Incidence and Outcome of Breast Biopsy Procedures During Follow-up After Treatment for Breast Cancer

Key Points

Question  How often will patients treated for breast cancer require invasive breast biopsies during follow-up?

Finding  In this cohort study of 121 879 patients with Medicare or commercial insurance, an estimated 14.9% to 23.4% of patients underwent breast biopsy by 10-year follow-up; breast biopsy was significantly associated with age at initial treatment, type of initial surgery or radiotherapy, and administration of chemotherapy. After biopsy, approximately 20% to 30% of women underwent additional cancer treatment.

Meaning  These biopsy rates provide meaningful information regarding treatment-related outcomes and expectations for patients treated for breast cancer.

Importance  No comprehensive data are available regarding the frequency of breast biopsies performed during follow-up of treatment for invasive breast cancer.

Objective  To determine how often patients treated for breast cancer require breast biopsies during follow-up.

Design, Setting, and Participants  This nationwide population-based cohort study included 41 510 patients 64 years or younger in a commercial insurance database and 80 369 patients 66 years or older in the Surveillance, Epidemiology, and End Results (SEER)–Medicare database. Patients were diagnosed with incident invasive breast cancer (stages I-III) from January 1, 2000, through December 31, 2011. Diagnosis and procedural codes were used to identify biopsy rates during follow-up. Data were analyzed from March 3 through October 3, 2017.

Main Outcomes and Measures  Cumulative incidence and adjusted risk of breast biopsy and subsequent breast cancer treatment were calculated using the Kaplan-Meier method and Cox proportional hazards regression. All statistical tests were 2 sided.

Results  Among the 121 879 patients in the study population, 5- and 10-year overall incidences of breast biopsy were 14.7% and 23.4%, respectively, in the commercial insurance cohort and 11.8% and 14.9%, respectively, in the SEER-Medicare cohort. The 5-year estimated incidence of breast biopsy was higher among women treated with brachytherapy (24.0% in the commercial insurance and 25.0% in the SEER-Medicare cohorts) than among those treated with whole-breast irradiation (16.7% in the commercial insurance and 15.1% in the SEER-Medicare cohorts) and persisted after multivariate adjustment in the commercial insurance (hazard ratio [HR], 1.53; 95% CI, 1.38-1.70; P < .001) and SEER-Medicare (HR, 1.76; 95% CI, 1.63-1.91; P < .001) cohorts. Adjuvant chemotherapy use (HR, 1.31; 95% CI, 1.25-1.37; P < .001) and patient age (>85 vs 66-69 years; HR, 0.40; 95% CI, 0.36-0.44; P < .001) in the SEER-Medicare cohort and endocrine therapy in the commercial insurance (HR, 0.88; 95% CI, 0.82-0.93; P < .001) and SEER-Medicare (HR, 0.91; 95% CI, 0.85-0.97; P = .002) cohorts were independently associated with biopsy. After unilateral mastectomy, the estimated 5-year contralateral breast biopsy rates were 10.4% and 7.7% in the commercial insurance and SEER-Medicare cohorts, respectively. Of the patients with breast biopsy, 1239 of 4158 patients (29.8%) in the commercial insurance cohort and 2258 of 9747 patients (23.2%) in the SEER-Medicare cohort underwent subsequent cancer treatment.

Conclusions and Relevance  These data on the need for breast biopsies during follow-up and subsequent treatments from a large cohort of women with commercial insurance and Medicare can be used in the context of therapy-planning discussions and survivorship expectations for patients with breast cancer.

When discussing initial surgical options for breast cancer, patients cite fear of having to endure the process of repeated breast biopsies and the associated stress related to the potential of another cancer diagnosis during follow-up for breast cancer as the reason for choosing mastectomy or bilateral mastectomy.¹ The continued increased risk for second breast cancers (local recurrence or second primary) after breast-conserving surgery (BCS) and radiotherapy is well known, with a rate of ipsilateral breast recurrence of 0.2% to 1.0% per year^2,3 (dependent on biological subtype^4-7) and a contralateral breast cancer incidence of approximately 0.6% per year.⁸ The published literature remains unclear on how often patients treated for invasive breast cancer will require breast biopsies during follow-up. To address this knowledge gap and to better provide patients and multidisciplinary treatment teams with this type of information, the goal of this study was to determine how often biopsies are performed in follow-up after previous breast cancer treatment in 2 large complementary databases consisting of younger and older women with insurance. The rate of contralateral breast biopsy was estimated among patients initially treated with unilateral mastectomy. A secondary study goal was to determine how breast biopsies are related to patient demographics and initial breast cancer treatments.

