After breast-conserving therapy (BCT) for the treatment of invasive breast cancer most ipsilateral local recurrences occur at or close to the original tumor site [, ]. Therefore, accelerated partial breast irradiation (APBI) has been studied as an alternative to whole breast irradiation (WBI) for patients with low-risk for local recurrence [, ]. With APBI radiotherapy (RT) is only targeted to the surgical cavity and a limited area of surrounding tissue instead of the whole breast. Also, a higher dose to a smaller volume can be delivered in a shorter period of time.
Several methods of APBI have been developed: the frequently used invasive techniques including intraoperative radiotherapy (IORT), interstitial or intracavitary brachytherapy and the non-invasive three-dimensional conformal external-beam radiation (3D-CRT). The invasive methods normally treat smaller volumes of normal tissue compared with 3D-CRT APBI, but a disadvantage of IORT is that pathology information at time of the irradiation is absent, including information about tumor free margins. Since the tumor is often not centrally located in the surgical resection specimen, this may lead to under- or overtreatment of the target area [, ]. Advantages of postoperative 3D-CRT APBI are the availability of pathology information at the time of treatment, optimization of the radiotherapy plan and more dose homogeneity, which can lead to less radiotherapy side effects including fibrosis. Furthermore, it is widely available, less physician-dependent, non-invasive and accessible to large groups of patients.
Several studies on postoperative 3D-CRT APBI showed variable results concerning toxicity and cosmetic outcome [, ,, , ]. The unfavourable results are associated with large treatment volumes. Large inter-observer variability in defining the tumor bed for postoperative irradiation has been shown in many studies, resulting in larger irradiated volumes [, , ].
In a previous study we showed less inter-observer variation in tumor delineation preoperatively compared to postoperatively . When RT is given preoperatively more accurate tumor delineation can be performed. After RT surgery is performed, thereby removing the area of the breast that received the highest RT dose. As a result, limited fibrosis and a good cosmetic result are expected.
In the multi-center phase II preoperative accelerated partial breast irradiation (PAPBI) trial (registered with ClinicalTrials.gov, Number NCT01024582) low risk patients of 60 years and older are irradiated preoperatively instead of postoperatively with external beam APBI. In total 120 patients will be included in the trial. In this paper we present the results of the interim-analysis of the first 70 patients. This interim-analysis was carried out in order to allow continuation of this trial, taking into account treatment complications, local control and cosmetic outcome. The main objective is to investigate the impact on local control, cosmesis and breast fibrosis of a short fractionated schedule given preoperatively.
Patients and methods
The PAPBI trial started in April 2010 at the Netherlands Cancer Institute-Antoni van Leeuwenhoek (NKI-AVL) in the Netherlands. Other institutes joined the trial later: Institut Gustave Roussy (IGR) in France, Karolinska Institutet in Sweden and University Medical Centre Utrecht (UMCU) in The Netherlands.
Patients were eligible for the PAPBI trial if they met the following criteria: age ⩾60 years; cT1–2 (tumor size ⩽3 cm) on magnetic resonance imaging (MRI); histologically proven adenocarcinoma; unifocal lesion on mammogram and MRI; pN0 (determined by a sentinel node procedure before start of RT); ECOG performance scale ⩽2. Patients were excluded in case of extensive microcalcifications on mammogram; ductal carcinoma in situ of the breast, without invasive tumor; lobular invasive breast cancer; treatment prior to radiotherapy; a planning target volume (PTV)/ipsilateral breast ratio >25%. Also, patients with a previous history of malignancy or synchronous malignant tumor in the other breast were excluded with the exception of pT1N0 contralateral breast carcinoma more than 5 years previously, basal cell carcinoma of the skin and adequately treated carcinoma in situ of the cervix. Written informed consent was obtained prior to start of the treatment protocol. The institutional review boards of the participating centers approved the study protocol.
Treatment and trial procedures
Prior to RT, tumor biopsies were taken for histopathological evaluation by a breast pathologist. At the time of the biopsies a marker was placed in the tumor, which was used for daily setup verification as well as guidance for the surgeon during the operation. RT consisted of 10 fractions of 4 Gray over 2 weeks. The gross target volume (GTV) was defined as the visible tumor on the planning computed tomography (CT) scan using additional information from MRI, mammogram and ultrasound. The clinical target volume (CTV) was constructed by extending the GTV with 2 cm. The PTV was constructed by extending the CTV with 0.5 cm. Dose distributions were planned according to the International Commission on Radiation Units and Measurements recommendations. Position verification was performed by cone-beam CT-scan or Electronic Portal Imaging according to the institutes’ protocol. Conformal 3D CRT, Intensity Modulated Radiotherapy (IMRT) or Volumetric Modulated Arc Therapy (VMAT) were used as RT techniques, depending on the institute. Six weeks after the last day of RT, a wide local excision was performed. In case of positive resection margins a re-excision was performed. Adjuvant systemic treatment was applied according to local guidelines in the participating institutes (based on tumor size on MRI and histological grading of the tumor on biopsies). To evaluate response to RT, MRI was performed prior to the start and six weeks after completion. A 18F-fluorodeoxyglucose positron emission tomography (PET)-scan was optional and was performed at the same time as MRI.
