Health service areas (HSAs) within Surveillance, Epidemiology, and End Results areas shaded by use of RT among patients who receive BCS for DCIS. Twenty-five HSAs with fewer than 20 patients with DCIS diagnoses during the study period are not shaded.
Mastectomy performed at the time of a second breast diagnosis (invasive cancer or DCIS) among women who receive breast-conserving surgery (BCS) without radiotherapy (RT) at initial diagnosis of DCIS by RT use (as shown in Figure 1). Odds ratios greater than 1 represent a higher likelihood of receiving mastectomy and were adjusted for age, race, ethnicity, income, educational level, residence, characteristics of the secondary diagnosis (cancer stage, estrogen-receptor status, and laterality), year of secondary diagnosis, and interval between diagnosis (Surveillance, Epidemiology, and End Results [SEER]) as well as Charlson comorbidity score, distance to nearest RT facility, chemotherapy for secondary breast event, and magnetic resonance imaging in the 6 months before the secondary breast event (SEER-Medicare). P values represent values for the health services area treatment intensity for primary DCIS variable in the adjusted model. Squares indicate point estimates; limit lines, 95% CIs.
eFigure 1. Flow Diagram for SEER Cohort
eFigure 2. Flow Diagram for SEER-Medicare Cohort
eTable 1. Characteristics of Patients Receiving BCS Without Radiation Therapy for Primary DCIS and Who Had Second Breast Event (DCIS or Invasive Cancer)
eTable 2. Association Between Patient Characteristics and Three-Level Cluster of Treatment Intensity for Primary DCIS
eTable 3. SEER Analysis of Treatment for Secondary Breast Cancer After Receiving BCS Alone for Primary DCIS
eTable 4. SEER-Medicare Analysis of Treatment for Secondary Breast Cancer After Receiving BCS Alone for Primary DCIS
eTable 5. Propensity Score Matched Samples: Association Between Patient Characteristics and Three-Level Cluster of Treatment Intensity for Primary DCIS
eMethods. Further Specification
Customize your JAMA Network experience by selecting one or more topics from the list below.
Punglia RS, Cronin AM, Uno H, et al. Association of Regional Intensity of Ductal Carcinoma In Situ Treatment With Likelihood of Breast Preservation. JAMA Oncol. 2017;3(1):101–104. doi:10.1001/jamaoncol.2016.2164
Copyright 2016 American Medical Association. All Rights Reserved.
Do regional practice patterns of radiotherapy for ductal carcinoma in situ (DCIS) affect the likelihood of eventual mastectomy in women who receive breast-conserving surgery (BCS) without radiotherapy for DCIS?
In this study using 2 population-based databases, Surveillance, Epidemiology, and End Results (SEER) and SEER-Medicare, residence in a health service area characterized by greater radiotherapy use for DCIS increased the likelihood of women undergoing mastectomy vs BCS at subsequent breast event even among those who had not previously received radiotherapy for DCIS.
This association of regional radiotherapy with the use of mastectomy suggests that provider-related factors are affecting the likelihood of breast preservation.
Large regional variation exists in the use of radiotherapy after breast-conserving surgery (BCS) for ductal carcinoma in situ (DCIS). Although patients who do not receive initial radiotherapy for DCIS are candidates for subsequent BCS if they experience a second breast event, many undergo mastectomy instead.
To examine whether regional practice patterns of radiotherapy for DCIS affect the use of mastectomy in these patients.
Design, Setting, and Participants
A retrospective analysis of population-based databases (Surveillance, Epidemiology, and End Results [SEER] and SEER-Medicare). Data were obtained for 2679 women in SEER with a diagnosis of DCIS between 1990 and 2011 and for 757 women in SEER-Medicare with a DCIS diagnosis between 1991 and 2009 who had not undergone radiotherapy for DCIS and experienced a subsequent breast cancer or DCIS diagnosis.
Treatment intensity for primary DCIS (high, medium, low), as defined by separating health service areas (HSAs) into 3 clusters based on radiotherapy use.
Main Outcomes and Measures
Mastectomy vs BCS at a second breast event defined as DCIS recurrence or new invasive cancer.
The median (SD) ages of the participants was 64 (13) years for the 2679 SEER population and 79 (6) years for the SEER-Medicare cohort. Residence in an HSA characterized by greater radiotherapy use for DCIS increased the likelihood of receiving mastectomy vs BCS at a subsequent breast event, even among women who had not previously received radiotherapy for DCIS. Adjusted odds ratios for receiving mastectomy were 1.43 (95% CI, 1.10-1.85) and 1.90 (95% CI, 1.27-2.84) in SEER and SEER-Medicare databases, respectively, among women residing in an HSA with the greatest radiotherapy use vs the least, corresponding to an adjusted increase from 40.8% to 49.6%, and from 38.6% to 54.5%.
