Long term radiographic findings of breast brachytherapy: Implications of surgical volume

Abstract

Purpose:

Little is known about long-term radiographic findings after treatment with accelerated partial breast irradiation (APBI).

Methods:

Univariate and multivariate analysis of factors leading to formation and resolution of seroma were performed in patients treated with lumpectomy and APBI.

Results:

Post-treatment images of 129 patients were reviewed by one radiologist. Median surgical excision volume was 108.9 cc (range 20.5-681.9). Primary mode of imaging was mammogram. Median time from end of RT to first and last surveillance image was 6 and 54 months, respectively. Median number of images was 7 (range 3-12). Seroma was identified in 98 (76%) patients, with median maximum diameter of 3.9 cm. Forty (41%) patients experienced resolution of seroma, at a median time of 29 months (range 6-74). On univariate analysis, surgical excision volume was associated with seroma formation, and tumor stage and margin re-excision were significant on univariate and multivariate analysis. No factors were associated with seroma resolution.

Conclusion:

Seroma formation after APBI resolves around 2.5 years for many patients, but persists for others possibly due to primary tumor and surgical excision volumes. With revised criteria on the definition of positive margins, smaller volumes may lead to decreased risk of seroma formation for future patients.

Introduction

While breast conservation therapy (BCT) with whole breast irradiation (WBI) has been accepted as the standard of care in the treatment of early stage breast cancer for several decades, shorter courses of focused radiation are being increasingly utilized.

Partial breast irradiation offers patients the convenience of a shorter course of therapy with good clinical and cosmetic outcomes. The final report of the American Society of Breast Surgeons (ASBS) MammoSite Breast Brachytherapy Registry revealed excellent clinical outcomes with 5 year recurrence rates of 3.8%, as well excellent/good cosmetic outcomes in 91.3% of patients.¹ The PROMIS experience at 10 years reported ipsilateral breast tumor recurrence of 7.6% in 1131 patients treated with APBI via multicatheter interstitial brachytherapy.² Despite these findings, acceptance, and utilization of accelerated partial breast irradiation (APBI) at academic institutions has been slow, where much of the data that is added to the literature is accumulated.³

APBI can be delivered via interstitial, device based, or external beam technique. Balloon based brachytherapy was an early mode of APBI delivery and has accumulated long-term safety and outcomes data. Seroma formation is known to be a common occurrence after APBI, with an incidence of 32-78% in various institutional series.^4–10 The ASBS MammoSite Breast Brachytherapy Registry reported that 29%, 13%, and 12% of patients were identified to have any seroma, symptomatic seroma, and seroma requiring drainage respectively.¹

While several retrospective institutional series have reported shorter radiographic findings, including seroma, calcifications, and architectural distortion, less is known about long-term radiographic findings. Additionally, many studies combine both radiographic and clinical diagnosis of seroma in the reported findings. The purpose of this study is to review long-term imaging findings of breast cancer patients treated with balloon based brachytherapy for in situ and invasive early stage breast cancer at one community academic cancer center. The goal is to report long term radiographic findings after breast brachytherapy, and to evaluate the impact of various patient, disease, and treatment factors on formation and resolution of seroma.

Material and methods

Records of patients treated with brachytherapy from 2005 to 2012 were reviewed after institutional review board approval was obtained. All patients were treated with segmental mastectomy followed by adjuvant APBI. The vast majority of surgical excisions were open cavity non-oncoplastic procedures (92%), and brachytherapy device placements were performed primarily by one surgeon. All patients had pre-treatment diagnostic imaging including a mammogram, as well a minimum of 2 years of surveillance imaging. Patients who had post-treatment imaging at outside institutions were excluded. Basic patient demographic, diagnostic, and treatment data were retrospectively captured.

All radiographic imaging was reviewed by one radiologist. Dates of initial presentation and resolution, as well as size of seroma were captured at each imaging time point. Most images were mammograms. However, if ultrasound or MRI studies were available, these were also used in conjunction with mammography for evaluation for seroma and other post treatment changes. Images were categorized within 3 months prior or after the dedicated timeframe of 6, 12, 18, 24, 36, 48, 60, and >60 months. Occurrence of skin thickening, presence of calcifications, and BIRADS classification were captured. Infection was divided into those that occurred prior to or during APBI versus those occurring more than 30 days after completion of APBI.

Balloon based catheters were placed days to weeks after surgical excision was completed. Most devices were MammoSite (Hologic Inc., Bedford, MA), but included a few Contura (Hologic Inc.). RT planning involved computed tomography (CT) based three-dimensional planning that occurred typically within 48 hours after placement of balloon catheter. Radiation (RT) planning was primarily performed by two radiation oncologists. Treatment planning was completed on Nucletron (Elekta Inc., Stockholm, Sweden). Radiation was prescribed 1 cm from the balloon surface to a dose of 34 Gray (Gy), administered over 5 days, twice daily. Planning target volume (PTV) was created using a uniform 1 cm margin around the surgical cavity site. Minimum balloon to skin distance was typically at least 7 mm for single lumen, but later accepted to be as little as 3 mm when multichannel devices were available. The device was removed after completing the prescribed dose. Prophylactic antibiotics were routinely prescribed to all patients. Adjuvant chemotherapy and hormonal blockade were delivered at the discretion of the respective medical oncologist.

