184 Background: The initial ASCO “Top 5” list, created as part of the Choosing Wisely campaign, recommends against use of imaging for staging of early stage breast cancer in asymptomatic women at low risk for metastasis. The objective of this study was to measure and compare use of imaging for staging in two large integrated health care systems, Kaiser Permanente (KP) and Intermountain Healthcare (IH). We also sought to distinguish whether imaging was used for routine staging or for diagnostic purposes. Methods: We identified stage 0-IIB breast cancer patients diagnosed between January 1, 2010 and December 31, 2012 with first primary malignancy from tumor registries in three KP regions (Southern California, Northwest, and Mid-Atlantic) and IH. Using the KP and IH electronic health records, we identified use of imaging tests (PET, CT, bone scan) during the staging window (30 days prior to diagnosis up to initial surgery). We performed chart abstraction on a random sample of patients who received an imaging test to identify indication. Results: For the total sample of 10,014, mean age at diagnosis was 60 (range 22-99); with 21% stage 0, 47% stage I, 32% stage II. Overall, 8% of patients (792 patients) received at least one imaging test during the staging window, including 8% at KP and 6% at IH (p=0.0005). Chart abstraction (N=129) revealed that overall, almost half of all imaging tests (48%) were performed to evaluate a symptom, sign or prior imaging finding, including 55% at KP and 32% at IH. Conclusions: Use of imaging for staging of low-risk breast cancer was very low in both health care systems, with clinically trivial differences between them. Approximately half of imaging services were in response to a sign or symptom. Strategies to reduce use of imaging at staging for early stage breast cancer patients within these health care systems are unlikely to yield meaningful improvement. [Table: see text]

Objective. Unfavorable histology endometrial carcinomas confer worse prognosis. We determined the association of adjuvant radiation on local recurrence and survival for unfavorable, early stage endometrial cancer.Methods. We retrospectively identified 125 patients who had a hysterectomy for early stage (FIGO IA), unfavorable histology (clear cell, papillary serous or grade 3 endometrioid), endometrial carcinoma treated between 1992 and 2011. Patients were restaged according to current FIGO 2009 guidelines. Primary endpoint was local control and secondary endpoints were distant recurrence and overall survival.Results. The median age of the cohort was 67 years old with a mean follow up 152 months. Adjuvant radiation was delivered in 60 patients (48%). There were a total of 24 recurrences; 5 had local-regional recurrences, 4 local and distant recurrence, 12 distant only recurrences, and 3 had unspecified recurrences. The 5-year local-regional control was 97.8% in patients who received radiation and 80.1% in patients who did not receive radiation (p = 0.018). The 5-year overall survival rate was 68.1% if patients did not receive radiation and 84.9% if they did receive radiation (p = 0.0062). On univariate analysis, only radiation (HR 0.12, 95% CI: 0.03 to 0.49, p-value = 0.018) was associated with a significant increase in local relapse free survival.Conclusions. Adjuvant radiation therapy was significantly associated with an improvement in local-regional control and overall survival in patients with unfavorable histology, early stage endometrial cancer. (C) 2014 Elsevier B.V. All rights reserved.

Objectives: To determine the effect of adjuvant radiation on local recurrence for high grade histology, early stage (FIGO IA) endometrial cancer.