Conclusions and Relevance
Areas with more radiotherapy use for DCIS had increased use of mastectomy at the time of a second breast event even among patients eligible for breast conservation. This association suggests that physician-related factors are affecting the likelihood of breast preservation.
Patients and their physicians are often confronted with a decision between more intensive or less intensive treatment for a particular diagnosis. Quality decision making between these options requires careful balancing of the risks and adverse effects, as well as weighing the expected outcomes and their associated value as assessed by the patient.
Although the incidence of ductal carcinoma in situ (DCIS) has risen dramatically,1 there is considerable debate about optimal treatment. Intensive therapies for DCIS, such as mastectomy (removal of the breast) or radiotherapy following breast-conserving surgery (BCS), reduce the likelihood of a second breast diagnosis2-5 but have not been shown to improve survival.6 In addition, radiotherapy usually necessitates mastectomy should a new cancer or DCIS develop in the same breast at any point during the patient’s lifetime, in addition to a small chance of long-term toxic effects. Previous radiotherapy can also complicate reconstructive options following mastectomy. The trade-off between risk of a second breast diagnosis and adverse effects and potential consequences of radiotherapy therapy underscores the need for patient preference–driven decision making.
Patients who receive BCS alone may be candidates for subsequent BCS if they have a second breast event in the same breast. One study7 suggests that some women choose not to have radiotherapy after DCIS because they want to have a breast preservation option should a second breast diagnosis occur. However, the likelihood of mastectomy vs BCS at the time of a new diagnosis in a previously unirradiated breast is variable.8-10 Whether a woman undergoes a second BCS for a new diagnosis may not only be a function of the stage of diagnosis but may also be determined by the regional treatment patterns used for management of DCIS. We sought to study whether regional intensity of radiotherapy use for DCIS treatment increases the likelihood of mastectomy at the time of a second breast event among women who have not received radiotherapy at initial DCIS diagnosis.
We identified 33 194 patients with DCIS between 1990 and 2011 who were treated with BCS without radiotherapy within registries captured by the Surveillance, Epidemiology, and End Results (SEER) database. Using SEER-Medicare diagnoses from 1990 to 2009 linked to Medicare claims through 2010, we identified 5320 such patients (eFigure 1 and eFigure 2 in the Supplement). This study was reviewed and approved by the Dana-Farber/Harvard Cancer Center Institutional Review Board, which also determined it to be exempt from the need for informed consent. SEER data are deidentified; SEER-Medicare data are limited.
We studied the receipt of mastectomy (vs BCS with or without radiotherapy) for a second breast diagnosis (stage 0-III breast cancer) among patients receiving BCS alone for primary DCIS. Multivariable logistic regression modeling was used with all variables of interest (eTable 1 in the Supplement) regardless of statistical significance.
Health service areas (HSAs) were assigned to 1 of the 3 clusters based on the observed proportion of radiotherapy use as coded by SEER or determined by claims in SEER-Medicare. The HSAs with the highest proportions of patients receiving radiotherapy were assigned to the high cluster; those with the lowest proportions, to the low cluster; and those in between, to the middle cluster. The HSAs with fewer than 20 patients diagnosed during the study period were not assigned (eMethods in the Supplement).
We identified 2679 women (median [SD] age, 64  years) in SEER and 757 women (79  years) in SEER-Medicare with stage 0 to III breast cancer after DCIS who had received BCS without radiotherapy for initial treatment (eFigures 1 and 2 in the Supplement). These patients resided within 1 of 166 HSAs separated into 3 clusters based on the use of radiotherapy after BCS at the time of the initial DCIS diagnosis according SEER data11 (Figure 1) or 97 HSAs according to SEER-Medicare data (eTable 2 in the Supplement).
Patients who lived in HSAs with the highest proportion of radiotherapy use after BCS had a 43% increased odds of receiving mastectomy relative to those within HSAs with the lowest use, corresponding to an adjusted increase in mastectomy use from 40.8% to 49.6% in SEER (odds ratio [OR], 1.43; 95% CI, 1.10-1.85) (Figure 2). In SEER-Medicare, patients within the highest proportion had a 90% increased odds of undergoing mastectomy compared those within HSAs with the lowest use (OR, 1.90; 95% CI, 1.27-2.84), corresponding to an adjusted increase in mastectomy use from 38.6% to 54.5% (eTable 3 and eTable 4 in the Supplement).
Analyses conducted with propensity score matching (eTable 5 in the Supplement) revealed similar associations, with corresponding ORs of 1.87 (95% CI, 1.14-3.06) in SEER and 1.56 (95% CI, 0.91-2.65) in SEER-Medicare. Restricting the SEER analysis to patients with an ipsilateral second diagnosis also showed a similar pattern, with a 63% increased odds in HSAs with the highest proportion of radiation use (OR, 1.63; 95% CI, 1.17-2.30).
