Michigan Breast Oncology Quality Initiative Research Articles

Factors Determining Anthracycline Use in Hormone Receptor Positive, Early-Stage Breast Cancer

BackgroundAnthracyclines are associated with significant toxicities whereas nonanthracyclines have proven to be better tolerated. A 21-gene assay allows clinicians to predict who will not benefit from adjuvant chemotherapy and avoid systemic toxicities. Physicians are using the recurrence score to guide chemotherapy selection, despite the lack of evidence. In this study we examined factors associated with prescribing patterns for an anthracycline-based chemotherapy in hormone receptor-positive stage I to III breast cancer. Materials and MethodsThis was a retrospective study using the Michigan Breast Oncology Quality Initiative data set (February 1, 2006 to December 31, 2015). Women with histologically confirmed stage I to III invasive breast cancer with estrogen receptor and/or progesterone receptor-positive, HER2/neu-negative receptor status were included. We used χ2 analysis to determine associations of these characteristics with the 21-gene assay score and anthracycline use. ResultsA total of 17,788 patients were evaluated. Most tumors were stage I (60%). Most procedures were lumpectomy with radiation (66%). Anthracyclines were used more often in stage III patients (69%), younger patients (40% for patients younger than 65 years), and those with higher 21-gene recurrence scores. Patients with low recurrence scores were more likely to receive anthracyclines if lymph node-positive (10%) than if lymph node-negative (1%; P < .001). Patients with high recurrence scores and lymph node-positive status were just as likely to receive an anthracycline-based as a nonanthracycline-based regimen (47.5% vs. 49.2%; P = .89). ConclusionThese data indicate that medical oncologists might be anticipating the results of Adjuvant Chemotherapy Guided by a 21-Gene Expression Assay in Breast Cancer study (TAILORx) and the Clinical Outcomes in ER+HER2-node-positive Breast Cancer Patients Who Were Treated According to the Recurrence Score Results: Evidence From a Large Prospectively Designed Registry (RxPonder) trials and are avoiding the potential serious complications associated with anthracycline treatment in patients least likely to receive benefit.

The Effect of the 2009 USPSTF breast cancer screening recommendations on breast cancer in Michigan: A longitudinal study

In 2009, the revised United States Preventive Services Task Force (USPSTF) guidelines recommended against routine screening mammography for women age 40-49years and against teaching self-breast examinations (SBE). The aim of this study was to analyze whether breast cancer method of presentation changed following the 2009 USPSTF screening recommendations in a large Michigan cohort. Data were collected on women with newly diagnosed stage 0-III breast cancer participating in the Michigan Breast Oncology Quality Initiative (MiBOQI) registry at 25 statewide institutions from 2006 to 2015. Data included method of detection, cancer stage, treatment type, and patient demographics. In all, 30008 women with breast cancer detected via mammogram or palpation with an average age of 60.1years were included. 38% of invasive cancers were identified by palpation. Presentation with palpable findings decreased slightly over time, from 34.6% in 2006 to 28.9% in 2015 (P<.001). Over the 9-year period, there was no statistically significant change in rate of palpation-detected tumors for women age <50years or ≥50years (P=.27, .30, respectively). Younger women were more likely to present with palpable tumors compared to older women in a statewide registry. This rate did not increase following publication of the 2009 USPSTF breast cancer screening recommendations.

Variation in the use of advanced imaging at the time of breast cancer diagnosis in a statewide registry.

Although national guidelines do not recommend extent of disease imaging for patients with newly diagnosed early stage breast cancer given that the harm outweighs the benefits, high rates of testing have been documented. The 2012 Choosing Wisely guidelines specifically addressed this issue. We examined the change over time in imaging use across a statewide collaborative, as well as the reasons for performing imaging and the impact on cost of care. Clinicopathologic data and use of advanced imaging tests (positron emission tomography, computed tomography, and bone scan) were abstracted from the medical records of patients treated at 25 participating sites in the Michigan Breast Oncology Quality Initiative (MiBOQI). For patients diagnosed in 2014 and 2015, reasons for testing were abstracted from the medical record. Of the 34,078 patients diagnosed with stage 0-II breast cancer between 2008 and 2015 in MiBOQI, 6853 (20.1%) underwent testing with at least 1 imaging modality in the 90 days after diagnosis. There was considerable variability in rates of testing across the 25 sites for all stages of disease. Between 2008 and 2015, testing decreased over time for patients with stage 0-IIA disease (all P < .001) and remained stable for stage IIB disease (P = .10). This decrease in testing over time resulted in a cost savings, especially for patients with stage I disease. Use of advanced imaging at the time of diagnosis decreased over time in a large statewide collaborative. Additional interventions are warranted to further reduce rates of unnecessary imaging to improve quality of care for patients with breast cancer. Cancer 2017;123:2975-83. © 2017 American Cancer Society.