Statistical analysis was performed using R software (Version 3.2.3). Univariate analysis of age, body mass index (BMI), log of surgical excision volume, log of largest surgical diameter to assess formation and resolution of seroma were performed using logistic regression. T stage, infection, and chemotherapy were analyzed using chi-squared test. Multivariate analysis of these factors was performed using logistic regression. Two sided P values of <0.05 were considered statistically significant.

Results

Records of 129 patients were reviewed. Median age at time of diagnosis was 62: 19% <50, 28% 51–59, 53% >60 years of age, and 87% of patients were Caucasian. Overall surgical stage was 0, I, and II in 26%, 63%, and 11% respectively. Pathologic and treatment characteristics can be found in Table 1. Median surgical excision volume including margin re-excision(s) was 108.9 cc (range 20.5-681.9). Final margin status was clear in all but 3 patients. Most patients were treated with MammoSite. RT dose was 34 Gy in 10 fractions delivered over a median of 8 calendar days. Median time from biopsy or segmental mastectomy to completion of RT was 1 month (range <1-4).

Table 1.

Disease and treatment characteristics

Histologic subtype
In situ	33 (25%)
Invasive	69 (53%)
In situ + Invasive	27 (21%)
Number of re-excisions
0	46 (36%)
1	52 (40%)
2+	31 (24%)
AJCC T stage
0	33 (25%)
1	81 (62%)
2	15 (11%)
Lymph Node evaluation	99 (77%)
Systemic therapy
Chemotherapy	25 (19%)
Hormonal blockade	93 (72%)
Receptor status
ER	103 (80%)
PR	89 (69%)
HER2	6 (5%)

Primary mode of imaging was mammography, with MRI, ultrasound, or MRI and ultrasound utilized in 7%, 38%, and 7% of patients respectively. Imaging included 870 mammograms, 105 ultrasounds, and 39 MRI. Median time from end of RT to first image was 6 months (range <1-13 months). Median time from end of RT to last scan was 54 months (range 21-106). Median number of images per patient was 7 (range 3-12). Median interval between images was 7 months (range 1-36). Seroma was identified in 98 (76%) patients. Time from end of RT to initial identification of seroma was 6.4 months (range 3-34). For all patients, maximal radiographic seroma diameter was 3.9 cm (range 0.4-10.3). Forty (41%) patients experienced resolution of seroma at a median of 29 months (range 6-74). In this cohort, maximal radiographic seroma diameter was 3.9 cm (range 0.4-8.3). Resolution of seroma at dedicated time periods can be found in Fig. 1. Resolution in patients with known imaging at designated time point can be found in Fig. 2.

Figure 1.

Patients with resolution of seroma over time (n = 40)

Figure 2.

Seroma resolution in patients with imaging available at selected timepoints

Radiographic skin thickening was noted in 14 (11%) patients, and benign calcifications were noted in 13 (11%) patients, which occurred with progression of time from completion of RT. BI-RADS was primarily 3, with overall trend of stability or improvement to BI-RADS2 with progression of time. Seventeen patients (13%) experienced infection at the treatment site, 10 occurring perioperatively to 30 days after APBI, and 7 occurring >30 days after completion of APBI.

On univariate analysis, tumor stage, surgical excision volume and need for re-excision of positive margins were significant for formation of seroma. Use of single versus multichannel device was not significant for formation or resolution of seroma. Tumor stage and need for surgical re-excision of positive margins were significant on multivariate analysis for formation of seroma. No factors were significant on univariate or multivariate analysis for resolution of seroma. Invasive or in situ pathology, BMI, chemotherapy, and infection were not associated with formation or resolution of seroma.

