Journal of Cancer Research and Therapeutics Medknow Publications on behalf of the Association of Radiation Oncologists of India (AROI) ISSN: 0973-1482 EISSN: 1998-4138 Vol. 5, Num. 2, 2009, pp. 93-101

Journal of Cancer Research and Therapeutics, Vol. 5, No. 2, April-June, 2009, pp. 93-101

Original Article

Accelerated partial-breast irradiation vs conventional whole-breast radiotherapy in early breast cancer: A case-control study of disease control, cosmesis, and complications

Wadasadawala Tabassum, Sarin Rajiv, Budrukkar Ashwini, Jalali Rakesh, Munshi Anusheel, Badwe Rajendra

Depatment of Radiation Oncology, Tata Memorial Hospital, Mumbai
Correspondence Address:Tata Memorial Hospital, Senior Registrar, Department of Radiation Oncology, Dr. Ernest Borges Road, Main building, Room No. 115, Parel (East), Mumbai - 400 012
drtabassum2004@yahoo.co.in

Code Number: cr09025

DOI: 10.4103/0973-1482.52794

Abstract

Context: Accelerated partial-breast irradiation (APBI) using various approaches is being increasingly employed for selected women with early breast cancer (EBC).
Aims: To conduct a case-control study comparing disease control, cosmesis, and complications in patients with EBC undergoing APBI using multicatheter interstitial brachytherapy vs those receiving conventional whole breast radiotherapy (WBRT).
Settings and Design: Women with EBC fulfilling the American Brachytherapy Society (ABS) criteria were selected as 'cases' if treated with APBI or as 'controls' if offered WBRT during the period from May 2000 to December 2004.
Materials and Methods: APBI patients were treated with high-dose-rate brachytherapy (HDR) to a dose of 34 Gy/10#/6-8 days. WBRT was delivered to the whole breast to a dose of 45 Gy/25# followed by tumor bed boost, either with electrons (15 Gy/6#) or interstitial brachytherapy (HDR 10 Gy/1#).
Results: At the median follow-up of 43.05 months in APBI and 51.08 months in WBRT there was no difference in overall survival (OS), disease-free survival (DFS), late arm edema, and symptomatic fat necrosis between the two groups. However, APBI resulted in increase in mild breast fibrosis at the tumor bed. Telangiectasias were observed in three patients of the APBI group. The cosmetic outcome was significantly better in the APBI group as compared to the WBRT group (P = 0.003).
Conclusions: This study revealed equivalent locoregional and distant disease control in the two groups. APBI offered better overall cosmetic outcome, though at the cost of a slight increase in mild breast fibrosis and telangiectasias.

Keywords: Accelerated partial-breast irradiation, whole-breast radiotherapy, American brachytherapy society, early breast cancer, interstitial brachytherapy

Introduction

Materials and Methods

To ensure appropriate matching and for minimization of bias due to various confounding factors, this case-control study was restricted to women participating in an ongoing randomized trial evaluating preoperative single-dose hormone therapy. All women in this randomized trial who underwent BCT during the period from May 2000 to December 2004 and fulfilled the American Brachytherapy Society (ABS) criteria were taken as ′cases′ ( n = 27) if they were treated with APBI or as ′control′ ( n = 67) if they were advised WBRT.

Surgery
All patients (whether undergoing APBI or WBRT) underwent wide local excision of tumor with at least a 1cm margin all around and complete axillary clearance (level I, II, and III nodes). The incision was placed directly over the tumor site. The tumor base was excised up to the pectoralis fascia.

Radiotherapy
In the APBI group, multi-catheter interstitial implant was performed intraoperatively in 18 patients and postoperatively in the remaining nine patients [Table - 3]. Using rules of Paris System for volume implants, the implant was configured to irradiate the excisional cavity with an additional margin of 1 cm. None of the patients underwent re-excision for close, positive, or uncertain margins in the APBI group, while only two patients from the WBRT group had to undergo revision of margins.