The decision between whether to pursue more or less aggressive treatment for a medical condition is ideally made by patients who weigh the advantages and disadvantages of each approach with respect to the outcomes that they value. However, regional treatment paradigms also influence these decisions, leading to unwarranted regional variation in use of therapies instead of use directed by patient preferences, which can be a marker of poor quality of care.12
We demonstrate that local treatment intensity for DCIS, defined by the proportion of women who undergo radiotherapy after BCS in an HSA, affects the likelihood of mastectomy at the time of a second breast event diagnosis among women who have not received radiotherapy for their initial DCIS. Patients who lived in an area with greater use of radiotherapy with BCS were more likely to undergo a mastectomy if they had a second breast event even though they were candidates for BCS, since they had initially foregone radiotherapy.
We used the SEER and SEER-Medicare databases to study the likelihood of mastectomy at the time of a second diagnosis after DCIS among women who received BCS alone. Limitations of the SEER database include the lack of sensitivity for capturing radiotherapy use13 and second breast events. However, our results were consistent among such patients in the SEER-Medicare database when claims linked to reimbursement were used to identify radiotherapy and second events. No data set can capture all of the complexity surrounding surgical decision making at the time of a second diagnosis. We did not have information about patient preferences, the availability of breast tissue for good cosmetic results after a second BCS, and clinical characteristics (beyond stage) about the second breast event that could affect the choice of the surgical procedure. Nevertheless, we controlled for stage in our analyses, and our results surrounding the effect of local treatment intensity on mastectomy use at the second diagnosis were stable when limiting the analyses to ipsilateral events in SEER. In addition, there is no reason to believe that these characteristics would vary systematically by region, which is the only way that they would bias our findings.
Most people would agree that the quality of life for patients with breast cancer has improved with the feasibility of BCS. An unintended consequence of using radiotherapy for DCIS is the increased use of mastectomy downstream among women who may have opted to receive less aggressive treatment for their initial DCIS (ie, BCS alone). Physicians in regions of high use of radiotherapy may guide patients with DCIS toward mastectomy because many of these patients are ineligible for BCS at the time of a second breast event—having already received radiotherapy—leading to mastectomy being recommended for patients who did not receive radiotherapy and are eligible for BCS. Awareness of this effect of practice patterns may be the first step toward its eradication and movement toward more patient-centered care.
Corresponding Author: Rinaa S. Punglia, MD, MPH, Department of Radiation Therapy, Dana-Farber Cancer Institute, Harvard Medical School, 450 Brookline Ave, Dana 1115C, Boston, MA 02115 (firstname.lastname@example.org).
Accepted for Publication: May 2, 2016.
Published Online: July 21, 2016. doi:10.1001/jamaoncol.2016.2164
Author Contributions: Dr Punglia and Ms Cronin had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Punglia, Greenberg, Frank.
Acquisition, analysis, or interpretation of data: Punglia, Cronin, Uno, Stout, Ozanne, Greenberg, Schrag.
Drafting of the manuscript: Punglia, Cronin, Frank.
Critical revision of the manuscript for important intellectual content: Uno, Stout, Ozanne, Greenberg, Frank, Schrag.
Statistical analysis: Punglia, Cronin, Uno, Schrag.
Obtained funding: Punglia.
Administrative, technical, or material support: Schrag.
Conflict of Interest Disclosures: None reported.
Funding/Support: The Patient-Centered Outcomes Research Institute provided funding for this study.
Role of the Funder/Sponsor: The Patient-Centered Outcomes Research Institute 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.
Disclaimer: This study used Surveillance, Epidemiology, and End Results (SEER) Program (http://www.seer.cancer.gov) Research Data (1973-2012), National Cancer Institute, Division of Cancer Control and Population Sciences, Surveillance Research Program, Surveillance Systems Branch, and the linked SEER-Medicare databases. The interpretation and reporting of these data are the sole responsibility of the authors.
Additional Contributions: Patient advocates Christine Tannous, PhD, and Sandra Finestone, PhD (compensation provided), and colleague Gregory Abel, MD, MPH (Dana-Farber Cancer Institute) (no compensation provided), offered important insights regarding our analysis and findings. Drs Tannous and Finestone received financial compensation; Dr Abel did not. We acknowledge the efforts of the National Cancer Institute; the Office of Research, Development, and Information, Centers for Medicare & and Medicaid Services; Information Management Services, Inc; and the SEER Program tumor registries in the creation of the SEER-Medicare database.
Create a personal account or sign in to: