Search

422 Background: ALP is a PI3Kα inhibitor and degrader approved with fulvestrant for the treatment (tx) of patients (pts) with PIK3CA-mutated, HR+, HER2– ABC. Hyperglycemia (HG) and rash are expected adverse events with ALP tx and remain a challenge for physicians and pts. Management guidance is primarily based on clinical trial experience, which is not necessarily reflective of real-world pts. Here we report guidance for optimizing prevention and management of HG and rash in pts taking ALP based on an integrated Delphi panel, a systematic, validated approach to organize consensus from experts in the absence of definitive evidence. Methods: Two modified Delphi panels were conducted, focusing on HG and rash, respectively. Each panel included 4 oncologists, 4 endocrinologists or dermatologists, 1 clinical pharmacist, and 1 pt advocate. Experts were asked to rate appropriateness of 908 interventions for HG and 348 for rash on hypothetical pt scenarios on a 1 (highly inappropriate) to 9 (highly appropriate) scale. Results were reviewed at virtual meetings, after which experts repeated the rating. The level of agreement or disagreement was determined using the median scores and dispersion from the final rating, and this level of agreement was used to develop consensus statements and tx algorithms. Results: Per the HG panel, (a) ALP tx is appropriate in individuals with HbA1c 6.5% to < 8% with a pre-tx endocrinology consult; (b) low carbohydrate diet is appropriate in all pts starting ALP; (c) prophylactic metformin is appropriate in pts with baseline HbA1c 5.7%-6.4%; may also be appropriate in pts with HbA1c < 5.7%; (d) after metformin, an SGLT2 inhibitor or a thiazolidinedione is an appropriate second-/third-line anti-HG agent (or first-line in metformin-intolerant pts), while insulin is not. Per the rash panel, (a) prophylactic nonsedating (NS) H1 antihistamines (standard dose) are appropriate for all pts; (b) starting/escalating NS H1 antihistamines and topical steroids (TS) is the preferred first step for managing rash; (c) it is appropriate to add, but not replace with, a sedating H1 antihistamine, if response to high-dose, NS option is inadequate, and to add an H2 antihistamine if response is still inadequate; (d) it is appropriate to hold ALP and start oral corticosteroids (OCS) if rash affects > 30% body surface area and is recurrent or has moderate/severe symptoms; (e) if angioedema is present, it is appropriate to either hold ALP and start OCS, or permanently discontinue ALP tx. Conclusions: Until further evidence is available, these expert recommendations provide guidance on prevention and management of HG and rash related to ALP tx in routine clinical practice.

e18584 Background: Cancer progression rates following diagnosis are readily measured. However, the progression rate of cancer during the preclinical sojourn time is generally unobserved. Understanding the duration of preclinical stages (“dwell time”) would allow clinicians to better identify appropriate screening intervals for cancer. We therefore elicited estimates of progression rate during the preclinical sojourn time for a wide variety of malignancies from a panel of clinical experts. Methods: We used a validated consensus methodology (RAND/UCLA modified Delphi panel method) to elicit per-stage dwell time estimates for 20 solid cancers and lymphoma from experts. Eleven experienced oncologists (general and subspecialists) from community and academic centers reviewed literature on the natural history of disease and estimated in number of years (

e18369 Background: Treatment regimens involving 2+ novel oncologics have improved health outcomes in several tumor types. These regimens have significantly higher costs than single agents and older treatments (e.g., chemotherapies), which are still widely used to treat different cancers. Our goal was to determine if current concepts of “value,” such as those described by ICER and other frameworks, adequately capture the value of combination therapies. Methods: Using a RAND/UCLA modified Delphi panel, 8 experts from various backgrounds were provided with a review of current concepts of value and asked to rate them on measurability, relevance, and necessity to determine a combination therapy’s value from 4 perspectives (patient, physician, payer, society). After the first round of ratings, panelists met in person and discussed areas of disagreement. Ratings were repeated, and results used to quantitatively summarize group opinion on concepts recommended for inclusion in a value definition. Results: In both rating rounds, experts agreed treatment, clinical evidence, and health outcomes as important domains in determining value. Experts disagreed on whether societal/cultural beliefs, disease factors (i.e., rarity of cancer, unmet need, burden of disease), and other elements of value (e.g., insurance value, reduction in uncertainty, treatment affordability) needed to be incorporated into value assessments of combinations. Responses differed by perspective. Concepts on which there was disagreement decreased post-meeting (23% to 9%). Conclusions: Experts agreed that “value” for 2+ novel oncologics would have a similar definition to value for all high cost oncology therapies. Four key research opportunities to characterize the value of combination therapy emerged: societal context and patient preference may affect value assessments but are not widely considered in current models; some important benefits are not recognized by patients and may be missed by traditional value assessments; given typical patient and societal preferences, cancer treatments may be systematically undervalued vis a vis other health conditions; and combination therapies may present challenges for recently promulgated value frameworks.