Site variation in assessment of biomarkers on biopsy specimens in patients with metastatic disease.

224 Background: Clinical practice guidelines in patients with metastatic breast cancer recommend assessment of biomarkers—estrogen receptors (ER), progesterone receptors (PR), and HER2—on metastatic lesions. The purpose of this study was to identify factors associated with assessment of these biomarkers in patients who had undergone confirmatory biopsies. We were particularly interested in variation between treatment sites. Methods: Eligible patients were those diagnosed with metastatic breast cancer who had undergone a confirmatory biopsy and who were part of the Michigan Breast Oncology Quality Initiative (MiBOQI) registry, a statewide registry of 25 health systems. Analyses investigated associations between assessment of ER, PR, HER2 and treating center, disease, clinical and non-clinical factors. Results: Between 2006 and 2015, 805 patients had a confirmatory biopsy. Of these, 66% had ER, 63% had PR, 62% had HER2, and 57% had all three assessed on the biopsy specimen. In bivariate analysis, characteristics associated with assessment of biomarkers were earlier stage and hormone receptor-positive primary tumors, private or military insurance, white race, longer time since initial diagnosis, and bone only, liver only, lung only, chest wall, skin, and lymph nodes, and multiple sites of metastases. Younger age was positively associated with assessment of HER2. Biomarker assessment also varied significantly between treatment sites and increased over time. In multivariate analyses, time since diagnosis of the primary (p &lt; 0.001) and treatment site (p &lt; 0.001) remained significant. With the exception of ER (p = 0.09), performance of biomarkers increased significantly over time. Conclusions: Across 25 health systems in a quality improvement collaborative, the proportion of patients with metastatic disease who underwent a confirmatory biopsy increased over time. Nonetheless, treatment site continued to be independently associated with the odds of having biomarkers assessed on biopsy specimens. Our findings suggest that there are opportunities for improvement to provide guideline-concordant care in patients treated in diverse settings.

Clinical and nonclinical factors associated with confirmatory biopsy in metastatic breast cancer.

249 Background: In patients with metastatic breast cancer, clinical practice guidelines include obtaining histologic confirmation of metastases when possible. The purpose of this study was to investigate patterns and correlates of receipt of a confirmatory biopsy. Methods: Data were abstracted from the records of patients diagnosed with metastatic breast cancer in the Michigan Breast Oncology Quality Initiative (MiBOQI), a statewide registry of 25 health systems. Patients with Stage IV disease at diagnosis were excluded. Analyses investigated associations between receipt of a confirmatory biopsy and disease, clinical and non-clinical factors, and treating health system. Results: Data were available for 1,231 (96%) of eligible patients between 2006 and 2015. Of these, 66% had a confirmatory biopsy; the proportion of patients having a biopsy varied between the 25 sites from 41% to 100% (p = 0.03). In bivariate analyses, younger age (p = 0.02), lower comorbidity (p = 0.007), longer time between the primary and recurrence (p &lt; 0.001), more recent year of recurrence (p = 0.01), having liver, skin, soft tissue, or multiple metastases (p &lt; 0.001), and private or government insurance (p = 0.002) were associated with biopsy. In multivariate analyses, longer time since the primary diagnosis (p &lt; 0.001), more recent year of recurrence (p = 0.03), initial site(s) of recurrence (p &lt; 0.001), and private or government insurance (p = 0.004) remained significant predictors of biopsy. Treatment site was no longer significant (p = 0.14). Minority status, obesity status, and disease characteristics (stage, estrogen receptor, progesterone receptor, HER2, grade) of the primary were not significant in either bivariate or multivariate analyses. Analyses were repeated without insurance with no change in the other findings. Conclusions: In a statewide collaborative, the proportion of patients having a confirmatory biopsy increased over the study period and was associated with several clinical factors. Insurance was an independent and significant predictor of receipt of what is considered standard care.