The primary endpoints were local recurrence, breast fibrosis and cosmetic outcome. Local recurrences should not exceed 4% at 5 years of follow-up.
An additional goal was to develop a classifier that predicts radiosensitivity. For this translational research gene expression profiling of the tumor will be correlated to response of the tumor as measured by pathologic response, MRI and PET-scan 6 weeks after irradiation.
The calculation of the number of patients to be included for the full phase II trial was based on the amount needed to develop a gene expression signature for radiosensitivity. It was estimated that 60 patients were needed in the training set. The classifier was considered to be clinically useful if the positive predictive value was >90% and the negative predictive value was >50%. Based on the assumption that the proportion of responders and non-responders would be close to 50% the number of patients needed in the validation set was 50. In total 120 patients will be included in the trial.
During RT acute skin toxicity (up to 1 month after treatment) was scored by the attending radiation oncologist or trained physician assistant according to the EORTC/RTOG acute radiation morbidity scoring scale. After RT, patients underwent clinical evaluation every 3 months in the first year. Thereafter, follow-up visits were conducted every 6 months and annually after 5 years. Patients underwent a screening mammography annually and in the first year of follow-up with an additional MRI. Postoperative complications were scored for all patients.
At each visit clinical induration was scored separately for the tumor bed area and the whole breast. Induration was used as a surrogate of the effect after radiotherapy and surgery (edema and fibrosis) [, ]. After 24 months induration was considered to be fibrosis. Also breast pain, rib pain and the presence of rib fracture(s) were evaluated according to the EORTC/RTOG/late radiation morbidity scoring.
Cosmesis was evaluated by the physician before start of RT, 6 months after completion of treatment and thereafter every year. The treated breast was compared to the contralateral breast for visible sequelae, telangiectasia and global cosmetic result scored as (0) excellent, (1) good, (2) fair, (3) poor or (4) not evaluable. Dutch patients received a questionnaire concerning cosmetic outcome yearly. Digital photos of both breasts were taken before the start of RT, 6 months after completion of treatment and thereafter every year. The photos will be used in the future for objective assessment of cosmetic outcome by the BCCT.core project software program.
Time intervals were calculated using the first day of RT as treatment day 1. Descriptive analyses were used to show the proportion of patients with grade 0, grade 1, ⩾grade 2 events at each follow-up visit.
Patients who underwent a mastectomy due to a local recurrence or incomplete resection were excluded from the cosmetic analysis after the mastectomy.
A total of 70 patients, treated between April 2010 and December 2013, were evaluated from all participating centers; NKI-AVL (n = 47), IGR (n = 14), Karolinska (n = 6) and UMCU (n = 3). The median follow-up was 23 months (range 3–44). The mean tumor size was 1.5 cm (0.4–3.2). All patients were node negative, determined by a sentinel node procedure before RT. At histological examination of the pre-treatment biopsy, ninety-four per cent of the tumors were classified as invasive ductal carcinoma (IDC), and the majority as histological grade 1–2 (96%), which was estrogen receptor (ER)/progesterone receptor (PR) positive and HER2 negative (Table 1).
|Variables||Number (n = 70)|
|Median age (years)||67.3 (60–80)|
|Median tumor size (cm)||1.3 (0.4–3.2)|
|Histological grade biopsy|
The mean GTV was 2.7 cc (SD 2.0, range 0.1–12.0), the mean CTV was 87 cc (SD 32, range 7–193) and the mean PTV was 121 cc (SD 42, range 13–263). The mean whole breast volume was 1028 cc (SD 474, range 107–2265) resulting in a PTV/ipsilateral breast ratio of 13.3% (SD 4.9, range 5–24). None of the 70 patients interrupted or stopped the treatment. 69 Patients underwent a wide local excision with negative resection margins. One patient had a focally positive resection margin. A subsequent mastectomy followed because a poor cosmetic result was expected with a re-excision. 46/70 patients (66%) received adjuvant hormonal treatment and 4/70 patients (6%) adjuvant chemotherapy.