Abstract Background: Prognostically relevant biomarkers in breast cancers can be assayed using nucleic acid (e.g. OncoType DX) or protein based (e.g. immunohistochemistry (IHC)) detection stratagies. Nucleic acid tests measure many more markers compared to immunohistochemistry, but are weighted heavily for ER, PR, and HER2 in their prognostic scoring algorithms. Recent data suggest that a panel of 4 IHC stains (IHC4) may have equivalent prognostic ability to multigene assays. The aim of this study was to retrospectively compare IHC4 scores to OncoType DX recurrence scores in a large cohort of patients in our statewide health system. Methods: Formalin-fixed paraffin-embedded breast cancer samples previously submitted for OncoType DX testing between 2008 and 2011 were obtained and sections stained for ER, PR, HER2, and Ki67 (the IHC4 staining panel). Stained slides were scanned digitally and scored using customized pathologist assisted image analysis algorithms (Leica Biosystems Aperio, Vista CA). The resulting data were used in the modified “Magee Equation 3” (Klein et al. Mod Pathol. 2013. PMID: 23503643) to calculate an IHC4 recurrence score on a numerically comparable scale to that used in the OncoType score. The IHC4 and Oncotype DX scores were directly compared for concordance and correlation. Results: 280 breast cancer samples comprised the study cohort. The mean (median) IHC4 score was 20.13 (19), range: 11-42. The mean (median) OncoType score was 18.48 (16), range: 0-75. Using cut-offs of 0-17 for low risk, 18-30 for intermediate risk, and >30 for high risk, the concordance data is shown in Table 1. Comparison of scores from IHC4 and OncoType by Risk Category OncoType RS LowIntermedHigh Low80201IHC4 RSIntermed816819 High029RS = Recurrence Score Pearson's correlation coefficient for the entire set was 0.6669. The overall concordance was 56.1%; one-step discordance 43.6%; and two-step discordance: 0.3%. Eliminating the intermediate scores, the concordance was 98.8%. Conclusions: Reasonable agreement was seen between risk scores derived by IHC4 and OncoType DX. Only one discrepancy among 280 scores would have resulted in a different treatment recommendation (low risk by IHC4 and high risk by OncoType). If the IHC4 score is clearly high or low, one would not expect a dramatically different result from Oncotype DX, and the Oncotype DX test may not be needed. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P6-06-34.

1582 Background: More than 95% of testicular cancer (TC) patients survive at 10 years; however, they develop many complications following diagnosis. Our study is toassess the risk of developing morbidities among TC survivors. Methods: We used the Utah Population Database to identify TC patients (age>13) who 1) were diagnosed between 1994 and 2008, 2) had medical records at the University of Utah and Intermountain Health Care hospitals and 3) survived at least 3 years. Cases were matched to five TC-free patients (controls) from the same hospitals on birth year, birth region and date of residence in Utah. Individuals (cases and controls) with morbidities of interest before the date of TC diagnosis were excluded. Adjusted Cox regression analysis was used to identify the risk of developing ischemic heart disease, renal failure, infertility, hearing loss, peripheral neuropathy, osteoporosis and restrictive lung disease among TC patients relative to controls. The models were adjusted for age, race and other confounders. We also compared the risk of these morbidities among patients with regional/advanced tumor relative to patients with in situ/localized tumor. The latter was adjusted for age and tumor histology. Results: We identified 955 TC patients and 4775 controls. TC stage was: 73.7 % in situ/localized and 26.3% regional/advanced. The hazard rate (HR) of developing any of the target morbidities was 2.57, 95% confidence interval (CI), 2.2-2.9 for patients with localized cancer, 3.85 (95% CI: 2.9-5.1) for patients with regional cancer and 3.3 (95% CI: 2.3-4.8) for patients with advanced cancer relative to controls. Risk of developing renal failure, infertility, peripheral neuropathy, hearing loss and restrictive lung disease was significantly higher among TC survivors (P<0.05). We also identified a significant association (HR 5.2; 95% CI 1.8-15.5) of cancer stage and the risk of developing hearing loss and non-significant positive association of TC stage with the other morbidities. Conclusions: TC survivors were more than twice as likely to develop morbidities following diagnosis when compared to controls. The risk increases when the comparison is stratified by stage, which might be related to the differences in the treatment used for different stages of TC.