Abstract P1-10-08: Factors affecting the administration of post-mastectomy radiation therapy (PMRT) in Michigan

Abstract Background: Evidence-based guidelines for locoregional therapy of invasive breast cancer treated with mastectomy include adjuvant PMRT for: ≥4 positive axillary lymph nodes (LN); T3 or above; or a positive surgical margin. We assessed PMRT uptake using data from the Michigan Breast Oncology Quality Initiative (MiBOQI), a Blue Cross Blue Shield of Michigan/Blue Care Network-sponsored collaborative quality initiative, and identified factors influencing its use in Michigan. Methods: We prospectively collected clinical data on all patients with stage I-III breast cancer in 25 health systems belonging to MiBOQI and identified patients who underwent mastectomy from 2008 to 2013. Patients with previous cancer, bilateral disease, or treated with neoadjuvant chemotherapy were excluded. Univariate and multivariate analyses were performed to identify independent factors associated with the use of PMRT in patients with 0, 1-3, and 4+ positive LNs. Covariates included age, hormone receptor status, HER2 status, surgical margin, T category, Charlson comorbidity index, and immediate reconstruction. Two-tailed p-values &amp;lt;0.05 were considered significant. Analyses were carried out using SAS software, version 9.4 (SAS Institute, Cary, NC). Results: We identified 6,596 patients with stage I-III invasive breast cancer. Of these, 4,455 had no positive axillary LNs; 1,481, 1-3 positive LNs; and 660, ≥4 positive LNs. There was wide variation in PMRT use across MiBOQI sites, from 13% to 63.% in patients with 1-3 positive LNs (overall 42%) and from 35% to 91% in patients with 4+ positive LNs (overall 69%). In multivariate analyses stratified by nodal status (0, 1-3, 4+), age ≥ 70 yrs was negatively associated with PMRT. We also noted lower PMRT use in women aged 51-69 with 0 and 1-3 positive LNs (Table 1). Table 1. Multivariate analysis: PMRT and age (+)LNs = 0(+)LNs = 1-3(+)LNs ≥ 4AgeOdds ratiop-valueOdds ratiop-valueOdds ratiop-value≤50 y (reference)1.0 1.0 1.0 51-69 y0.70 (0.49 - 0.98)0.030.69 (0.54 - 0.88)&amp;lt;0.00011.03 (0.63 - 1.65)&amp;lt;0.0001≥70 y0.60 (0.37 - 0.96) 0.26 (0.18 - 0.37) 0.31 (0.18 - 0.54) In the 0 and 1-3 positive node groups, PMRT use was strongly associated with T category and close or positive margin status (Table 2). Table 2: Surgical characteristics and PMRT uptake (+)LNs = 0(+)LNs = 1-3(+)LNs ≥4VARIABLEOdds ratiop-valueOdds ratiop-valueOdds ratiop-valueMargin status Negative (reference)1.0 1.0 1.0 Positive16.7 (10.5 - 26.7)&amp;lt;0.00012.27 (1.32 - 3.90)0.0010.69 (0.32 - 1.47)0.90Close (&amp;lt;1 mm)4.63 (3.03 - 7.10) 1.90 (1.29 - 2.80) 0.97 (0.55 - 1.69) T category T0 to T2 (reference)1.0 1.0 1.0 T3 to T426.2 (16.4 - 41.8)&amp;lt;0.00013.24 (2.19 - 4.78)&amp;lt;0.00010.86 (0.57 - 1.28)0.45 Finally, there was no association between PMRT use and hormone receptor status, HER2 status, Charlson comorbidity index, or reconstructive surgery at the time of mastectomy. Conclusions: PMRT use across Michigan was lower than the American College of Surgeons Commission on Cancer target of 90% in patients with 4 or more positive LNs. Contrary to common belief, immediate reconstructive surgery was not independently associated with decreased PMRT uptake. Understanding reasons for nonuse of PMRT may lead to interventions to increase its use by MiBOQI member institutions in patients for whom it is indicated. Citation Format: Gorski DH, Braun T, Munir K, Griggs JJ, Breslin TM, Henry NL. Factors affecting the administration of post-mastectomy radiation therapy (PMRT) in Michigan [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-10-08.

Associations between use of the 21-gene recurrence score assay and chemotherapy regimen selection in a statewide registry.