Treatment-related toxicities and complications
Acute skin toxicity was scored for all 70 patients; 39 patients (56%) had no acute toxicity, 30 patients had grade 1 (43%) and one patient (1%) grade 2 acute skin toxicity. Postoperative complications were noted in 11 of the 70 patients (16%); two patients had direct post-operative bleeding requiring re-surgery performed on the same day, one patient developed a hematoma two months after surgery and needed re-surgery, eight patients developed a postoperative wound infection, one of whom needed re-surgery for a wound abscess and in one patient a small fistula was formed which closed within ten months after treatment with antibiotics. The other six patients were successfully treated with oral antibiotics. No other wound healing problems were observed. Seven patients (10%) developed persistent seroma (see Fig. 1).
In four patients transient edema in the whole breast was observed, in all other patients the breast tissue outside the irradiated tumor area showed no side effects. In the first year an increase of induration was seen at the tumor area: from 52% (31/60 patients) to 69% (41/59 patients) to 80% (40/50 patients) after 3, 6 and 9 months respectively. Over time the proportion of patients with (any grade of) induration-fibrosis declined as well as the grade (Fig. 2 and Supplemental Table 1). At 12 months induration was scored for 57 patients; 11 (19%) had none, 40 (70%) mild and six (11%) moderate induration. At 24 months 19 of 41 patients (46%) had mild fibrosis and one (2%) patient moderate fibrosis. At 30 months the majority of patients (15/23 patients) had no fibrosis (65%). After 36 months fibrosis was scored for only 11 patients, all had none-mild fibrosis. The area of fibrosis was limited to a clinically estimated volume of 1–2 cm.
During total follow-up, 27 of 70 patients (39%) reported grade 1 breast pain (transient in 21, persistent in six patients) and seven patients (10%) grade 2 breast pain. In one patient the breast pain diminished from grade 2 to grade 1 and in five patients the pain was transient (grade 0). One patient reported grade 3 breast pain, which had diminished to grade 2 at last follow-up date. 11 patients (16%) reported rib pain, of whom nine scored grade 1 (six transient) and two patients grade 2 (persistent). No rib fractures were observed.
Cosmesis, as scored by the physician, improved over time (Fig. 3 and Supplemental Table 2). After 0.5 years global cosmetic outcome was scored for 66 patients; 51 (77%) had a good to excellent, 13 (20%) a fair and two (3%) a poor cosmetic result. After 1 (n = 56), 2 (n = 41) and 3 years (n = 13) the proportion of patients with a good to excellent cosmetic outcome was 89%, 88% and 100% respectively. The cosmetic assessment by the Dutch patients (n = 50) was also evaluated (Fig. 4 and Supplemental Table 3). The majority of patients was satisfied to very satisfied with the cosmetic result; 81%, 86%, 80% and 79% after respectively 0.5, 1, 2 and 3 years. Examples of cosmetic outcome are shown in Fig. 5.
In two patients an ipsilateral tumor recurrence (IBTR) was diagnosed. In both cases the primary tumor was grade 1, ER/PR positive, HER2 negative; in one patient the IBTR was diagnosed at the skin entry of the biopsy tract after 12 months, outside the radiation field; one patient was diagnosed with an IBTR in another quadrant of the breast after 26 months. For both patients, the IBTR was outside the radiation field and loss of heterozygosity-analysis showed true recurrences. No distant failure has occurred. One patient died due to pneumonia, not related to the breast cancer or PAPBI treatment.
This is, to the best of our knowledge, the first study of preoperative as opposed to postoperative APBI in low risk breast cancer patients. Seventy patients treated with preoperative APBI show low complication rates, limited fibrosis in a small volume and good to excellent cosmetic results with an acceptable local recurrence rate.
Despite preoperative irradiation, our postoperative complication rate compares favourably to conventional radiotherapy with modern techniques. In the Cambridge IMRT Trial, 648 patients were treated with BCT of which 19.7% had a postoperative infection and 7.9% postoperative hematoma , although lower infection rates have been reported in brachytherapy trials. We observed an overall postoperative infection rate of 11% and a 5.7% postoperative hematoma rate.