We used the MarketScan Commercial Claims and Encounters database (Truven Health Analytics) to identify women younger than 65 years who had incident invasive breast cancer (International Classification of Diseases, Ninth Revision [ICD-9] diagnosis code 174.x) diagnosed from 2000 to 2011 using a modified claims-based algorithm (eTable 1 in the Supplement).⁹ Health plans in this commercial insurance database include employer-sponsored, private, fee-for-service, and capitated insurance to employees and covered dependents. This longitudinal database tracks all patient-level inpatient and outpatient claims as long as the employees remain with their employers. The Surveillance, Epidemiology, and End Results (SEER)–Medicare database was used to identify incident cases of breast cancer in patients 66 years or older. The SEER database captures clinical, pathologic, and Medicare claims for incident cancers diagnosed in individuals who reside within 1 of the 16 geographic areas accounting for 26% of the US population. The case ascertainment is approximately 98%, and 93% of patients in SEER are successfully linked to Medicare claims.¹⁰ This study was granted exempt status by the institutional review board of The University of Texas MD Anderson Cancer Center.

The cohorts were limited to patients with complete insurance coverage from 12 months before through 12 months after diagnosis to enable determination of prediagnosis comorbidity and to ensure that patients completed their initial breast treatment (breast surgery with radiotherapy and/or chemotherapy). Because staging data are not available in the commercial insurance database claims, tumor stage was assigned as follows: patients with invasive breast cancer (ICD-9 code 174.x) without an accompanying code for stage IV disease (ICD-9 codes 196.X-199.X, excluding 196.0, 196.3, 198.2, and 198.81) were assumed to have had stages I through III cancer. Patients with claims for other simultaneous cancers diagnosed within 12 months after the breast cancer diagnosis were excluded.

The SEER-Medicare cohort was limited to patients with T1 through T4 and N0 to N3 breast cancer using SEER breast-adjusted American Joint Committee on Cancer, 7th edition, stages I, II, and III.¹¹ The analytic cohorts were further limited by excluding patients with a history of breast cancer and bilateral mastectomy. By using these criteria, we identified 41 510 patients from the commercial insurance database and 80 369 patients from the SEER-Medicare database. Current Procedural Terminology treatment and national drug codes were used as previously reported by Smith et al.¹¹

Procedure codes were used to identify breast biopsies performed during follow-up. These biopsies were classified as surgical, percutaneous with image guidance (eg, stereotactic, ultrasonography-guided, magnetic resonance imaging–guided, and not further specified), and percutaneous without image guidance (eTable 2 in the Supplement). The biopsy procedure was included if listed with a breast-related diagnosis for the claim or if listed as a breast-specific biopsy code. Subsequent treatment after biopsy was defined as a claim for breast surgery (lumpectomy, mastectomy), new systemic therapy, any radiotherapy within 3 months after the breast biopsy, and/or any combination of subsequent treatments (eTable 3 in the Supplement). The initial unilateral mastectomy treatment group was used to approximate the contralateral biopsy rate.

Demographic covariates included age, ethnicity/race (only available in the SEER-Medicare cohort), health plan, and year of diagnosis. Severity of comorbid disease was based on a modified Charlson comorbidity index¹¹ derived from claims during the 12 months preceding diagnosis and grouped as 0 for none, 1 for mild, 2 for moderate, or 3 or higher for severe comorbidity. Additional treatment-related variables derived from claims included receipt of chemotherapy, trastuzumab, and adjuvant endocrine therapy.¹¹ The last available claim date was used for the commercial insurance cohort to determine the end of follow-up. Final follow-up was completed on December 31, 2011.