337 Background: This study aimed to assess treatment adherence, healthcare resource utilization and costs in patients with GI NETs who initiated pharmacologic treatments. Methods: In 2 US commercial claims databases, patients ≥18 years with ≥1 inpatient or ≥2 outpatient claims for GI NETs were identified. The first claim for pharmacologic treatments (e.g. somatostatin analogues [SSAs], cytotoxic chemotherapy [CC], targeted therapy) following diagnosis and between 7/1/2009 and 12/31/2013 was defined as the index date. A 6-month clean period before index date and a 6-month pre- and a ≥1-year post-index enrollment were required. Proportion of days covered (PDC) was calculated during the follow up period. Outcomes were reported separately for patients with 1 and 2 years of post-index enrollment. Descriptive statistics including means, standard deviations, and frequencies and percentages for continuous and categorical data, respectively, were reported. Results: Of 1,322 patients with 1-year of follow-up, 847 initiated SSAs, 397 CC, 35 targeted therapies, 2 interferon, and 41 various combinations. Due to sample sizes, remaining results focus only on SSAs and CC. Mean PDC (SD) was 0.669 (0.331) for SSAs and 0.466 (0.236) for CC; SSA users had 20.5 (13.5) office visits and 0.59 (1.03) hospitalizations, CC users had 30.5 (19.8) and 0.89 (1.45) office visits and hospitalizations respectively; total annual cost for SSA-treated patients during the 1st year was $99,691 (82,423) and $134,912 (116,078) for CC. Among 685 patients with 2 years of follow-up, the annual mean costs in the second year were $8,071 and $58,944 lower than the first year for SSAs and CC, respectively. Conclusions: In this descriptive non-comparative study, we reported the resource utilization and costs associated with different treatment therapies. Overall, costs were higher in the first year than in the second year. This 2-database study offers new information on the magnitude and trends in the cost of pharmacologically treated GI NETs. Additional research with a larger sample size would be needed to better understand real-world utilization and costs for GI NET patients treated with different pharmacological therapies.

e15641 Background: CC pts with biliary stents or catheters are prone to have obstruction/infection and studies show 20-40% present with fever/or jaundice requiring urgent treatment for which there is no uniform guideline. We aimed to develop an expert panel consensus on this topic using the modified RAND/UCLA Delphi process to rate treatment appropriateness. Methods: We recruited 13 physician experts from relevant specialty, geography, and practice settings. Patient scenarios were developed and panelists rated the therapies before and after a face-to-face discussion. The appropriateness of various therapies was rated on a 1-9 scale and classified as appropriate, inappropriate, or uncertain. Scenarios with > 2 ratings of 1-3 (inappropriate) and > 2 ratings of 7-9 (appropriate) were considered to have disagreement and were not assigned an appropriateness rating. Results: Panelists were from all US regions (92%) and the UK (8%); had practiced for a mean 16.5 years (4-33 years). Panelists rated 480 scenarios before the meeting, but re-rated only 288 clinical scenarios as they felt that ongoing treatment with chemotherapy did not influence decision-making. Disagreement decreased from 37.5% before the meeting to 10.4% after. Consensus statements are summarized in the table below. Conclusions: The Delphi process produced consensus guidelines to fill an unmet need in urgent management of ascending cholangitis in pts with CC. (Support: The Cholangiocarcinoma Foundation). [Table: see text]

e18333 Background: This cohort study estimated cost and healthcare use among DLBCL patients who progressed beyond 1L therapy. Methods: Using the MarketScan database, patients ≥18 years old who had R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone) as 1L therapy and a claim for DLBCL (ICD-9-CM 200.7X) in the year before or 90 days after 1L start were selected for 2 cohorts: nonrelapsers who had no second line (2L) treatment for ≥2 years and 2L patients with non-R-CHOP chemotherapy after 1L. Patients not continuously enrolled for 6 months before and ≥2 years after 1L initiation, or those treated with maintenance rituximab or lenalidomide were excluded. 2L was subgrouped based on time between 1L end and 2L start: refractory (≤3 months), early (4-12 months), and late relapse (>12 months). Descriptive statistics compared cost and use between groups. Subsequent 2-year and 3-year data will be presented. Results: There were 217 (15.8%) 2L and 1,157 (84.2%) nonrelapse patients identified. Among 2L patients, 61 had hematopoietic stem cell transplant (HSCT) within 1 year. Of these HSCTs, 42.6% were early relapse patients vs. 24.6% late and 32.8% refractory (p=0.05). Mean total cost was $210,488 (SD $172,851) in 2L and $25,044 (SD $32,441) in non-relapsers (p

