p<0.05 | ||
0.4 | 3.2 | 3.1 | 4.7 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | p<0.05 | p<0.05 | ||
0.7 | 3.1 | 3.1 | 1.9 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | n.s | n.s | ||
0.3 | 0.2 | 0.2 | 0.0 | 0.013 | n.s | n.s | n.s | n.s | n.s | n.s | ||
1.2 | 3.8 | 3.7 | 3.8 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | n.s | n.s | ||
5.6 | 8.5 | 8.5 | 7.8 | <0.001 | p<0.05 | p<0.05 | p<0.05 | n.s | n.s | n.s |
3D-CRT = three dimensional conformal radiotherapy; MHD = mean heart dose; MLD = mean lung dose; n.s. = non-significant; RA = intensity-modulated arc therapy; SS and SW = „step and shoot” and „sliding window” intensity-modulated radiotherapy
With respect to the ipsilateral breast, the best results for dose reduction were achieved with RA technique which was also confirmed by the conformation number. The V50% values of non-target breast for 3D-CRT, SS, SW and RA were 39.6%, 29.9%, 29.6%, 19.8%, respectively. In the non-target ipsilateral breast the maximum doses were 105.1%, 105.8%, 103.8% and 106.4%, respectively.
With respect to the ipsilateral lung, we received significantly better results for dose-volume values for non-coplanar 3D-CRT in low dose ranges such as 10% of the prescribed dose, while RA resulted in the largest low dose bath (V10%). For higher dose ranges the two static IMRT techniques achieved significantly better results for V40% than the other two.
The heart was better protected from low doses by the 3D-CRT technique. For the V15% values RA technique still provided the worst results, while the static-filed IMRTs reached the lowest doses for the heart.
Both contralateral lung and contralateral breast received low dose with all techniques. Noticeable values and differences can only be found at 5% of the prescribed dose, and only very small volume received a higher dose than 10%. For both organs the 3D-CRT had significantly better results. With 3D-CRT technique, the average maximum doses for the contralateral lung and the contralateral breast were 0.8 Gy and 2.1 Gy, respectively. With static-field IMRT techniques the average values were 3 Gy and 3.1 Gy and with RA 3.5 Gy and 2.5 Gy, respectively.
The PQI values of 3D-CRT, SS, SW and RA techniques were 0.50, 0.36, 0.34 and 0.43. By definition, low PQI values mean high overall plan quality. The two static-field IMRT PQI values were significantly better than the ones for the other two techniques, and RA was also significantly better compared to 3D-CRT. With respect to monitor units, 3D-CRT (677 MU) and SS (814 MU) techniques were significantly lower than with SW (1077 MU) and RA (1136 MU). The difference between SW and RA was also significant.
The results of this study in comparison with other dosimetric studies in APBI can be found in Table 320, 21, 22 and Table 4 for the 3D-CRT and IMRT techniques, respectively.
Comparison between our and other 3D-CRT data published in APBI studies
Target |
Ipsilateral breast (%) |
Non-target breast (%) |
Heart (left-sided lesions) |
Ipsilateral lung |
||||||
---|---|---|---|---|---|---|---|---|---|---|
VPTV (ccm) | VPTV / VIB | V95% (%) | V100% | V50% | V50% | MHD (cGy) | V5% (%) | MLD (cGy) | V30% (%) | |
n.a | n.a | n.a | 26 | 52 | n.a | n.a | n.a | 370 | n.a | |
n.a | 0.17 | 99.9 | 32.8 | 57.6 | 40.9 | n.a | n.a | n.a | n.a | |
236.2 | n.a | n.a | 20.3 | 46.8 | n.a | 76.1 | 6.4 | 193.2 | 4.4 | |
176.4 | n.a | 96.4 | n.a | n.a | 23.2 | 200 | 15.9 | 370 | n.a | |
187.5 | 0.243 | 96 | 19.9 | 47 | 34.5 | n.a | n.a | n.a | n.a | |
155.5 | 0.168 | 99.5 | 14.3 | 47.5 | 39.6 | 90.3 | 8.7 | 205.3 | 3.9 |
MHD = mean heart dose; MLD = mean lung dose; VIB = volume of ipsilateral breast
Comparison between our and other IMRT data published in APBI studies
Target |
Ipsilateral breast (%) |
Non-target breast (%) |
Heart (left-sided lesions) |
Ipsilateral Lung |
|||||||
---|---|---|---|---|---|---|---|---|---|---|---|
VPTV (ccm) | VPTV / VIB | V95% (%) | V100% | V50% | V50% | MHD (cGy) | V5% (%) | MLD (cGy) | V30% (%) | ||
n.a | 0.17 | 99.4 | 27.2 | 50.3 | 33.3 | n.a | n.a | n.a | n.a | ||
236.2 | n.a | n.a | 17.3 | 45.3 | n.a | 35.9 | 0.4 | 123.6 | 1.4 | ||
123 | 0.21 | 96 | n.a | n.a | 37.5 | n.a | n.a | n.a | 9.7 | ||
243 | 0.14 | 99.1 | 16.2 | 39.9 | n.a | n.a | 14.6 | n.a | 4.6 | ||
187.5 | 0.243 | 88.8 | 9.4 | 42.1 | 28.1 | n.a | n.a | n.a | n.a | ||
155.5 | 0.168 | 99.8 | 12.1 | 39.9 | 29.6 | 103.6 | 17.5 | 272.1 | 3.6 | ||
236.2 | n.a | n.a | 18.2 | 44.9 | n.a | 54.4 | 1 | 148 | 2 | ||
176.4 | n.a | 98.9 | n.a | n.a | 19.7 | 100 | 4.3 | 190 | n.a | ||
155.5 | 0.168 | 99.8 | 12.7 | 31.6 | 19.8 | 198.1 | 38.9 | 364.7 | 6 |