All patients treated with WBRT were simulated in supine position, using a breast board for patient positioning and immobilization and a plaster-of-Paris (POP) cast for obtaining breast contour and various beam parameters for bitangential technique. The dose was prescribed at the isocenter for the tangential fields. Two-dimensional planning on Sunrise/ PLATO planning system was done for obtaining a uniform distribution of isodoses, such that 95% isodose would adequately cover the entire breast. Wedges were used as compensators in order to improve the dose homogeneity whenever necessary. The entire breast was treated to a dose of 45 Gy in 25 fractions over 5 weeks, followed by a boost to the tumor bed to a dose of 15 Gy in six fractions. Five patients received a boost to the tumor bed in the form an interstitial implant to a dose of 10 Gy single fraction [Table - 4]. RT was started 2-4 weeks after completion of systemic chemotherapy.

Adjuvant systemic treatment
Eighteen patients (66.7%) in the APBI group and 63 patients (94%) in the WBRT group received systemic adjuvant chemotherapy. A larger proportion of patients (74%; n = 20) in the APBI group received adjuvant hormonal treatment than in the WBRT group (67%; n = 45). The possible reasons for this difference in the use of adjuvant systemic therapy are discussed later in this article.

Follow-up strategy
Patients were seen every 6 months during the first 2 years and every 12 months thereafter both in the surgical as well as the RT clinic. Baseline mammography was performed 6 months after completion of radiotherapy and was repeated yearly thereafter. Cosmetic results were evaluated by the treating radiation oncologist using the Harvard criteria. ^[14]

Statistical analysis
Statistical analysis was performed using SYSTAT (SPSS, version 14). Analysis was done for local recurrences, regional recurrences, distant metastasis, disease-free survival, overall survival, cosmesis, and complications. The Kaplan-Meier method was used to evaluate the rates of ipsilateral local or regional failure, disease-free survival, and overall survival. The statistical significance of difference between actuarial curves was calculated using the log rank test. Tumor-related categorical variables were compared within the two groups using the chi square test and continuous variables were compared using Student′s unpaired t test. P ≤ 0.05 was considered statistically significant. All time intervals were calculated from the date of registration.

Results

Patient, tumor, and treatment characteristics
The details of patient, tumor, and treatment-related characteristics for both APBI and WBRT are presented in [Table - 3],[Table - 4],[Table - 5]. All patients fulfilled the ABS criteria for partial-breast irradiation, i.e., age ≥ 45 years, pathologic tumor size ≤3 cm, pathologically uninvolved nodes, clear margins, and no evidence of Extensive Intraductal Componenet (EIC). Although the same ABS criteria were used for both cases and controls, patients in the WBRT group had more unfavorable prognostic factors than patients in the APBI group,, more were Estrogen Receptor (ER) negative (71.1% vs 44.4%; P = 0.031), more had lymphovascular emboli (17.9% vs 3.7%; P = 0.038), and the mean pathological tumor size was slightly larger (2.47 cm vs 2.20 cm; P = 0.058). This difference in prognostic factors could have led to the increased use of chemotherapy in the WBRT group as compared to the APBI group (94% vs 66.7%; P = 0.001).

Outcome parameters
No local or regional recurrences were observed in both the groups, i.e., there was 100% locoregional control. No distant metastases were encountered in the APBI group, while the rate of distant metastases was 7.5% in the WBRT group. The disease-free survival and overall survival were 100% in the APBI group and 93.8% and 92.3% in the WBRT group respectively [Figure - 1] and [Figure - 2].

Cosmesis
The cosmetic outcome recorded at the time of the last follow-up was significantly better in the APBI group as compared to the WBRT group [Table - 6]. [Figure - 3] compares the percentage of ′excellent-good′ cosmesis (88.9% vs 56.0%) and ′fair-poor′ cosmesis (11.1% vs 44%) in the two treatment groups. There was a statistically significant difference between the two groups ( P = 0.003) favoring APBI. Examples of ′excellent′ and ′poor′ cosmesis are depicted in [Figure - 4] and [Figure - 5].