e15690 Background: Carcinoid syndrome (CS) is associated with a significant symptom burden and deleterious impacts on quality of life. Patients may experience delays in CS diagnosis of 5 to 7 years from initial onset of symptoms, as symptoms may be mistaken for other diseases, such as irritable bowel syndrome or menopause. The current study attempts to detect factors predictive of CS prior to CS diagnosis among patients with GI NETs. Methods: Using a US administrative claims database (IMS PharMetrics) and a retrospective case-control study design, patients (≥18 years) who were newly diagnosed with GI NETs without CS (controls) were exactly matched to patients with CS (cases) based on NET diagnosis date (month and year) at a 3-to-1 ratio. For CS patients, the study index date was the first CS diagnosis date. Each control patient was assigned to have the same index month and year as the matched case patient. The most frequently observed conditions other than symptoms/diagnoses known to be associated with CS during the 1 year prior to the index date were assessed. These conditions were entered into a forward-stepwise logistic regression model to derive predictive factors. Similar methods were applied to another administrative claims database (Truven Health Analytics MarketScan) to validate these predictors. Results: In the development database 1,004 patients with GI NETs were identified, among whom 251 (25%) had CS and 753 (75%) were controls. In both databases, three factors prior to CS diagnosis were found consistently associated with higher risks for CS, including liver disorder [odds ratio (95% CI): 3.38 (2.07 - 5.51)], enlargement of lymph nodes [2.13 (1.10 - 4.11)], and abdominal mass [3.79 (1.87 - 7.69)]. Conclusions: By assessing patients with GI NET from two independent US claim databases, this study suggested that patients diagnosed with CS were 2-3 times more likely to be diagnosed with liver disorder, enlargement of lymph nodes, or abdominal mass, than those without CS during the one year prior to CS diagnosis. Future studies using patients’ medical charts are warranted to validate and interpret the findings. These findings, when validated, may aid physicians to diagnose CS patients earlier.

6603 Background: Little is known about the validity and reliability of value assessment frameworks. Methods: Eight panelists used the ASCO, ESMO, ICER, and NCCN frameworks to conduct value assessments of 15 drugs for advanced lung and breast cancers and castration refractory prostate cancer. Panelists received instructions and published clinical data to complete the assessments, assigning each drug a numeric or letter score. We used Kendall’s W coefficient to measure convergent validity by cancer type among frameworks and intraclass correlation coefficients (ICC) to measure framework inter-rater reliability across cancers. Panelists were surveyed on their experiences. Results: Kendall’s W for breast, lung, and prostate cancer drugs were 0.560 ( p= 0.010), 0.562 ( p= 0.010), and 0.920 ( p< 0.001), respectively. Pairwise and subdomain W are shown in the table. ICC (95% CI) for ASCO, ESMO, ICER, and NCCN were 0.800 (0.660-0.913), 0.818 (0.686-0.921), 0.652 (0.466-0.834), and 0.153 (0.045-0.371), respectively. Panelists generally agreed the frameworks were logically organized and easy to use. Conclusions: Convergent validity among the frameworks was fair to excellent, increasing with clinical benefit subdomain concordance and simplicity of drug trial data. Inter-rater reliability, highest for ASCO and ESMO, improved with clarity of instructions and specificity of score definitions. Continued use, analyses, and refinements of the frameworks will bring us closer to using value-based treatment decisions to improve patient care and outcomes. [Table: see text]