Data were analyzed from March 3 to October 3, 2017. Analyses were conducted using SAS software (version 9.3; SAS Institute) and assumed a 2-tailed α = .05. The primary outcome was the percentage of women who underwent breast biopsies during follow-up. For the secondary outcome, we examined associations between breast biopsies and patient age, health plan type, comorbidity, treatment year, and initial breast cancer treatment, including surgery, chemotherapy, trastuzumab, and endocrine therapy using Pearson χ² tests. Subsequent breast treatment after breast biopsy constituted another secondary outcome.

The rate of breast biopsy was estimated using the Kaplan-Meier method (log-rank test) for the commercial insurance cohort, because survival status was not reported in this cohort, whereas cumulative incidence (Gray test) was used for the SEER-Medicare cohort, considering the relatively high competing risk of death in this cohort of individuals 66 years or older. Patients in the SEER-Medicare cohort were censored at the last available medical care service date or death. We used Cox proportional hazards regression and Fine-Gray competing risk regression models^12,13 to test the association between breast biopsy and subsequent breast treatment and adjusted for covariates identified a priori based on clinical significance and univariate analyses (P < .05). The proportional hazards assumption was confirmed by inspection of log (-log [survival]) curves. The SEER-Medicare and commercial insurance data sets were assembled with different selection methods. Therefore, no statistical comparisons or inferences were conducted between these 2 groups.

Baseline characteristics from 121 879 patients (41 510 in the commercial insurance and 80 369 in the SEER-Medicare databases) diagnosed with and treated for breast cancer from January 1, 2000, through December 31, 2011, are enumerated in Table 1. The median follow-up was 3.7 years (interquartile range, 2.1-5.6 years) for the commercial insurance cohort and 5.8 years (interquartile range, 3.5-8.7 years) for the SEER-Medicare cohort. The mean (SD) patient age was 52.8 (7.4) years for the commercial insurance cohort and 76.3 (6.8) years for the SEER-Medicare cohort. In the commercial insurance cohort, 24 341 patients (58.6%) initially received BCS and 17 169 (41.4%) received mastectomy. Of the 80 369 SEER-Medicare cohort patients, BCS was initially performed in 47 578 (59.2%) and mastectomy was performed in 32 791 (40.8%).

Overall, a breast biopsy was performed during breast cancer follow-up in 4158 patients in the commercial insurance cohort (10.0%) and 9747 in the SEER-Medicare cohort (12.1%). In the commercial insurance cohort, a total of 3635 patients (8.8%) had at least 1 biopsy during follow-up and 523 (1.3%) had 2 or more biopsies. In the SEER-Medicare cohort, 8164 patients (10.2%) had 1 biopsy and 1583 (2.0%) had at least 2 biopsies. The 5- and 10-year estimated biopsy rates are 14.7% and 23.4%, respectively, for the commercial insurance cohort and 11.8% and 14.9%, respectively, for the SEER-Medicare cohort (Table 2) and were associated with patient age and initial treatment characteristics (Figure and eFigure in the Supplement).

The 5-year estimated breast biopsy rate in the commercial insurance cohort was 10.4% among those initially treated with unilateral mastectomy compared with 16.7% after initial BCS with whole-breast irradiation (WBI), 24.0% after BCS with brachytherapy, and 22.0% after BCS alone. In the SEER-Medicare cohort, the estimated 5-year biopsy rates were 7.7% after initial mastectomy, 15.1% after BCS with WBI, 25.0% after BCS with brachytherapy, and 9.4% after BCS alone (Figure). Breast brachytherapy was associated with a higher risk of breast biopsy, with an estimated 5-year cumulative incidence of 24.0% in the commercial insurance cohort and 25.0% in the SEER-Medicare cohort vs 16.7% and 15.1% among patients treated with WBI in the respective cohorts (P < .001) (Figure).