The 21-gene recurrence score (RS) assay predicts response to adjuvant chemotherapy in patients with early-stage, hormone receptor-positive, human epidermal growth factor receptor 2 (HER2)-negative invasive breast cancer, but to the authors' knowledge, the role of the assay in guiding the selection of chemotherapy regimen has not been established. The current study was conducted to examine patterns of use of the RS assay for selecting chemotherapy regimens across a statewide registry from 2006 through 2013. Demographic, pathologic, and treatment data were abstracted from medical records for 16,666 women with breast cancer who were treated at 25 hospital systems across Michigan that were participating in the Michigan Breast Oncology Quality Initiative. Treatment patterns were examined based on the RS assay test result. Approximately 25% of patients with lymph node-negative disease who underwent testing with the RS assay and who were treated with chemotherapy received an anthracycline-based regimen, compared with 49% of patients with lymph node-negative disease who were treated with chemotherapy and who had not undergone testing with the RS assay. Of those patients with lymph node-positive disease who underwent testing with the RS assay and who received chemotherapy, 31% received an anthracycline-based regimen. In comparison, 71% of patients with lymph node-positive, chemotherapy-treated disease who did not undergo testing received an anthracycline. From 2006 through 2013, there was a statistically significant decrease in the use of anthracycline-containing regimens in both patients with lymph node-negative and lymph node-positive disease. Use of anthracycline-containing chemotherapy regimens in eligible patients appears to vary with use of the RS assay, despite the lack of evidence supporting use of the assay to guide regimen selection. Results of ongoing prospective trials should help to define the role of the RS assay in this setting. Cancer 2017;123:948-56. © 2016 American Cancer Society.

Unwarranted variation in time to initiation of adjuvant chemotherapy in a statewide collaborative.

264 Background: The purpose of this study was to investigate clinical and non-clinical factors associated with time to initiation of breast cancer adjuvant chemotherapy following the final definitive surgery in a state-wide collaborative. Methods: We collected prospective data from all patients who received adjuvant chemotherapy for Stages I - III breast cancer in 25 health systems in the Michigan Breast Oncology Quality Initiative (MiBOQI) between 2006 - 2014. MiBOQI is a Blue Cross Blue Shield of Michigan/Blue Care Network-sponsored quality collaborative. The dependent variable was time to initiation of adjuvant chemotherapy. Time to chemotherapy was examined according to clinical and non-clinical patient and treatment factors and by health system, both with and without adjustment for patient and treatment variables. Results: Complete data were available for 7,513 of 8,236 eligible (91%) patients. The mean time to chemotherapy was 43 days (range 5-390). In multivariate analyses, clinical factors independently associated with longer time to chemotherapy included increasing age, greater comorbidity, and obesity (p value &lt; 0.0001 for all). Higher stage disease was associated with shorter time to chemotherapy (p &lt; 0.0001); there was no association with hormone receptor or HER2 status. Treatment factors associated with longer time to chemotherapy were breast reconstruction, receipt of genomic assays, and receipt of advanced imaging (p value &lt; 0.0001 for all). Non-clinical factors—minority status, higher area-level poverty, and non-private insurance—were also independently associated with longer time to chemotherapy (p value &lt; 0.0001 for all). In addition, time to chemotherapy varied significantly between health system (p &lt; 0.0001); this difference persisted after adjustment for all patient and treatment factors. Conclusions: Clinical and non-clinical factors, including site of care, are associated with variation in time to initiation of chemotherapy. Identifying patient-level factors and processes of care that facilitate or impede timely initiation of chemotherapy may decrease unwarranted variation in care delivery across a collaborative network.

Variation in use of advanced imaging at the time of breast cancer diagnosis in a statewide registry.

287 Background: Advanced imaging for diagnosis of metastatic disease at the time of diagnosis of stage 0-II breast cancer is not recommended according to national guidelines. The purpose of this study was to examine the variation in ordering of CT chest, abdomen, and pelvis, PET, and bone scans across the state of Michigan, and to investigate clinical and non-clinical factors associated with test ordering. Methods: We prospectively collected data from all patients diagnosed with stage 0-II breast cancer from 2008-2014 in 25 health systems in the Michigan Breast Oncology Quality Initiative (MiBOQI), a Blue Cross Blue Shield of Michigan/Blue Care Network-sponsored quality initiative. Demographic, imaging, and pathologic data were abstracted from the medical record. The primary endpoint was the percentage of patients with at least one advanced imaging test performed within 90 days after diagnosis of breast cancer. Analyses were performed using the chi-squared test of association. Results: Of the 27,011 patients, 4799 (17.8%) had at least one staging exam performed within 90 days of diagnosis. Imaging tests were primarily performed post-operatively. The average percentage of patients who had at least one staging exam performed varied by stage and by site (Table). The rate of testing decreased over time for stage 0-IIA cancers but not for stage IIB. On multivariate analysis, younger age, black race, hormone receptor negativity, HER2 positivity, and higher grade were associated with increased likelihood of undergoing imaging. Conclusions: There is considerable variability in ordering of staging scans at the time of diagnosis of early stage breast cancer across Michigan. Factors underlying use of imaging in stage IIB disease should be examined in order to develop interventions for rate reduction. [Table: see text]