113 Background: Mammographic screening for women 40-49 years of age remains controversial based on results from earlier large scale, controlled mammography trials. Methods: From 2002-2006, 871 women aged 40-49 were diagnosed with breast cancer at Intermountain. The charts of all patients without a record of a screening mammogram at Intermountain (n= 436) were reviewed to confirm that they had not had a screening mammogram in the prior two years at any facility (Interval Cancers) and their survival was compared to 435 women who had their cancer diagnosed on a screening exam (Screen Detected). All patients were followed for at least 5 years via the tumor registry. Results: Stage distribution for Screen Detected/Interval cancers was 25.3/6.4% stage 0, 36.8/24.8% stage I, 24.6/35.6% stage II, 6.9/20.4% stage III and 0.5/3.4% stage IV. Overall, 67 patients (7.7%) did not have complete staging data. Overall survival was significantly better (p<.0001) for 40-49 year old women with Screen Detected compared to those with Interval cancers. 702 (79.6%) had ER/PR status recorded (83.5% ER/PR positive). Women with DCIS or LCIS did not have tissue sent for markers. 679 patients (76.9%) had HER2 status recorded (78.8% HER2 neg). Of the patients with both HER2 and ER/PR status recorded 10.4% were “triple negative.” Survival following screening mammography was significantly enhanced for women who were ER/PR positive (p<0.0001), HER2 negative (p=0.0065), or HER2 positive (p=0.0013). Survival was not improved by screening mammography for women who were ER/PR negative (p=0.3818) or for women who were triple negative (p=0.416). Conclusions: A minority of women age 40-49 who develop breast cancer (13%) have biologic features suggestive of aggressive disease and, after 5 years of follow up, they are not benefitted by screening mammography. The remaining 87% are clearly benefitted by screening mammography. Our results suggest that the discrepancies noted in the screening mammography trials in 40-49 year old women may have resulted from population variations in the proportion of women with unfavorable biology. Based in part on these results, we continue to recommend regular screening in the 40-49 year old cohort.

135 Background: Data is essential to achieve meaningful quality improvement. A variety of commercial products are currently available for sophisticated data collection. However, data systems alone are not sufficient to improve quality. Additional resources are required in order to leverage electronic clinical data for meaningful improvement. This abstract outlines the necessary requirements and available methods for data-based oncology quality improvement using the experience of Intermountain Healthcare. Methods: The organizational components required for quality improvement are complex. Successful quality improvement begins with project feasibility, data availability and clinical leadership. Clinical processes are reviewed, data accuracy and availability are confirmed and clinical goals are established. Data collection, validation, and analysis are standardized across multiple facilities and providers. Data must be analyzed and presented in a way that clearly illustrates differences in current performance compared to quality goals, should be tracked over time to ensure real and sustainable progress, and must be combined with other improvement strategies to maximize impact. Results: Once quality reports are generated, a physician champion presents them to clinical staff along with education materials, national guidelines and current evidence from peer-reviewed literature. Clinicians are presented with individualized data comparing their personal performance to de-identified performance of their peers, the facility and the system. Providers are given updated data on a regular basis, the data system is monitored for outliers and the need for subsequent interventions, and additional metrics are tracked to ensure process changes don’t negatively impact quality in other areas. Conclusions: Oncology quality improvement requires both clinical and data infrastructure. Electronic clinical data systems are essential for quality improvement, but are not sufficient by themselves. Additional resources are required to capture, extract, validate and analyze clinical data. Appropriate use of these resources transforms existing electronic data systems into a powerful quality improvement tool.

141 Background: Accountable care is defined as moving the incentives for health care from a system that rewards volume and procedures to one that rewards improvements in the quality of care for a defined population. To prevent this process from deteriorating into solely a cost reduction exercise, physicians, and hospitals need to develop a valid, reproducible, and effective means of measuring quality and impacting behavior to reduce variation and improve quality of care. The Intermountain Healthcare Oncology Clinical Program’s (OCP) experience with Oncology Quality Improvement (OQI) offers several key lessons for enabling this process. Methods: OQI initiatives are developed by a multidisciplinary physician-based team tasked with directing standardization and ensuring optimal care delivery. The team uses clinical knowledge, peer-reviewed literature, and data from an enterprise data warehouse to develop goals. Performance is measured against a goal which focuses on variation between physicians and facilities. Individual physician data is compared to de-identified data of peers, facilities, and the system. A physician champion performs academic detailing for physician groups across the system and is critical to the success of the program. Results: Over the past decade, the OCP initiated over 30 projects designed to measure and improve quality of oncology care delivery. Breast cancer projects included breast conservation in surgical management, reducing axillary dissection for ductal carcinoma in situ and sentinel node biopsy rather than axillary dissection. The OCP also explored standardizing lymph node resection during colorectal cancer surgery and subsequently the utilization of adjuvant chemotherapy. Imaging based goals included improving mammography callback rates and using PET/CT during preoperative assessment of lung cancer. In most instances the process resulted in significant, sustainable OQI. Conclusions: The investment in program and clinician staff is significant, and the requirements and costs for a sophisticated data system are real. However, an OQI program can provide meaningful improvements in the quality of cancer care and is an important step to facilitate the transition to accountable care.

