Abstract
Background
Early detection of recurrence improves the survival rate of patients treated with breast conservation therapy (BCT). Therefore, ultrasonography (US) may be useful for metachronous ipsilateral breast tumor recurrence (MIBTR) obscured on mammography by dense breast tissue and distortion.
Purpose
To evaluate clinical, radiologic, and pathologic findings of MIBTR retrospectively, and to assess the role of surveillance US additional to mammography for MIBTR detection.
Material and Methods
During 2000 to 2012, 28 MIBTR were collected and reviewed among 2958 women treated for primary breast cancer with conservation surgery. The detection rates of imaging studies for identifying metachronous ipsilateral lesions were assessed and compared. MIBTR tumor staging was evaluated according to imaging modality for detection of MIBTR, palpability, and recent imaging surveillance.
Results
No significant difference was observed in the detection rate between mammography and US for overall MIBTR (84.2% vs. 85.7%; P = 0.898) or non-palpable MIBTR (88.2% vs. 81.0%; P = 0.566). US alone identified 33.3% of non-palpable MIBTRs (seven of 21). Among these cases, two had negative mammograms. All 14 MIBTRs with recent imaging surveillance were stage T2 or less, and all seven MIBTRs detected by US alone were in situ or T1; 33% of MIBTRs without recent imaging surveillance were T3 or T4.
Conclusion
The overall MIBTR detection rate by US was not higher than the detection rate of mammography, although combined surveillance with US and mammography found MIBTRs slightly earlier than mammography alone.
Keywords
Introduction
As breast cancer diagnosis improves and longer survival rates are achieved, there has also been increasing interest in improving the quality of life of breast cancer survivors. Breast conservation therapy (BCT) is associated with a high quality of life and minimizes the amount of removed breast tissue (1). Patients who receive BCT can have survival rates comparable to patients who receive mastectomy. BCT surgically excises breast tumors leaving a margin of normal breast tissue, followed by moderate-dose, whole-breast radiation therapy (2). Currently, breast conservation surgery with radiation therapy is widely used as the treatment of choice for women with early stage breast cancer based on randomized controlled studies showing similar long-term survival rates for BCT and radical mastectomy (1,3,4).
However, local recurrence occurs in approximately 1–2% of BCT patients per year during a 10-year period after initial treatment (3,5) and early detection of breast tumor recurrence is an important prognostic factor for better survival of patients treated with BCT (6,7). Imaging is useful for the early detection of non-palpable ipsilateral breast tumor recurrence (6). However, diagnostic methods and examination intervals vary by institution (3) and surveillance protocols are not well established.
Surveillance mammography is widely used and is effective for the early detection of metachronous ipsilateral breast tumor recurrence (MIBTR) (5,6). Approximately 50% of local recurrences after BCT are detected by mammography (8). However, the residual breast tissue after BCT is less sensitive to examination than untreated breast tissue, because BCT distorts the breasts and results in loss of symmetry (7). Therefore, mammography misses half of the existing MIBTRs. Moreover, hematomas, fat necrosis, or scar formation can increase local breast density and thicken skin thickness, creating abnormalities on mammography (3,8). These abnormalities are similar to mammographic features of local recurrence, presenting another challenge in the interpretation of mammography after local surgery and radiotherapy (3,8). MIBTR can also be obscured on mammography by dense breast tissue (9). Therefore, additional imaging techniques are needed. The most commonly used additional imaging techniques are ultrasonography (US) and magnetic resonance imaging (MRI) (10). Compared with MRI, breast US is widely accessible, relatively inexpensive, and does not require injection of a contrast agent (10). Recently, breast US was reported to be useful in the evaluation of the chest wall and axilla, which cannot be visualized by mammography (11,12). The purpose of this study was to retrospectively compare clinical, radiologic, and pathologic findings of MIBTR and compare surveillance US to mammography in the detection of MIBTR.
Material and Methods
This retrospective study was approved by the Severance Hospital Institutional Review Board and informed consent was waived.