Abstract P5-08-07: Appropriate use of the 21-gene recurrence score (RS) assay across Michigan

Abstract Background: The 21-gene RS assay is used to assess prognosis and to predict response to adjuvant chemotherapy in patients with early stage hormone receptor positive, Her2 negative invasive breast cancer. The National Comprehensive Cancer Network (NCCN) first recommended consideration of testing of appropriate patients with the RS assay in 2008. We examined trends in the use of testing with the RS assay in hospitals across Michigan from 2006 through 2013 using data from the Michigan Breast Oncology Quality Initiative (MiBOQI), a Blue Cross Blue Shield of Michigan/Blue Care Network-sponsored quality initiative. Methods: Demographic, pathologic, and treatment data for women with breast cancer treated at all 25 hospitals participating in MiBOQI were abstracted from the medical record. Patients were excluded if they had stage 0 or IV disease at diagnosis, received neoadjuvant therapy, had bilateral breast cancer, or had a prior history of breast cancer. The primary outcome was the percentage of patients eligible for testing according to NCCN criteria (version 2010) who underwent testing with the RS assay. Analyses were performed using the statistical software R, Version 3.0.1. Results: Of the 18,046 patients in the MiBOQI Registry from 2006-2013 who met inclusion and exclusion criteria, 7133 (39.5%) met the NCCN criteria for testing (eligible). The rate of testing increased from 2006 to 2013 in both the eligible and ineligible cohorts, and varied by site. 21-Gene Recurrence Score (RS) Testing in MiBOQI, 2006-2013FactorEligible (n=7133)Ineligible (n=10913)Tested with RS assay3920 (55.5%)1424 (13.0%)Testing rate in 200643.8%7.3%Testing rate in 201362.3%19.6%Testing by site, 2006-2013 (range)35.8% - 73.3%6.8% - 26.2% Testing of the eligible cohort was statistically significantly associated with younger age, lower tumor grade, and lack of nodal involvement. Overall, 73.4% of patients whose tumors were tested with the RS assay met the NCCN criteria for testing and were deemed appropriately tested. This rate of appropriate testing ranged from 60.8% to 85.4% across sites. Of all patients who underwent testing, 498 (9.3%) had 1 or more positive lymph nodes (&amp;gt;0.2 cm). Receipt of chemotherapy was lower in eligible patients who were tested compared to those not tested (25.5% vs 29.9%, p&amp;lt;0.001). Of the 2387 eligible patients with RS &amp;lt; 18, 117 (5.5%) received chemotherapy, which ranged from 0% - 13.6% across the 25 sites. Of the 341 patients with RS &amp;gt; 30, 56 (9.8%) did not receive treatment with chemotherapy, which ranged from 0% - 50% across the sites. Of the 1192 patients with RS 18-30, 502 (45.7%) received chemotherapy, ranging from 14.5% for RS 18 to 72.5% for RS 30. Conclusions: In sites across Michigan the majority of patients whose tumors were tested with the RS assay were in accordance with the NCCN guidelines, although there was considerable variability across sites. The rate of testing for patients who do not meet the NCCN criteria is increasing. There is very low inappropriate use of the recurrence score for making chemotherapy treatment decisions. Citation Format: Ali HY, Munir K, Braun T, Griggs JJ, Silver SM, Gorski DH, Breslin TM, Henry NL. Appropriate use of the 21-gene recurrence score (RS) assay across Michigan. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P5-08-07.

Abstract P5-08-08: Use across Michigan of the 21-gene recurrence score (RS) assay in lymph node positive patients with breast cancer