MHD = mean heart dose; MLD = mean lung dose; VIB = volume of ipsilateral breast
The ratio of the target volume and the ipsilateral breast is a strict selection criteria in APBI, because good cosmetic outcomes are only achievable with the limitation of the irradiated volume. In our randomly selected 40 patients the median of tumour bed volume was 13.5 cm3 (3–40 cm3). Different values can be found in the literature with respect to tumour bed sizes. Vicini
In order to achieve a good cosmetic result, the dose to the ipsilateral breast must be kept under a limit. According to the NSABP B-39/RTOG 0413 protocol, the V50% value of the ipsilateral breast has to be less than 60%.17 However, in a retrospective study Jagsi
The coverage of the target volume was excellent with each technique. Based on our protocol, the V95% value of the PTVeval had to be higher than 99.5%, while the maximum dose should be less than 40.6 Gy. With 3D-CRT in case of 5 patients the coverage criterion was not fulfilled. We note that only Moon
Considering homogeneity, the best results could be achieved with the SW technique, and the difference is significant compared to the other three. With regards to the conformation number, the average value of the RA was the highest and was significantly better than for the other techniques, while in this respect the 3D-CRT was significantly the worst.
Regarding the mean dose of the ipsilateral lung in our study the 3D-CRT plans were significantly better compared to other techniques and at least as good as reported by others. As to volumes irradiated by higher doses, such as V30%, the RA resulted in the largest volumes, however these values were still in the range of the volumes published by others.
The best heart protection was achievable with 3D-CRT technique, especially in the low dose region, and our results are very similar to the data available in the literature. In this dose region (2 Gy or less) our results with intensity modulated techniques are slightly worse compared to data published by others, but well below the clinical limits. The outcome of static-field IMRT plans were better compared to RA. Other published data about dose to heart for left sided patients are also summarized in Table 3 and 4.
As for the contralateral lung and breast, each plan resulted in low doses. The 3D-CRT was significantly better as there are no fields intersecting with these organs. When applying intensity modulated techniques, only small portion of dose reaches the contralateral volumes, the average of maximum point doses varied between 2.5 and 3.5 Gy.
As for the PQI index, which takes into account the homogeneity, conformity and the dose to the organs at risk simultaneously, the two static-field IMRT techniques (SS and SW) achieved the best results; they were significantly better than the 3D-CRT and RA, but RA was significantly better compared to 3D-CRT. However, this index only takes into consideration plan quality and ignores factors, such as the overall number of MUs, treatment time or the couch rotation during deliveries of the beams.
With regard to the number of monitor units, the 3D-CRT and SS technique proved to be significantly better than the SW and RA (Table 1). The Table rotation applied with the 3D-CRT technique on one hand makes the treatment time longer and on the other hand can cause unintended movement of the patient, increasing the chances of intrafractional patient positioning errors, accordingly. Moreover, during planning it is difficult to take into consideration which gantry and couch angle combinations may result in collision. The RA technique provided the shortest overall treatment time, SW technique came very close to that and SS plans still could be delivered much faster than 3D-CRT plans.
The advantage of the 3D-CRT technique with respect to dose to majority of OAR arises from its non-coplanar tangential beams arrangement which in the same time causes its biggest disadvantage, the long treatment time with higher uncertainty of patient positioning. Owing to its high conformality, the RA technique minimizes the potential side-effects for the ipsilateral breast, provides the shortest treatment time, however, higher volumes of organs at risk are irradiated with low doses. The two static-field IMRT techniques have a relatively short treatment time, excellent homogeneity, good conformity and the doses to the organs at risk are well below the protocol constraints. The technique with the best outcome for a specific case can be anatomy, tumour bed shape, -size and –location dependent, but taking all aspects into consideration, the SW IMRT is our recommended technique for accelerated partial breast irradiations.