Patient selection
From the database of a breast imaging center at Severance Hospital, 7706 consecutive examinations of 2958 women from January 2000 through May 2012 with a history of BCT for primary breast cancer were analyzed. Among 2958 women in the database, 138 received imaging-guided biopsy (US-guided in 134 and mammography-guided in 10) for 144 ipsilateral breast lesions. Lesions were considered metachronous if they were diagnosed 6 months after BCT. Among 144 ipsilateral lesions, 143 breast lesions in 137 patients were included with the exclusion of one synchronous lesion that was found within 6 months of surgery, and we retrospectively reviewed the clinicopathologic records of 143 metachronous ipsilateral breast lesions. Among 143 lesions, 115 (80.4%) were benign and the remaining 28 (19.6%; mean size, 19.0 mm; range, 4–100 mm) were malignant, which were included in the analysis.
Imaging surveillance
Mammography was performed using dedicated equipment (DMR; GE HealthcareA, Milwaukee, WI, USA) before April 2005 and a full-field digital mammography system (Selenia; Lorad/Hologic, Danbury, CT, USA) from May 2005 onward. Standard craniocaudal and mediolateral oblique views were routinely obtained and additional views were obtained as needed.
US was performed using a sonography unit (ATL HDI 5000 or 3000; iU22; Philips Medical System, Bothell, WA, USA) with a 5–10-MHz or a 5–12-MHz linear-array transducer to interpret both breast US and mammography at the time of this study by one of nine radiologists with a range of 1–15 years of experience in performing breast sonography, who performed 200–300 breast ultrasounds per month. Bilateral whole-breast US was performed with patients in the supine position with arms raised. Scanning was of the ipsilateral breast first and then of the contralateral breast for each examination. Examinations took approximately 10 min (range, 5–25 min). US was performed on the same day the mammogram was reviewed, and the mammogram and US were interpreted together.
Following the protocol of our institution, after BCT, follow-up clinical examinations were performed every 6 months for the first 2–3 postoperative years and annually thereafter. Patients with previous BCT were recommended for annual mammography and bilateral whole-breast US every 6 months during the first 2 years and annual mammographic and sonographic evaluations thereafter. Imaging evaluation intervals were somewhat variable because of examination schedules and patient preference.
Sonographic and mammographic findings were classified according to the American College of Radiology Breast Imaging-Reporting and Data System (BI-RADS) lexicon (13). BI-RADS categories for mammography and US were used in the original radiology reports. BI-RADS US lexicon did not exist before 2003 but sonographic findings at our hospital had been categorized according to breast cancer risk similar to mammographic BI-RADS categories since 1999.
Data analysis
We evaluated clinical, imaging, and pathologic findings of included patients to assess metachronous ipsilateral cancer detection methods. Reference standards were based on specimen pathologies from core biopsies or surgery. For comparison with metachronous lesions, we defined index cancers as previously treated cancers. We assessed detection according to the parenchymal pattern on mammography and our study population was divided according to parenchymal pattern. The fatty breast category had patients with fatty breasts or scattered fibroglandular tissues (BI-RADS pattern A or B). The dense breast category had patients with heterogeneously or extremely dense breast tissue (BI-RADS pattern C or D). Cancer staging was compared according to the examination modality used to detect metachronous ipsilateral breast cancers. To determine breast cancer size, masses similar to cancers were evaluated with only calcifications being excluded. The size of the lesion was measured on US. Lesion size and staging were compared between palpable and non-palpable MIBTR. We also analyzed whether patients with MIBTR underwent recent imaging surveillance (surveillance performed in 1 year or less,) before MIBTR diagnosis.
The imaging detection rate for metachronous ipsilateral lesions was assessed and clinical, mammographic, sonographic, and pathologic findings were compared. Statistical comparisons used the generalized estimating equation to compare the diagnostic performance of imaging studies by sensitivity and specificity and the McNemar test to compare index cancer stage and metachronous ipsilateral breast cancer stage. The Chi-square or Fisher’s exact tests were used for other non-parametric variables and the Student’s t-test and Mann–Whitney test were used for parametric inference. Statistical significance was assigned to P values less than 0.05. All statistical analyses used the SAS system for Windows (version 9.2, SAS Institute).
Results
The included 28 MIBTRs in 28 patients (mean age, 49.2 years; age range, 30–76 years) were 20 invasive ductal carcinomas, seven ductal carcinomas in situ (DCIS), and one adenoid cystic carcinoma.