On multivariate analysis, brachytherapy vs WBI remained associated with an increased risk of subsequent biopsy in the commercial insurance (hazard ratio [HR], 1.53; 95% CI, 1.38-1.70; P < .001) and SEER-Medicare (HR, 1.76; 95% CI, 1.63-1.91; P < .001) cohorts (Table 3). Compared with WBI, having a unilateral mastectomy was associated with a reduced risk of subsequent biopsy for the commercial insurance (HR, 0.60; 95% CI, 0.56-0.64; P < .001) and SEER-Medicare (HR, 0.53; 95% CI, 0.50-0.55; P < .001) cohorts. Adjuvant chemotherapy remained associated with an increased risk for subsequent biopsy in the SEER-Medicare cohort (HR, 1.31; 95% CI, 1.25-1.37; P < .001), whereas endocrine therapy was associated with reduced risk of subsequent biopsy in the commercial insurance (HR, 0.88; 95% CI, 0.82-0.93; P < .001) and SEER-Medicare (HR, 0.91; 95% CI, 0.85-0.97; P = .002) cohorts. Being older (vs age 66-69 years) remained associated with reduced biopsy risk in the SEER-Medicare cohort (HR for ≥85 years, 0.40; 95% CI, 0.36-0.44; P < .001). Charlson comorbidity index of at least 2 (vs 0) was associated with an increased biopsy risk in the commercial insurance cohort (HR, 1.27; 95% CI, 1.01-1.59; P = .04), whereas it was associated with a decreased biopsy risk in the SEER-Medicare cohort (HR, 0.91; 95% CI, 0.85- 0.97; P = .01).

After biopsy, 1239 of 4158 patients (29.8%) in the commercial insurance cohort and 2258 of 9747 patients (23.2%) in the SEER-Medicare cohort underwent additional breast cancer treatment within 3 months of biopsy (Table 4). This treatment consisted of surgery alone in 723 patients (58.4%) in the commercial insurance cohort and in 1302 (57.7%) in the SEER-Medicare cohort. In the commercial insurance cohort, surgery combined with chemotherapy was used in 136 patients (11.0%) and surgery combined with radiotherapy in 53 (4.3%). In the SEER-Medicare cohort, 93 patients (4.1%) had surgery combined with chemotherapy and 118 (5.2%) had surgery combined with radiotherapy. Surgery combined with endocrine therapy was administered in 65 patients (5.3%) in the commercial insurance cohort and in 44 (1.9%) in the SEER-Medicare cohort. Initial chemotherapy alone was administered in 108 patients (8.7%) in the commercial insurance cohort and 344 (15.2%) in the SEER-Medicare cohort. Patients receiving a breast biopsy during follow-up were significantly more likely to receive additional breast cancer treatments if they were initially treated with BCS alone compared with other initial treatments in the commercial insurance (179 of 361 [49.6%]) and SEER-Medicare (370 of 847 [43.7%]; P < .001) cohorts (Table 4).

In this comprehensive sample of younger and older women who completed their initial breast cancer treatment, the 5- and 10-year estimated overall incidence of breast biopsy were 14.7% and 23.4%, respectively, in the commercial insurance cohort and 11.8% and 14.9%, respectively, in the SEER-Medicare cohort. The risk of breast biopsy was significantly increased in both cohorts after initial brachytherapy and in the SEER-Medicare cohort after adjuvant chemotherapy. Initial endocrine therapy and older patient age were associated with reduced biopsy risk. Higher comorbidity was associated with increased risk of biopsy in the commercial insurance cohort, whereas it was associated with decreased biopsy rate in the SEER-Medicare cohort. The overall biopsy rates provide an estimate of the patient-level rate of any biopsy (ipsilateral and contralateral biopsy); it cannot distinguish between ipsilateral and contralateral procedures. However, among patients receiving unilateral mastectomy, the estimated 10-year contralateral biopsy rates in these cohorts varied from 10.1% to 16.7%, indicating that a substantial number of invasive breast procedures will occur during follow-up of the contralateral breast. The present study represents, to our knowledge, the first comprehensive, population-based study on biopsy rates in follow-up after breast cancer treatment.