Abstract Background: Standard of care for women with lymph node positive breast cancer includes treatment with chemotherapy. The 21-gene RS assay is indicated to assess prognosis and to predict response to adjuvant chemotherapy in patients with early stage hormone receptor-positive, HER2-negative invasive breast cancer. Findings from the SWOG S8814 clinical trial, published in 2010, suggested utility for using the RS assay for treatment decision making in node-positive patients in order to withhold chemotherapy. The Center for Medicare and Medicaid Services subsequently approved coverage for the use of the RS assay in women with up to 3 involved lymph nodes, and the National Comprehensive Cancer Network recently recommended consideration of testing in this population. We examined trends in the use of testing of patients with node positive breast cancer with the RS assay in hospitals across Michigan from 2006 through 2013 using data from the Michigan Breast Oncology Quality Initiative (MiBOQI), a Blue Cross Blue Shield of Michigan/Blue Care Network-sponsored quality initiative. Methods: Demographic, pathologic, and treatment data for women with breast cancer treated at all 25 hospitals participating in MiBOQI were abstracted from the medical record. Patients were excluded if they had stage 0 or IV disease at diagnosis, received neoadjuvant therapy, had bilateral breast cancer, or had a prior history of breast cancer. The primary endpoint was the percentage of patients with lymph node positive, hormone receptor-positive, HER2-negative breast cancer who underwent testing with the RS assay. Analyses were performed using the statistical software R, Version 3.0.1. Results: Of the 30,992 patients included in the MiBOQI Registry from 2006-2013, 2526 (10.7%) had hormone receptor positive, HER2 negative, lymph node positive disease and met the eligibility criteria. The rate of testing with the RS assay in this patient cohort increased from 0% in 2006 to 32.5% in 2013, including an increase from 15.4% in 2010 to 28.3% in 2011. Median age of the tested cohort was 60 (range 26-87). On multivariate analysis, testing was statistically significantly associated with older age, smaller tumor size, 1-3 involved lymph nodes, and lower Charlson Comorbidity Index. Receipt of chemotherapy was lower in those patients who underwent testing compared with those not tested (40.0% vs 82.0%, p&amp;lt;0.001). Chemotherapy was administered to 105 (27.4%) of the patients with RS &amp;lt; 18, 91 (51.4%) of the patients with RS 18-30, and 49 (92.4%) of the patients with RS &amp;gt; 30. Conclusions: Use of the RS assay for assessment of women with involved lymph nodes is increasing over time, primarily in older patients and patients with lower risk disease, and is associated with decreased treatment with chemotherapy. Results of the ongoing SWOG S1007 clinical trial, which is assessing the impact of use of the RS assay on breast cancer outcomes, are eagerly awaited. Citation Format: Henry NL, Ali H, Braun T, Munir K, Silver SM, Griggs JJ, Breslin TM, Gorski DH. Use across Michigan of the 21-gene recurrence score (RS) assay in lymph node positive patients with breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P5-08-08.

Abstract P1-01-13: Patterns of definitive axillary management in the era prior to reporting ACOSOG Z0011: comparison between NCCN Centers and hospitals in Michigan

Abstract Background: The results of the ACOSOG- Z0011 trial have had potential practice changing implications for the management of patients with positive sentinel lymph node (SLN) undergoing lumpectomy and radiation for breast cancer. However, some evidence suggests a shift in axillary management even prior to the initial report of data supporting sentinel lymph node biopsy (SLNB) alone in mid-2010. We analyzed data in the National Comprehensive Cancer Network (NCCN) outcomes database from NCCN centers and the Michigan Breast Oncology Quality Initiative (MiBOQI) hospitals to examine institutional practice patterns with respect to use of completion axillary dissection (CALND) for SLN positive breast cancer in the years leading up to publication of these trial results. We hypothesized that CALND would be omitted more frequently in women treated at NCCN centers compared to those treated at MiBOQI programs. Methods: We identified 2,172 women with clinical T1/T2 N0 breast cancer who underwent breast surgery and SLNB and had a positive SLN from 2007 through 2010 at one of 12 participating NCCN centers or 12 MiBOQI sites. Patient and tumor characteristics, definitive breast procedure, year of diagnosis, and institutional affiliation were analyzed as predictors of use of SLNB alone in univariate Chi-Square and multivariable logistic regression models. Results: CALND was omitted in 314 (14.5%) of the 2,172 patients. Over time, there was a dramatic increase in the use of SLNB alone (12% in 2007 to 23% in 2010). In the univariate analyses, increased patient age, later year of diagnosis, lower T stage, and lower pathologic N stage were significant predictors of use of SLNB alone (all p &amp;lt; .0001). There was no association between definitive breast surgery type, hormone receptor status, Her-2 Neu status, or institutional affiliation and use of SLNB alone. In the multivariable model, older age at diagnosis, later year of diagnosis, and lower pathologic N stage remained significant independent predictors of SLNB alone. There were no significant differences in rates of omission of CALND between NCCN and MIBOQI sites. Conclusions: Omission of CALND occurred frequently in women with SLN positive breast cancer cared for in both NCCN and MiBOQI institutions in advance of reporting results of ACOSOG-Z0011. This shift was seen in management of patients undergoing lumpectomy as well as mastectomy. Further study is warranted to determine the extent of durable practice changes as well as any impact on survival and local-regional control. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P1-01-13.