The mean interval from diagnosis of the first cancer to detection of the MIBTR was 69.1 months (range, 6–264 months).
Clinical, imaging, and pathologic findings
Clinical and pathologic findings of 28 metachronous ipsilateral breast cancers.
Breast Imaging-Reporting and Data System.
Fatty breast means ACR A and B and Dense breast includes ACR C and D.
According to the American Joint Committee on Cancer.
All patients underwent surgery. The mean pathologic size of the MIBTR was 19.0 mm (range, 4–100 mm). The median pathologic size of non-palpable MIBTRs was 10 mm (range, 4–100 mm) and the median pathologic size of palpable MIBTRs was 20 mm (range, 6–30 mm). Among 28 patients, 19 women underwent mammography, all at the same time as US.
Detection rate by imaging modality
Twenty-eight breast cancer staging at MIBTR diagnosis according to the detection mode.
According to the AJCC.
N staging according to the American Joint Committee on Cancer.
A 55-year-old woman with breast conservation surgery of the left breast 4 years ago. (a, b) Mediolateral-oblique (a) and craniocaudal (b) mammogram shows skin folding due to a postoperative scar that has shown no interval change during serial previous mammograms. (c) Sonogram shows a 7-mm oval, hypoechoic nodule with a microlobulated margin at the upper outer quadrant of the left breast. It showed invasive ductal carcinoma at core-needle biopsy. At subsequent surgery, the nodule was staged pT1N0M0.
Pathologic stages of metachronous ipsilateral breast tumoral recurrences according to recent imaging surveillance.
Performed in 1 year or less.
T&N staging according to the AJCC.
Three of four patients had negative mammographic results 6 months before MIBTR diagnosis.
The cancer staging of the MIBTRs according to the American Joint Committee on Cancer (AJCC) was compared with index cancers in 28 patients (14). Although the index cancers and MIBTRs showed no significant differences in staging (P > 0.05), all MIBTRs detected by US alone and 73.9% of all MIBTRs were designated with the same (n = 10) or earlier stage (n = 7) than the index cancers.
Discussion
Postoperative local tumor recurrence occurs within 10 years after primary cancer treatment in approximately 1–2% of patients receiving BCT (3,5). Recht et al. showed that MIBTR risk peaks at 2.5% per year between 2 and 6 years after treatment and then decreases thereafter (15). However, women remain at risk even 10 years after therapy (15–17). Our study showed similar results to this study. MIBTR tended to develop in 5 years or less after treatment (60.7%, n = 17; Table 1). Of these women with MIBTR, 10 (58.8%) were diagnosed within 3 years.
Over a period of 12 years at our institution, the cancer detection rate was 0.9% per patient (28 of 2958) and 0.4% per US examination (28 of 7706 US examinations), which was lower than previously reported (3.0–4.7% for 5–6-year follow-up) (3,5). Our study might have had a lower local recurrence rate in BCT patients for several reasons. First, our patients usually underwent preoperative ultrasound staging in addition to mammography, which could have resulted in the low prevalence of local recurrence. Second, the study period was 12 years, but the entire population was not followed for the full 12 years. The study period was based on when ultrasound surveillance was performed at our institution and not all patients underwent ultrasound surveillance for 12 years. While some patients did undergo repeated US for 12 years, others were evaluated for just a certain portion of the study period. Finally, our study population was all Asian women, who have small breast volumes (18). Therefore, BCT was tried only in cases with strictly determined localized disease. In 2000, only 27% of patients with primary breast cancers underwent BCT in Korea, which is lower than the percentage performed in Western countries (19). Our low prevalence of local recurrence was consistent with the rate reported by Buist et al. (20). Although their study did not include statistical analysis examining the difference in local recurrence prevalence according to race or ethnicity, the local recurrence in Asian women with a history of breast cancer surgery was reported to be the lowest among several ethnicities: less than half the recurrence of white women and one-quarter of the recurrence of black women.