Despite the considerable clinical and sociodemographic differences between the 2 nonoverlapping cohorts, we observed concordance in the results of our analysis. In both populations, brachytherapy was associated with increased risk of subsequent biopsy during follow-up of 53% in the commercial insurance cohort and 76% in the SEER-Medicare cohort, compared with the standard of practice of BCS with WBI. These synchronous findings among the 2 cohorts provide cross-validation that enhances the scientific validity of these findings. The higher biopsy rate after brachytherapy is likely explained in part by the difficulties encountered in the interpretation of surveillance mammograms after brachytherapy owing to seroma, fat necrosis, and scarring.^14-16 In addition, as routinely practiced in the United States, breast brachytherapy may also be associated with higher rates of subsequent cancer treatment, as suggested by data from the present study. Previous data from the SEER-Medicare data set showed an increased subsequent mastectomy rate after brachytherapy compared with standard WBI, although the claims data are limited for confirming actual cancer recurrence.^15,17 However, the results of several trials using interstitial brachytherapy show no difference in recurrence between strictly selected patients who underwent brachytherapy and those who underwent WBI.^18-23 After BCS alone, local recurrence rates are known to be increased.^24-27 We also found that after BCS alone (compared with whole-breast radiotherapy), subsequent treatment after biopsy in the present study is concordant with this finding.

After unilateral mastectomy, the estimated biopsy rates during the first 5 and 10 years were 10.4% and 16.7%, respectively, in the younger cohort and 7.7% and 10.1%, respectively, in the older cohort. For patients who received a mastectomy, we assumed that the breast biopsy was performed on the remaining contralateral breast without cancer during follow-up. The mastectomy group thus served as a convenience sample to estimate the true contralateral biopsy rate. The estimated 10-year ipsilateral biopsy rate after BCS with WBI was 10.8% (27.5% minus 16.7%) for the younger cohort and 9.3% (19.4% minus 10.1%) for the older cohort.

The increased HRs of 1.31 after adjuvant chemotherapy and 1.24 after neoadjuvant chemotherapy in the SEER-Medicare cohort represent the more advanced disease stages, which are associated with a higher risk of local recurrence. Endocrine therapy has a protective effect on the local recurrence rate and contralateral breast cancer and was thus associated with a lower biopsy rate in both cohorts (HRs, 0.88 in the commercial insurance cohort and 0.91 in the SEER-Medicare cohort). Being younger was associated with an increased biopsy rate in the present study. Vapiwala et al²⁸ examined factors predictive of ipsilateral breast biopsy after BCS with radiotherapy in 193 treated breasts and found that the biopsy rate varied inversely with the age at the original diagnosis. This finding is consistent with the well-established correlation of younger age with a greater risk of local recurrence.

Houssami et al²⁷ identified age at the primary breast cancer diagnosis younger than 40 years as the strongest predictive factor of an invasive breast cancer diagnosed during follow-up (interval cancer) with BCS without radiotherapy, increased mammographic breast density, and first-degree family history of breast cancer. Systemic therapy of any type did not affect the risk of an interval breast cancer, whereas endocrine therapy reduced the risk for screen-detected invasive breast cancer and for ductal carcinoma in situ. Buist et al²⁹ followed up 17 286 women for up to 5 years and found that women with adjuvant hormonal therapy and those who were 80 years or older had the lowest rates of second events. Women with a family history or greater breast density had higher rates of second primary cancers, and women who received BCS without radiotherapy had higher recurrence rates. Lee et al³⁰ prospectively evaluated a cohort of women with stages 0 to II unilateral primary breast cancer and found a 5-year risk of interval invasive second cancer of 0.60%. The 5-year risk ranged from 0.07% to 6.11% for women with the most favorable (screen-detected primary breast cancer, grade I, treated with mastectomy, and fatty breast tissue on surveillance mammogram) and least favorable (clinically detected primary breast cancer, grade II, treated with lumpectomy without radiotherapy, and extremely dense breast tissue on mammography) profiles.³⁰

Previous breast surgery and radiotherapy for breast cancer might complicate interpretation of surveillance imaging. Postsurgical fluid collections are often visualized and reflect the surgical technique. In the study by Mendelson,³¹ 50% were still present at 6 months and 20% were demonstrated at 9 months. Scar formation and cancer may present as spiculated, poorly marginated soft-tissue densities that require a biopsy to differentiate. Increased breast density can also be caused by edema and skin thickening seen after radiotherapy. A recurrence might also manifest by calcifications. Mendelson³¹ found new calcifications within 6 to 12 months in 28% of 110 patients. An analysis of 2773 patients with suspicious calcifications³² identified in the ipsilateral breast within 24 months after BCS demonstrated a 7% cancer yield. In a retrospective review of 342 patients who had early post-BCS mammograms performed, Barron et al³³ showed that 19 (5.5%) had findings that resulted in a biopsy with a yield of only 1 malignant neoplasm (0.3%), which was not enough to justify early mammograms. They recommended the first screening mammogram to be closer to 1 year, rather than 6 to 12 months.