The majority of the patients in the PAPBI trial had no or only mild fibrosis. During the trial it was noted that the volume of fibrosis was very small, beyond expectation. The volume was clinically estimated 1–2 cm in most patients, generally at the boundaries of the previously irradiated area, suggesting a leftover of the irradiated volume. After conventional BCT with WBI, fibrosis rates are higher; in the EORTC 22881-10882 boost–no boost trial the 10 year rate of moderate or severe fibrosis increased from 13% to 28% with a boost . These rates are similar to the study of Hepel et al., which showed moderate to severe fibrosis in 25% of the patients after a median follow-up time of 15 months in a study of 64 patients . The presence of fibrosis was the strongest correlate of a fair/poor cosmetic outcome. The results concerning toxicity of other external beam APBI studies are varied. Several studies show limited toxicity, comparable with our results [, , ,, ]. In the multicenter study of Berrang et al. with a median follow-up of 3 years, grade 2 fibrosis was the most prominent late effect and was observed in 7 of 87 patients (8%) .
The global cosmetic outcome, scored by the physician, was good to excellent in 77% at 0.5 years and 100% at 3 years in our patients. Cosmesis did not deteriorate since treatment with a tendency to improvement over time, in contrast to other external beam APBI trials where cosmetic outcome gets worse with longer follow up [, ]. Results concerning cosmetic outcome in other trials, concerning postoperative RT, range from good to excellent cosmetic outcome in 97% of the patients  to 21% unacceptable cosmesis at a median follow up of 2.5 years . Most published results involve single-center studies with small numbers of patients and short follow up. Larger trials include the RAPID trial, which showed significantly more poor cosmetic results at 3 years in the 3D-CRT APBI group compared to WBI (29% versus 16.5%) . The volume of breast that receives 95% of the prescribed dose was restricted to <35% in that trial. It was speculated that this volume may have been too large in some patients. Hepel et al. and Jagsi et al. have also linked their poor cosmetic results to the large volumes of breast tissue receiving relatively high irradiation doses [, ]. In the trial of Hepel et al. a mean PTV/ whole breast volume of 18% in patients with excellent/good cosmesis and 24% in patients with fair/poor cosmesis was seen.
It is expected that preoperative external beam APBI results in smaller treatment volumes compared to postoperative APBI because of more accurate target definition. The mean PTV/breast ratio in the PAPBI study was 13%. Compared to the study of Hepel et al. the mean tumor size in the PAPBI was larger (1.5 cm in the PAPBI versus 0.9 cm in the study of Hepel et al.). Although similar CTV and PTV margins were used, we found significantly smaller PTV volumes in our preoperatively treated patients (mean PTV 122 cc versus 296 cc in the Hepel et al. series). This was also seen in a study of Palta et al. . A virtual plan for preoperative single fraction external beam APBI resulted in a substantial reduction in ipsilateral breast tissue dose compared with postoperative APBI. Also Nichols et al. showed that preoperative APBI could decrease the planning target volume and the dose to normal tissues compared to postoperative APBI in a cohort of 40 patients . Our good cosmetic results are likely to be explained by the small irradiated volume (mean PTV 122 cc) and the surgical removal of the breast tissue that received the highest radiation dose, with very limited fibrosis as a result.
Comparison between studies can be difficult because varying toxicity and cosmetic result scoring systems and dose and fractionation regimens are used. For instance, we applied an RT scheme of 10 times 4 Gray once a day while other external beam APBI trials used schedules with RT twice a day. Furthermore, dosimetric constraints differ between studies. For most patients in the PAPBI trial IMRT is used, while in other APBI trials most patients were treated with conformal 3D-CRT.
The treatment efficacy results of postoperative external beam APBI phase I and II studies are limited and overall show low local recurrence rates after limited follow up time [, , , , ]. Most long-term data concerning APBI originate from trials using brachytherapy. In the study of Polgar et al. equal local control for multicatheter brachytherapy or electron beam irradiation after 10.2 years was observed [, ]. Other trials that randomised between external beam RT and IORT showed higher IBTR rates in the IORT group [, ]. In our PAPBI trial, two patients had an IBTR. Longer follow-up is needed to conclude more about breast relapse rates in our and other external beam APBI studies. Therefore, adequate patient selection remains important [, ].
Limitations of the current study are the non-randomized set-up and the relatively small sample size. As a consequence of the low number of patients and incomplete follow-up the progress of cosmesis and fibrosis must be carefully interpreted. A clinical disadvantage of preoperative compared to postoperative APBI, is that a sentinel procedure has to be obtained before RT, requiring an additional surgical procedure.
The first results of this preoperative APBI trial show low complication rates, limited induration-fibrosis in a small volume and good to excellent cosmetic results. Preoperative APBI is a feasible and widely available technique with promising results for low risk breast cancer patients.
This study was financially supported by the Dutch Cancer Society-KWF Grant NKI 2009-4389 and by the Cancer Society in Stockholm.
Conflicts of interest
The authors have no conflicts of interest to declare.
We thank Erik van Werkhoven for collecting and providing data.