122 Background: Quality improvement initiatives and outcomes research activities at Intermountain Healthcare are regularly conducted across 9 clinical domains, including oncology. The purpose of this abstract is to define some of the projects initiated in the primary management of breast cancer and critically analyze the barriers and catalysts allowing for adoption of quality improvement processes. Methods: Between 2003 and 2010 we initiated approximately 15 projects designed to measure and improve quality of care delivery in the treatment of breast cancer. These projects included increasing the use of breast conservation for primary surgical management, reducing axillary dissections for patients with ductal carcinoma in situ and the adaptation of sentinel node biopsy rather than axillary dissection for primary breast cancer. Quality improvement projects involving diagnostic imaging included evaluation of the time from abnormal mammogram to biopsy and the utility of follow-up mammography following breast conservation. Results: Measured variability among practicing clinicians in many instances was substantially reduced or eliminated. Adoption of clinical improvements was most often rapid and noticeable across the entire health care system, however, in some instances quality processes were not completely adopted. The adoption of quality improvement initiatives occurred where obvious clinical benefit could be documented, technical expertise was enhanced and peer supervision existed. Adoption occurred less frequently when financial barriers to adoption existed, clinical benefit was less well defined and peer encouragement was less vigorous. Conclusions: Quality improvement processes across a large healthcare system can be instituted and adopted by a variety of facilities and practicing clinicians. Barriers and catalysts to adoption do exist and this presentation will attempt to document and outline opportunities for success using quality improvement in a large healthcare system.

6053 Background: Intermountain Healthcare is a network of 24 urban and rural hospitals including 9 comprehensive cancer centers serving Utah and southeastern Idaho. This project builds on the institution's long history of quality improvement and was undertaken to standardize the treatment of colon cancer with respect to surgery, pathology and medical oncology. Methods: A multidisciplinary physician-based development team identified the treatment of colon cancer as an area of interest, collected baseline data to assess current practice, and developed system wide goals to ensure optimal staging and treatment for patients. Following a review of existing literature and an evaluation of national guidelines, the team set a goal to remove and evaluate at least 12 lymph nodes in 90% of stage I-III colon cancer patients who undergo resection with curative intent. The next year the development team added a goal to ensure that 80% of stage III colon cancer patients receive chemotherapy. Both goals were developed and implemented using a similar process of ongoing data collection, analysis and feedback, physician education, regular review of noncompliant charts and constant monitoring by the development team. Both goals are still monitored on an ongoing basis. Results: Lymph node removal data were collected on all stage I-III colon cancer patients undergoing potentially curative colon cancer resection after January 1, 2007. Following standardization, the percentage of cases in which a minimum of 12 lymph nodes were evaluated increased from 78.8% (n = 99) to 95.1% (n = 102) (p = 0.001). Likewise, the percent of stage III colon cancer patients undergoing surgery beginning January 1, 2008 who received chemotherapy increased from 68.2% (n = 44) to 86.5% (n = 37) (p = 0.068). Following standardization, all 37 stage III patients were seen by a medical oncologist in consultation and offered chemotherapy. Conclusions: Practice standardization is crucial to ensuring the adoption and adherence to best practice guidelines and can be achieved in multiple hospitals and among multiple physician specialties. This requires validated organizational processes, physician leadership, ongoing data collection and constant monitoring. No significant financial relationships to disclose.