In interpreting results of this study, information regarding the side of the initial cancer treatment and subsequent biopsy is not available from these databases, and information regarding whether the biopsy was ipsilateral or contralateral was not definitively known except in cases in which the initial treatment was unilateral mastectomy. Notwithstanding, the information obtained from this study regarding overall rates (ipsilateral and or contralateral) of breast procedures that occur after treatment for breast cancer in younger and older patients is of value for discussions with patients and provides novel data regarding this subject for the literature. Second, the study is limited by the nature of claims data, which rely on the accuracy of codes and are not comprehensive of all clinician and patient characteristics. Third, the precision estimates degrade quickly in the commercial insurance cohort with increasing follow-up time owing to loss of patients. Subsequent treatments were assumed to occur within 3 months after breast biopsy, and these treatments were associated with cancer. Although the first 3 months after biopsy were used to report subsequent treatment, some patients might have had chemotherapy or endocrine therapy after this time frame. The definition of invasive cancer and radiotherapy were claims based and could be subject to misclassification bias. Definitions were, however, based on a previously validated algorithm and sought to maximize specificity.⁹ Finally, for many women, surveillance for recurrence lasts longer than 10 years, and this study does not address lifetime rates of the need for breast biopsy after treatment.

In a comprehensive cohort of insured younger and older women, the estimated biopsy rate after breast cancer treatment varied from 14.9% to 23.4% and was followed by subsequent treatment for cancer in approximately 20% to 30% of patients. These data can be used in the context of therapy-planning discussions and to help establish survivorship expectations for patients with breast cancer.

Accepted for Publication: October 19, 2017.

Corresponding Author: Henry M. Kuerer, MD, PhD, Department of Breast Surgical Oncology, Division of Surgery, The University of Texas MD Anderson Cancer Center, Unit 1434, 1400 Pressler St, Houston, TX 77030 (hkuerer@mdanderson.org).

Published Online: January 31, 2018. doi:10.1001/jamasurg.2017.5572

Author Contributions: Dr Smith had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: van la Parra, Smith, Yang, Leung, Kuerer.

Acquisition, analysis, or interpretation of data: van la Parra, Liao, Smith, Leung, Giordano, Kuerer.

Drafting of the manuscript: van la Parra, Liao, Yang, Kuerer.

Critical revision of the manuscript for important intellectual content: van la Parra, Smith, Yang, Leung, Giordano, Kuerer.

Statistical analysis: Liao, Smith.

Obtained funding: Smith, Kuerer.

Administrative, technical, or material support: van la Parra, Yang, Giordano, Kuerer.

Study supervision: van la Parra, Smith, Yang, Kuerer.

Conflict of Interest Disclosures: Dr Smith reports receiving research funding from Varian Medical Systems. Dr Leung reports serving on the Scientific Advisory Board for Hologic, Inc. Dr Kuerer reports receiving publishing patents, royalties, and other intellectual property from the New England Journal of Medicine Group and McGraw-Hill Publishing, serving on the speaker’s bureau for Physicians’ Education Resource, LLC, and formerly receiving research funding from Genomic Health. No other disclosures were reported.

Funding/Support: This study was supported by a Clinical KWF Fellowship from the Dutch Cancer Society (Dr van la Parra), the PH and Fay Etta Robinson Distinguished Professorship in Cancer Research Endowment (Dr Kuerer), Cancer Center Support Grant CA16672 from the National Institutes of Health, the MD Anderson Clinical Research Funding Award Program (Dr Kuerer), the Andrew Sabin Family Fellowship (Dr Smith), grant RP 160674 from the Cancer Prevention and Research Institute of Texas (Dr Smith), and grant R01 CA207216-01 from the National Cancer Institute (Dr Smith).

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.