Complications
The difference in the incidence of arm edema and first symptomatic fat necrosis in the two groups was not significant ( P = 0.129 and P = 1.000, respectively). Of the two patients who developed symptomatic fat necrosis in the WBRT group, one patient received interstitial boost and another received electron boost to the tumor bed. Telangiectasias and mild breast fibrosis were significantly higher in the APBI group than in the WBRT group ( P = 0.025 and P = 0.000, respectively). However, it is notable that the fibrosis related to APBI was only grade I and limited to the tumor bed and did not significantly affect overall cosmesis.

Discussion

The concept of APBI for EBC has gained popularity over the past decade because of the attractive technical and logistic advantages. If it proves to be a successful strategy, this may mark a major shift after a century of adherence to the Halstedian paradigm of whole-breast treatment. ^[9] As with any new therapeutic approach, APBI cannot become a standard treatment option for selected patients until a sound radiobiological and clinical rationale can be defined for this approach and, most importantly, it can be backed up by mature data from well-conducted, large, randomized controlled trials (RCT). At present, only one well-conducted (though small) RCT has reported early satisfactory results with APBI. ^[15] The 5-year results of ongoing randomized trials from other large cooperative groups will be available only after 6-8 years. ^{[16],[17],[18],[19],[20],[21],[22]}

There are several differences in the profile of patients in this study (which is the only Indian study) as compared to patients in the studies of APBI reported from the West. The main differences are that, in our study, very few tumors were detected by screening or were < 1 cm in size, the histological features were less favorable, and there was a higher use of systemic chemotherapy. ^[23] To the best of our knowledge, no other Indian center has as yet initiated a systematic programme of APBI. There is a need to establish the safety and efficacy of APBI in selected Indian women who are potentially eligible for APBI as per the ABS and ASBS criteria but, in general, have a somewhat different profile as compared to the typical patients treated with APBI in the West. Ideally, this issue should be addressed through a well-designed and adequately powered randomized study conducted in India. However, to rule out a small difference (of a few percent) in disease control with APBI as compared to WBRT, we would need an RCT with over a thousand patients, as in most of the ongoing RCTs in the West. With the relatively small number of women who are potentially suited for APBI at our center, it is not feasible for us to conduct an RCT. Hence, in the Indian context, the best option at present is a prospective study comparing patients treated with APBI with matched controls who have been treated with conventional WBRT by the same set of clinicians and with comparable standards of pathology reporting during the same time period.

In the present study, we did not include the entire cohort of 160 women who had undergone APBI at our center during the period of 2000-2006 as the cases. The general principle in selecting controls is to select subjects who satisfy the inclusion criteria for the study similar to cases but were offered standard treatment. We decided to restrict our cases and controls to women who fulfilled all the ABS criteria but underwent either APBI (cases) or WBRT (controls) and were originally from a cohort of women participating in another ongoing randomized controlled clinical trial designed to assess the utility of single-dose preoperative hormone therapy in improving distant disease control and survival. The use of preoperative hormone injection was equally balanced between the groups ( P = 0.977). Before restricting our cases and controls to subjects from this study we made the important assumption that use of hormone injection prior to surgery is unlikely to significantly influence the incidence and pattern of breast recurrence, irrespective of the volume of breast irradiated. In order to ensure a reasonably good follow-up period, we further restricted this study to patients treated during May 2000 to 2004, and their records were updated in August 2007. The merit that we saw in selecting our cases and controls from among the women participating in this hormone therapy study was that it was a very tightly monitored RCT, with a good clinico-radiological follow-up and maintenance of detailed records of all disease- and treatment-related outcomes. By assigning all the 94 women who underwent BCT during the study period in this trial and fulfilled the ABS criteria as either cases ( n = 27) or controls ( n = 67), we have ruled out the possibility of any intentional or investigator bias.