Use of diagnostic testing for patients with early-stage breast cancer in a multi-hospital quality collaborative in Michigan.

242 Background: In 2012, ASCO identified five opportunities to improve the quality of cancer care and reduce costs. Two of these recommendations focused on limiting use of advanced diagnostic imaging with computed tomography (CT), Positron Emission Tomography (PET), and radionuclide bone scans (BS) at diagnosis and during surveillance for women with early stage breast cancer with a low risk of metastasis. We describe the use of diagnostic imaging among patients with early stage breast cancer treated at hospitals in Michigan. Methods: The Michigan Breast Oncology Quality Initiative is a collaboration between Michigan hospitals, the University of Michigan Comprehensive Cancer Center, and the National Comprehensive Cancer Network (NCCN). This collaborative collects longitudinal demographic, staging, treatment, and follow up data using the NCCN Breast Outcomes Database platform. We analyzed use of advanced diagnostic testing with CT, PET, BS and contrast it with use of traditional imaging (mammography/ultrasound) at diagnosis and during follow up in patients with stage 0, I, II breast cancer between 1998 to 2009. Patients who died or recurred were excluded. Results: The cohort included 7,632 patients (19.6% stage 0, 48.1% stage I, 32.3% stage II) treated at 17 hospitals. Use of traditional imaging was documented in nearly 100% of patients at diagnosis and throughout the follow up period. With respect to advanced diagnostic imaging, 20.3 % underwent CT, 4.7% underwent PET, and 11.6% underwent BS at diagnosis. Advanced testing use decreased over time (Table). Conclusions: Despite published guidelines, which recommend against their routine use, advanced diagnostic imaging use was prevalent among patients treated for early breast cancer at MiBOQI hospitals. Education efforts should target physicians and patients on the lack of proven benefit and potential risks of using advanced diagnostic imaging in this population. [Table: see text]

Improving breast cancer care through a regional quality collaborative

Regional collaborative organizations provide an effective structure for improving the quality of surgical care. With low complication rates and a long latency between surgical care and outcomes such as survival and local recurrence, quality measurement in breast cancer surgery is ideally suited to process measures. Diagnostic biopsy technique for breast cancer diagnosis is measurable and amenable to change at the provider level. We present initial results from our analysis of institutional variation in surgical and core needle biopsy use within a regional breast cancer quality collaborative. Established in 2006, the Michigan Breast Oncology Quality Initiative (MiBOQI) consists of 18 hospitals collecting data on breast cancer care using the National Comprehensive Cancer Centers Network (NCCN) Oncology Outcomes Database Project platform to analyze and compare breast cancer practices and outcomes amongst member institutions. Institutional review board approval is obtained at each site. Data are submitted electronically to the NCCN and analyzed for concordance with practice guidelines. Aggregate and blinded data are shared with project directors and institutions at collaborative meetings, and ongoing practice patterns are observed for change. We analyzed variation in breast biopsy technique for initial cancer diagnosis over time and between institutions. Diagnostic biopsies were categorized as core needle, surgical excisional, surgical incisional, and other surgical biopsy. Procedural data for 8,066 patients treated for breast cancer between November 1, 2006 and December 31, 2009 were analyzed. The mean patient age was 59.5 years (range, 25.4-90.0 years). Within MiBOQI, 21% of patients underwent surgical biopsy for initial diagnosis. The percentage of patients undergoing surgical biopsy ranged from 8% to 37%, and the majority of surgical biopsies were classified as excisional biopsies. Patients with ductal carcinoma in situ were more likely to undergo surgical biopsy compared to those with invasive cancer (30.4% vs 17.8%; P < .001). There was no association between biopsy type and patient age, race, or comorbidity. Data on biopsy technique were shared with site project directors and a target surgical biopsy rate of <15% was chosen by consensus. Site project directors disseminated the data to their institutions and developed action plans for provider and patient education. Over the study period, the percentage of cases undergoing surgical biopsy for the entire MiBOQI collaborative decreased from 21% to 15% (P < .001). The regional quality collaborative model can be used to collect, analyze, and disseminate surgical breast care quality data to organizations and treating physicians. These data can be used to describe patterns of care and make comparisons over time and between organizations. These data can also be used to set regional quality standards and provide an avenue for physician-led quality improvement.

Clinical presentation of breast cancer: Age, stage, and treatment modalities in a contemporary cohort of Michigan women.