Abstract Background: Hormone receptor testing is important in the management of women with breast cancer. We previously reported potential adverse effects of variable specimen and prolonged specimen handling conditions on ER (estrogen receptor) test results (2005 SABCS, abstract#5107).Objective: To compare prevalence of ER and PR (progesterone receptor) negative test results following standardization of pre-analytical specimen handling conditions at Intermountain facilities.Methods: Prospective, quasi-experimental study design of 6 Intermountain facilities. Facilities were separated into 2 categories: experimental (2 facilities) and control (4 facilities) groups. Pre-analytical specimen handling conditions (including recording of time to fixative and duration of fixation in neutral buffered formalin) were standardized at experimental facilities but not at control facilities. Standardization consisted of educating operating and grossing room staff about appropriate specimen handling and the value of recording time to fixation and fixative duration as a way to improve pre-analytic standardization. OR staff was called in any cases where times were not recorded. Study population includes women undergoing breast cancer surgery and who were tested for ER/PR status between January 2008 and January 2009. Specimen handling conditions and ER/PR test results were collected manually. Covariates were retrieved from cancer registry and included age, grade, positive lymph nodes, specimen type, and tumor stage. Multivariate logistic regression was used to compare prevalence of ER and PR negativity between experimental and control facilities after controlling for covariates.Results: 1054 women with breast cancer were tested for ER/PR status during the study period. The average age was 60.2 years (59.2 years for control cases and 61.1 years for the experimental group). The overall prevalence of ER and PR negative tests was respectively 18.5% and 27.3%. Average time to fixative at experimental facilities was 18.4 minutes (SE=3.1; 95% CI, 12.2-25.6) and average time in fixative was 18.0 hours (SE=0.4; 95% CI, 17.2-18.8). Compared to experimental facilities, both the prevalence of ER and PR negativity was higher (16.9% vs 19.7%) and (23.9% vs 30.0%) at control facilities. After controlling for covariates there was no difference in prevalence of ER negativity (p=0.13) between the two groups. However, the prevalence of PR negativity remains significantly higher (p<0.01) at control facilities compared to experimental facilities even after controlling for covariates.Conclusions: The prevalence of ER and PR negative results was lower following staff education and recording of pre-analytical specimen handling conditions. Our data suggest that staff education and recording of pre-analytical specimen handling conditions has the potential to optimize hormone receptor test results. It also shows the feasibility of fixing tissue routinely after less than an hour interval between time of breast cancer tissue removal and fixation and underscores the value of standardization of pre-analytic handling as a method to improve ER and PR testing on breast cancer specimens. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 4154.

Abstract Background:Oncotype DX has been validated to quantify the risk of distant recurrence and predict the benefit of chemotherapy (CT) in ER positive, node negative breast cancer treated with tamoxifen. A retrospective study was undertaken to assess the use of the Oncotype DX test at Intermountain Healthcare. Intermountain is a not-for-profit healthcare system with 21 inpatient facilities, 9 comprehensive cancer centers and 42 affiliated medical oncologists ranging from single practitioners to multiphysician groups.Methods: This study contains a group of T1-3 N0 ER+ breast cancer patients who received an Oncotype DX test paired with a control group of non-tested T1-T3 N0 ER+ patients receiving hormone therapy (HT) from the same period. Data comes from a supplemental database containing treatment and follow-up data from individual physician offices combined with data from Intermountain's cancer registry. To ensure data completeness, Genomic Health provided a list of relevant Oncotype DX results for study patients. The analysis was done using multivariate logistic regression and controlled for age, tumor size, grade and T stage.Results: From 2005 to 2008, Oncotype DX testing was performed on tumor specimens from 285 patients. 8 patients had positive nodes and 4 patients were ER negative. In addition, 9 patients did not and will not receive HT (5 refused, 4 contraindicated) and 9 patients have yet to begin HT. 11 patients were lost to follow-up. Of the remaining 244 patients who form the study group, six patients were Her-2 neu positive (1 high recurrence score (RS), 3 intermediate and 2 low). Tumor size ranged from <1 cm to 7 cm, but 80% were <2 cm. 120 study patients (49%) had a low RS, 95 (39%) intermediate and 29 (12%) high. Only 2% of patients in the low RS group received CT, whereas 93% of patients in the high RS group and 40% in the intermediate RS group received CT.An analysis of potential factors affecting CT treatment decision making in the intermediate RS group showed that 60% of the 15 patients under age 50 received CT, compared to 36% of the 80 patients age 50 and over. CT was given in 35% of grade 1, 40% of grade 2 and 50% of grade 3 tumors. 43% of patients with a <1 cm tumor received CT compared to 42% of 1-2 cm tumors and 28% of tumors >2 cm. Compared to our control group of 688 patients, Oncotype DX-tested patients are younger (p<0.01), less likely to have T1a (p=0.03), more likely to have T1c (p<0.01) or T2 (p=0.03) tumors and less likely to undergo CT (p<0.01). Low RS patients are less likely to receive CT (p<0.01) and high RS patients more likely to receive CT (p<0.01), whereas intermediate RS patients showed no significant difference (p=0.07) but were trending toward receiving less CT than the control group.Conclusions:Virtually all patients with a low RS received only HT, while most patients with a high RS also received CT. In patients with an intermediate RS, younger age and higher grade may impact clinicians' decisions to administer CT, while a larger tumor size does not. Overall, patients undergoing Oncotype DX testing are less likely to receive CT. These data indicate that the use of Oncotype DX testing facilitated appropriate therapeutic decisions in most patients. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 6058.