In this study, by using the ABS criteria which takes into account several patient- and tumor-related parameters, we assumed that both cases and controls would be very well matched for these as well as other important parameters. However, the findings of our study highlight the inherent pitfalls of a case-control study. Despite all the cases as well as controls being women over the age of 45 years, with tumors less than 3 cm in size and uninvolved axillary nodes and surgical margins, a significantly higher proportion of controls had unfavorable pathological features (i.e., were ER negative or LVI positive, or had larger tumor size) and, as a result, there was greater use of chemotherapy in the control group. A lower disease-free survival in the control group, though not statistically significant, is perhaps a reflection of the unequal distribution of these prognostic variables. It is worth noting that the difference in the mean pathological tumor size between the two groups was merely 2.6 mm. Reducing this difference between the two groups would have required very precise matching (to the last millimeter), which would have been very impractical; the problem would have been compounded by the need to match for all the other histological parameters.

The other limitation of this study is the small sample size. This small sample size may have prevented us from confirming as real the important trends noted for the effect of several therapeutic and tumor parameters on disease outcome, cosmesis, and QOL.

Locoregional control
While the relatively small sample size of our study may make it unlikely that we will be able to identify the all-important slight worsening of tumor control with the experimental treatment (APBI), especially at longer follow-up, it is reassuring to note that at a median follow-up of 43 months, none of the 27 women had any recurrence. This is an indication of not only very appropriate selection criteria but also of the use of a good surgical and RT technique. In addition to applying the standard ABS selection criteria, we also excluded from this study those patients with EIC or lobular carcinoma or those with strong family history of carcinoma breast. It was important to exclude these patients as most of these characteristics predispose to multicentricity and increased risk of recurrence outside the tumor bed. However, when the use of APBI was first initiated at our institute, similar to other investigators, family history of breast cancer was not used as a criterion for excluding patients from APBI treatment,. But later, we sometimes excluded patients who gave a very strong family history of hereditary breast cancer as there is a known high predilection for multicentric involvement leading to higher ipsilateral breast recurrences in such cases. ^[39] In other studies that have used poor case selection or poor radiotherapeutic technique, unacceptably high breast recurrence rates were observed [Table - 7], and many of these recurrences were within the first 3 years of treatment. Our results with APBI are comparable with those reported from other institutions [Table - 8]. The case-control study reported by Vicini et al ., comparing 174 cases with 1388 matched controls, was able to reliably show the equivalence of APBI to WBRT for tumor control in their highly selected cohort. ^[23] The mean tumor size in their study was however only 11 mm, much smaller than the 22 mm in our cases.

Cosmetic outcome
[Table - 9] shows the cosmetic outcome and incidence of grade 3 or 4 late complications in the APBI group in our study as compared to the reports from other institutions. In the present study, the overall cosmetic outcome was better in the APBI group than in the WBRT group. This could be related to the late effect of RT to whole breast and the higher total dose delivered in case of WBRT as compared to APBI. For tumor control, if we use an α/β value of 10 Gy, the typical value for many tumors, in the linear-quadratic cell survival model, the biologic equivalent dose (BED) computed for the hypofractionated schedule (34 Gy/10#) would be substantially lower (46 Gy) than that for the standard treatments. ^[28] However, if we apply an α/β value for breast carcinoma of 4 Gy, as suggested by experiments involving irradiation of human breast cancer cell lines and clinical trials evaluating hypofractionated WBRT, a dose of 34 Gy given in 10 fractions of 3.4 Gy per fraction over 5 days is radiobiologically equivalent in tumor control to a dose of 50 Gy given in 25 fractions of 2 Gy over 5 weeks, which is the dose commonly used in studies of the National Surgical Adjuvant Breast and Bowel Project, i.e., it is identical to that of the standard treatment without a boost. ^[6],[7] At the same time, this hypofractionated scheme results in the same BED for late breast tissue complications (including fibrosis, erythema, and telangiectasia) as that of 60 Gy in 30 fractions, a regimen used at many institutions, including ours, to treat the tumor bed (46-50 Gy to the whole breast plus a boost of 10-14 Gy), which has been reported to have excellent cosmetic results.