Discussion

Seroma formation and resolution

While much is known about the clinical outcomes of balloon based APBI, less is known about the long-term radiographic findings. Seroma is an expected radiographic event after APBI, with variably reported rates across single institution series, yet most studies combine clinical and radiographic reporting. In a study comparing WBI to APBI, a statistically significant difference in seroma formation rate of 18.5% versus 46.5% at 5 years was found. Seromas were either clinically or radiographically detected.⁴ Monticciolo et al reported a 69% versus 7% radiographic seroma rate in patients treated with MammoSite versus conventional external beam radiation, although median duration of follow-up is unclear in their report.⁸ In 38 patients treated with MammoSite APBI inserted at time of segmental mastectomy, Evans et al found that 76.3% had detectable seroma at 17 months and 68.4% persisting >6 months, with median seroma size of 4.4 cm. In this study, however, seroma was identified clinically or radiographically.⁷ In another study of 109 patients with median follow-up of 36 months, Watkins et al reported a seroma formation rate of 41%, with clinical significance in 13%.¹⁰ Chao et al reported a single institutional series with a seroma formation rate of 45%, although this was clinically or radiographically identified, with no significant changes note over time.⁵ The ASBS MammoSite Brachytherapy Registry identified a rate of 28.9%, 13.4%, and 12.2% rate of any seroma, symptomatic seroma, and seroma requiring drainage, with no specific criteria provided to sites for reporting of seroma events. Beyond 24 months, symptomatic seroma rates were 0.6%.¹ Our rate of seroma formation was higher than reported in many studies in the literature, which may be due to the use of MRI and ultrasound (when available) in conjunction with mammography for evaluation of post procedural seroma, as these supplemental modalities are more sensitive for fluid collection. Additionally, meticulous and methodical imaging review specifically for the purpose of this study likely increased the diagnosis of post treatment seroma, as registry studies do not have access to the actual radiological images, and clinical surveillance imaging is performed by multiple radiologists. Many studies co-report clinically and radiographically detected seromas, which makes interpretation of findings difficult. Our study reports seromas identified on imaging only. Additionally, we found that resolution of seromas occurred at a median of 29 months. This is similar to findings by Patel et al who found 65% of patients had resolution of seroma by 2 years on serial sonography.⁹

Factors influencing seroma formation

Volume of resection

Extent of surgical excision, timing of device placement, and infection control procedures have been associated with seroma in various institutional series. The general surgical literature has reported lower rates of seroma with decreasing excision volume.^11–14 Similar findings have been found in the APBI studies. Larger balloon volumes (>50 cm3) were associated with seroma formation by Chao et al.⁵ This is in line with our findings both on univariate and multivariate analysis. With guidelines issued by American Society of Radiation Oncology and the, Society of Surgical Oncology recommending against re-excision for when negative at inked margins, we may see a decrease in the frequency of seromas as size of resection cavities become smaller.¹⁵

Timing of catheter placement

The above mentioned reports by Evans, Watkins, and Chao et al found that timing of catheter placement was associated with seroma.^{5, 7, 10} Catheters placed at the time of surgical excision were more likely to lead to seroma formation than those placed after wedge resection. Our catheters were not placed at the time of surgical excision, yet a large percentage of our patients experienced radiographic seroma. To our knowledge, this has not been reported in the literature before.

Infection

There have been conflicting reports with regards to incidence of infection in APBI both in institutional series and larger population based reports. Monticciolo et al. reported a higher infection rate of 37% versus 20% in APBI versus conventional radiation.⁸ Earlier population and claim based retrospective analyses have identified increased risk of infection in patients treated with APBI in comparison to conventionally fractionated radiation.^{16, 17} However, a more recent analysis by Huo et al revealed no increase in the risk of infection from brachytherapy after 2007.¹⁸ This is likely due to the increasing use of prophylactic antibiotics and improved catheter infection prevention practices. In one institutional series, prophylactic antibiotics significantly reduced the risk of post-treatment infections from 37.5% to 6%. This is important, as they found infection to be associated with development of clinically significant seroma.¹⁰ Our infection rate is well within that reported in the literature, at 13%. We did not find infection to be associated with seroma in our series.

Other radiographic findings

Ahemd et al reported time to development of peak mammographic distortion was 21 weeks following MammoSite therapy in 22 patients with early stage breast cancer.¹⁹ Dragun reported that 74% of patients were found to have BI-RADS 3, with only 32.6% due to seroma in a cohort of 38 patients with 126 mammograms. Only 10 additional interventions required based on findings.⁶ Most images were BI-RADS 3 in our study, which remained stable, or improved to BI-RADS 2 with the progression of time.

Limitations

Limitations of our study are its retrospective nature, as we did not have various clinical factors, such as smoking, cardiovascular, or diabetes status that may have contributed to poor wound healing and possibly increased rates of seroma formation. Finally, dosimetric factors may have been implicated in resolution of seroma, but that data were not available for our analysis. As we move toward an era of patient reported outcomes, there would be benefit in identifying seromas that impact patient quality of life, rather than what is felt to be significant in the eyes of the treating clinician.

Conclusion

In patients who experience resolution of seroma, the time frame is typically around 2.5 years. We found that various factors pertaining to size of disease and excision impacted the formation of seroma. With new guidelines issued by several oncologic organizations recommending against re-excision when negative at inked margins, smaller volumes may lead to decrease in formation of seroma. Future results looking at other patient, provider, and treatment level factors could be beneficial in stratifying seromas into those that impact patient quality of life.