1 Background: The U.S. Preventative Services Task Force (USPSTF) revised their breast cancer screening recommendations in 2009 changing their stance on age and frequency of mammography screening to biannual exams starting at age 50. The purpose of this study is to analyze data from a statewide breast cancer registry managed by the Michigan Breast Oncology Quality Initiative (MiBOQI) to determine the impact of these new recommendations on diagnosing and treating breast cancer. Methods: De-identified data were collected on women participating in the MiBOQI registry at seventeen statewide institutions from 2006-2009. Data included method of detection, cancer stage, age at detection, treatment type, and patient demographics. Participants were stratified based on age, and data analyzed using NCSS software (chi-square and t-tests). Results:5903 women with breast cancer with an average age at diagnosis of 59.4 years were included. 65.5% of breast cancers were detected via mammography, 29.8% by palpation, and 4.7% by other methods. In women under 50 years of age, cancers were detected by mammography in 48.3% (69.3% &gt;50yr) and by palpation in 46.1% (24.1% &gt;50yr). Patients with palpable presentations were younger (55.8 yr vs. 61.2 yr; p &lt;0.001). Cancers with a palpable presentation were diagnosed at higher stages (50.0% stage II) than with mammography (52.5% stage I; p&lt;0.001). Breast conservation surgery occurred more frequently than mastectomy (palpation 54.2%; mammography 72.9%, p&lt;0.001) but cancers detected through palpation were more likely to undergo mastectomy procedures. Conclusions: Results of this study validate the importance of annual screening mammography in women older than 50 years, and women aged 40 to 49 years recently omitted from screening guidelines. There was an increased prevalence of palpation for the method of detection in women less than 50 years of age. If screening mammography is omitted in this group, cancers when detected may be of a more advanced stage and result in more mastectomies. This study also supports the use of palpation as a method of detection despite recent recommendations against teaching self breast exams by USPSTF.

Use of a 21-gene reverse transcriptase-polymerase chain reaction (RT-PCR) assay to guide therapy in 14 Michigan Breast Oncology Quality Initiative (MiBOQI) sites.

6039 Background: A 21-gene RT-PCR assay (the assay) has been recommended to predict the risk of recurrence and likelihood of adjuvant chemotherapy (AC) benefit in a subgroup of newly diagnosed patients with breast cancer (BC) in National Comprehensive Cancer Network guidelines (NCCN GLs) since 2008. MiBOQI, a statewide quality improvement initiative, used the NCCN BC Outcome Database Project (BCOD) model to determine the extent and appropriateness of assay use and impact of recurrence scores (RS) on AC use. Methods: 4,594 women diagnosed with unilateral breast cancer with completed follow-up were enrolled between November 2006 and June 2009 from 14 sites. The BCOD prospectively collects data on newly diagnosed BC including the results of gene expression analysis. The NCCN BC GL considers the assay an appropriate option when evaluating BC primary tumors as 0.6-1.0 cm with unfavorable features or > 1 cm, and node-negative, hormone receptor-positive and HER2-negative. Results: 471 out of 1,458 (32%) women appropriately had the assay. 90% of those were < 70 years old and 53% had a low RS, 37% an intermediate RS, and 9% a high RS. In women < 70 with appropriate characteristics (n = 1,035), 36% who received the assay had AC, while 44% who did not have the assay received AC (p = 0.006). 31/217 (14%) with low RS, 79/152 (52%) with intermediate RS, and 33/37 (89%) with high RS received AC (p < 0.001 between RS groups). For women < 70 who had at least one characteristic not consistent with NCCN GL recommended assay use, 142/2,289 (6%) received the assay. Among patients < 70 who were otherwise appropriate assay candidates, those with HER2-positivity and those with tumors < 0.6 cm or 0.6-1.0 cm with no unfavorable features had assay use rates of 17% (95% CI, 12-25%) and 18% (95% CI, 14-22%). Conclusions: The 21-gene RT-PCR assay is frequently used in an appropriate group of women with BC. Those who receive the test have a significant reduction in the use of AC. HER2-positivity and small tumors were associated with nonconcordant assay use. The assay guided use of AC, but 14% and 11% of patients in low and high RS groups had nonconcordant use of AC. Author Disclosure Employment or Leadership Position Consultant or Advisory Role Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration American Society of Clinical Oncology, BlueCross Blue Shield of MI, National Comprehensive Cancer Network Genomic Health Genomic Health BlueCross BlueShield of MI, BlueCross BlueShield of Michigan, National Comprehensive Cancer Network