Additional imaging modalities or intensive surveillance have been considered as adjunct imaging to mammography because postoperative distortion and density superimposed by postoperative changes such as hematomas and scar formation can obscure or mimic local recurrence on mammography. US was expected to be a useful adjunct to mammography for detecting non-palpable tumoral recurrences in women with dense breast parenchyma, similar to its usefulness in screening dense breasts. However, in this study, the detection rate of MIBTR by US was not higher than the detection rate by mammography (P = 0.898) and was lower than mammography for non-palpable MIBTR although differences were not significant (P = 0.566). In a study by Kim et al. (7), the detection rate of US for ipsilateral tumoral recurrence was around 85% and it was about 10 percentage points lower than the detection rate for contralateral breast cancer. This finding supports the hypothesis that detection by US can be affected by postoperative changes. The sensitivity of surveillance mammography was reported to be 64–67% for detection of ipsilateral tumoral recurrence (8) and the detection rate for dense breasts by mammography in our study was higher than in other studies (7). This finding affected the comparison of detection rates between mammography and ultrasound. Ultrasonograms and mammograms were read at the same time in our study. Therefore, mammography interpretations might have influenced the US findings.
Non-palpable local recurrence detected by US alone was 25% of all MIBTRs (7 of 28; Table 2) in our study. However, of these, only two lesions were confirmed as negative by mammography. The remaining five were not examined by mammography due to the absence of a confirming mammogram at the time of diagnosis. Therefore, in these cases, MIBTR was confirmed by US-guided biopsy after US surveillance and determined to be malignant. Mastectomy was recommended to these patients and mammograms were inevitably not performed because the treatment option was already determined. According to our institution protocol, US was recommended every 6 months for patients with a history of BCT and annual mammography. Every other US was done without mammography. In 14 patients with previous recent imaging surveillance, seven lesions were identified by US alone. Of these, six lesions (42.9%) had a negative mammogram at the time of MIBTR diagnosis (n = 3) or 6 months before (n = 3). Those six lesions had all N0 status with T0 or T1 stage, while MIBTRs with positive mammograms were stage T2 in 28.6% (2 of 7) of patients with lymph node metastasis in 14.3% of cases (1 of 7, Table 3). Although these data did not show significant differences due to the limited sample size, our results supported the detection of earlier-stage MIBTR by US surveillance since cancer stage is considered a surrogate marker for overall survival (20). Our results were consistent with the small tumor size of mammographically occult breast cancers detected by US in previous studies (9,21). We analyzed the differences between patients who underwent US surveillance 6 months or 12 months prior to MIBTR diagnosis.
Although NCCN guidelines recommend only annual mammography for patients who received BCT (22), imaging surveillance can be done in several ways and protocols may vary by institution even in a single country, such as mammography being recommended every 6 months for 2–5 years after BCT (23). The benefits of intensive surveillance such as 6-month follow-up mammography have been debated (5,24). Arasu et al. reported that biannual mammography is preferable for detecting ipsilateral local recurrence at an earlier stage than noncompliant annual surveillance, similar to our results (5). However, we detected no difference in tumor size. Although we found more and smaller MIBTRs with US, the number of MIBTRs detected was limited (cancer detection rate, 0.4% per US examination) with a large number of US examinations over 12 years. The overall detection rate of US was not higher than the detection rate of mammography, which is not sufficient evidence to support intensive surveillance with US, even at 6-month intervals.
This study had some limitations. First, the design was retrospective and did not include a large number of MIBTRs, although data were collected over 12 years. Second, long-term follow-up was not evaluated for survival benefits. Therefore, future prospective studies are required with a larger group of patients. Also, we did not investigate characteristics according to intrinsic MIBTR subtype. In addition, about 30% of patients (nine of 28) did not have mammographic information at the time of MIBTR diagnosis, because every other US was done without mammography according to our institution protocol.
In conclusion, the overall MIBTR detection rate of US was not higher than the detection rate of mammography, although combined surveillance with US and mammography found MIBTRs slightly earlier than mammography alone.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Basic Science Research Program of the National Research Foundation of Korea funded by the Ministry of Science, ICT& Future Planning, Republic of Korea (grant no. 2013R1A1A3013165) and by a faculty research grant of Yonsei University College of Medicine for 2013 (6-2013-0094). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.