6521 Background: Intermountain Healthcare is a system of 9 major community hospitals based in Salt Lake City, Utah. The system has a long history of quality improvement and previously has standardized the use of breast preservation and the surgical management of ductal carcinoma in situ (DCIS). This project was undertaken to standardize evaluation of the axilla; specifically the incorporation and utilization of sentinel lymph node biopsy as the standard of care. Methods: Baseline data for sentinel node evaluation was obtained from 9 hospitals showing substantial variation between facilities and among surgeons. A physician-based surgical quality improvement team with system wide representation defined a 75% sentinel node biopsy rate as a minimally achievable goal. This allowed for legitimate omissions of sentinel node evaluation in a minority of patients. Processes put in place to increase compliance included: improved infrastructure, physician mentoring and increased accountability. The project was presented at each facility and surgeons were provided with data on their personal performance. Results were confirmed by individual chart review on a quarterly basis. Results: Sentinel node utilization was tracked in 265 patients treated by 67 surgeons from January 1, 2007 to June 30, 2007. Utilization increased from 69.1% to 82.6% system wide and in one facility increased from 39.1% to 90.9%. Utilization of sentinel node evaluation increased the axillary evaluation of patients with DCIS; an unexpected consequence of this project. Conclusions: Variation in surgical management can be decreased in a large hospital system through validated organizational processes. This process requires constant monitoring to assure gains are maintained and other quality indicators are not adversely affected. No significant financial relationships to disclose.

The RAS/RAF signaling pathway is an important mediator of tumor cell proliferation and angiogenesis. The novel bi-aryl urea BAY 43-9006 is a potent inhibitor of Raf-1, a member of the RAF/MEK/ERK signaling pathway. Additional characterization showed that BAY 43-9006 suppresses both wild-type and V599E mutant BRAF activity in vitro. In addition, BAY 43-9006 demonstrated significant activity against several receptor tyrosine kinases involved in neovascularization and tumor progression, including vascular endothelial growth factor receptor (VEGFR)-2, VEGFR-3, platelet-derived growth factor receptor beta, Flt-3, and c-KIT. In cellular mechanistic assays, BAY 43-9006 demonstrated inhibition of the mitogen-activated protein kinase pathway in colon, pancreatic, and breast tumor cell lines expressing mutant KRAS or wild-type or mutant BRAF, whereas non-small-cell lung cancer cell lines expressing mutant KRAS were insensitive to inhibition of the mitogen-activated protein kinase pathway by BAY 43-9006. Potent inhibition of VEGFR-2, platelet-derived growth factor receptor beta, and VEGFR-3 cellular receptor autophosphorylation was also observed for BAY 43-9006. Once daily oral dosing of BAY 43-9006 demonstrated broad-spectrum antitumor activity in colon, breast, and non-small-cell lung cancer xenograft models. Immunohistochemistry demonstrated a close association between inhibition of tumor growth and inhibition of the extracellular signal-regulated kinases (ERKs) 1/2 phosphorylation in two of three xenograft models examined, consistent with inhibition of the RAF/MEK/ERK pathway in some but not all models. Additional analyses of microvessel density and microvessel area in the same tumor sections using antimurine CD31 antibodies demonstrated significant inhibition of neovascularization in all three of the xenograft models. These data demonstrate that BAY 43-9006 is a novel dual action RAF kinase and VEGFR inhibitor that targets tumor cell proliferation and tumor angiogenesis.