The other factor which has the potential to worsen the cosmetic outcome is the use of adjuvant systemic therapy. Results from studies evaluating the effect of chemotherapy have reported that in women receiving doxorubicin-based chemotherapy there was a higher risk of fat necrosis, lower rates of excellent cosmetic results, ^[35] and the occurrence of acute radiation recall reactions. ^[27] The inferior cosmetic results in the WBRT group could be related to the higher proportion of patients receiving adjuvant chemotherapy in our study. As mentioned earlier, our small sample size may have failed to find any small, yet significant, impact of these factors on the cosmesis. Apart from these factors, numerous surgical and radiation factors are also known to affect cosmesis and should be considered while evaluating cosmetic outcome. ^[36]

Long-term complications
Breast edema is only rarely a long-term complication of RT and tends to resolve by about 12 months. It was not commonly seen in either of the treatment groups. The incidence of arm edema, breast symptoms, and symptomatic fat necrosis was similar in the two groups. Upper extremity lymphedema occurs when fluid accumulates in the interstitial space and causes enlargement with, usually, a feeling of heaviness in the limb. Axillary surgery contributes considerably to the incidence of lymphedema, with the incidence and severity of swelling being related to the number of lymph nodes removed. ^[40] Lymphedema after standard axillary lymph node dissection can occur in up to approximately 50% of patients. However, the risk of lymphedema is decreased substantially with newer sentinel lymph node sampling procedures, which is especially relevant for early tumors. ^[41],[42] Adjuvant RT to the breast or lymph nodes increases the risk of lymphedema, which has been reported in 9-40% of these patients. ^[43] Management of lymphedema requires a multidisciplinary approach to minimize the effect on the patient′s QOL. In the present study, though a greater number of WBRT patients developed mild to moderate lymphedema, the difference was not statistically significant. Due to complete avoidance of axillary RT as the present standard, the horrendous consequence of severe lymphedema is not seen in present-day practice.

Telangiectasia, a delayed skin sequela of RT, typically occurs late after RT, usually after about 18 months. But telangiectasia, unless extremely severe, is uncommonly the only cause of a poor cosmetic result. A higher incidence of telangiectasia has been reported with the use of MammoSite or interstitial brachytherapy in cases when skin-to-radioactive source distance was < 5 mm ^[44] In our study, telangiectasias were found in 11.1% of patients in the APBI group and in none in the WBRT group. These telangiectasias were seen only at the entry and exit sites of the catheter and not in any other part. As these were not severe, the overall cosmesis was not significantly affected and the difference between the two groups was not statistically significant.

The APBI group experienced greater grade 1 or 2 skin fibrosis as compared to the WBRT group (44.4% vs 8.7%; P = 0.000), as is to be expected of any brachytherapy procedure. However, the fibrosis was restricted to the tumor bed and there were no increased rates of grade 3 or 4 skin toxicity. When large fraction sizes are used, differences in normal-tissue radiosensitivity are likely to be magnified. There are currently no predictive markers to determine which patients will develop radiation-induced late toxicity. Li and colleagues detected a significant correlation between pretreatment plasma levels of tumor growth factor-β-1 (a multifunctional cytokine implicated in tissue fibrosis) and the risk of severe fibrosis among patients treated with BCT. ^[38] Other studies have revealed that specific polymorphisms of the tumor growth factor-β-1 promoter gene could be associated with the development of severe fibrosis. Hopefully, studies of ′radiation genomics′ may result in a panel of markers that can be used to prospectively detect ′fibrosis-prone′ individuals.

Conclusion

References

Copyright 2009 - Journal of Cancer Research and Therapeutics

The following images related to this document are available:

Photo images