210 Background: With the accumulation of RWD in healthcare, CER continues to expand. RWD are becoming increasingly relevant in oncology, particularly since the onset of care pathways and CMS’ Oncology Care Model pilot program; yet, CER in oncology presents several challenges as disease biology differs by cancer type and many RWD sources do not capture clinical response, progression or survival. This review examined common endpoints reported in RWD studies on CER and TxP in MBC, focusing on HER2-negative and Triple Negative (TN) MBC. Methods: PubMed (2006-January 2016) and 4 conferences (2011-January 2016)—ASCO and SABC meetings/symposiums—were searched using MeSH/keywords, e.g., metastatic breast cancer, treatment, and comparative effectiveness. RWD CER and TxP studies in U.S. patients with HER2-negative or TNMBC were included; clinical trials were excluded. Results: Of1,782 total records, 17 articles and 9 conference abstracts were included. Studies using RWD increased over time with 2 studies published in 2010, 1 in 2012, 6 in 2013, 6 in 2014, 10 in 2015, and 1 as of January 2016. Of these, 8 were CER and 18 examined TxP. Most studies were retrospective chart reviews (7 CER; 10 TxP studies), others were retrospective secondary database analyses (1 CER; 6 TxP) and physician surveys (2 TxP). RWD sources included commercial insurance claims, SEER-Medicare, and California Cancer Registry data. Nineteen studies reported results in patients with HER2-negative MBC and 7 reported in TNMBC patients. Primary endpoints included overall survival (OS), progression-free survival (PFS), and treatment duration (TD). CER studies most commonly reported TD and survival outcomes (e.g., OS, PFS), each reported in 75% of the studies. TxP studies also most commonly examined survival outcomes (61% of studies), in addition to various treatment patterns and duration outcomes. Conclusions: This literature review indicates that in parallel to the availability of RWD, published CER studies and analysis of treatment patterns have grown in the last 5 years. The most commonly reported outcomes include OS, TD and PFS.

452 Background: CC pts with biliary stents or catheters often present with fever and/or jaundice requiring urgent treatment for which there is no uniform guideline. We aimed to develop an expert panel consensus on this topic using the modified RAND/UCLA Delphi process to rate treatment appropriateness. Methods: We recruited 13 physician experts from relevant specialty, geography, and practice settings. Patient scenarios were developed based on a literature review, and therapies were rated by the experts before and after a face-to-face discussion. The appropriateness of various therapies was rated on a 1-9 scale and classified as appropriate, inappropriate, or uncertain. Scenarios with > 2 ratings of 1-3 (inappropriate) and > 2 ratings of 7-9 (appropriate) were considered to have disagreement and were not assigned an appropriateness rating. Results: Panelists were from all US regions (92%) and the UK (8%); had practiced for a mean 16.5 years (4-33 years) and reported seeing an average of 120 unique CC patients a year (0-900 pts). Panelists rated 288 clinical scenarios. Experts decided that ongoing treatment with chemotherapy did not influence decision-making. Disagreement decreased from 37.5% before the meeting to 10.4% after. Consensus statements are summarized in table 1. Conclusions: The Delphi process produced consensus guidelines to fill an unmet need in urgent management of ascending cholangitis in pts with CC. Studies of the impact of these guidelines on cost of care and pt outcomes are warranted. (Support: The Cholangiocarcinoma Foundation)[Table: see text]

387 Background: NET comprise a broad set of rare tumors. Almost 2/3 arise in the gastrointestinal (GI) tract. NCCN guidelines for unresectable GI NET recommend somatostatin analogues (SSA) first, however, guidelines do not recommend a particular sequence for remaining therapies. Our objective is to describe real world treatment patterns of GI NET. Methods: This retrospective study combined 2 claims databases to examine newly pharmacologically treated patients using tabular and graphical techniques. Treatments included SSA, cytotoxic chemotherapy (CC), targeted therapy (TT), interferon (IF) and combinations. Patients ≥ 18 years, with ≥ 1 inpatient or ≥ 2 outpatient claims for GI NET who received pharmacologic treatment from 7/1/09-6/30/14 were identified. A 6 month clean period prior to first treatment ensured patients were newly treated. Patients were followed until end of enrollment. Results: We identified 2,258 newly treated GI NET patients. 59.6% started first line therapy with SSA monotherapy, 33.3% CC, 3.6% TT, and 0.5% IF. The remainder received combination therapies. Mean follow up was 576 days. Mean duration of first line therapy was 361 days for all newly treated patients (449 SSA, 215 CC, 267 TT). 58.9% of patients had no subsequent pharmarcological treatment after discontinuation of first line therapy.The most common second line was combination therapy with SSA (i.e., CC or TT added). In patients who did not begin with SSA, most received SSA as second line. In graphical pattern analysis, there was no clear pattern visible after first line therapy. Conclusions: More than 1/2 of pharmacologically treated patients began treatment with SSA and 1/3 with CC, with duration of use > 1 year and just over 6 months, respectively. We found treatment patterns after first line were unclear, and more than 1/2 of patients had no subsequent pharmacological treatments after discontinuing first line therapy. It is unclear whether there is underutilization of the pharmacological therapy or this is due to stable disease status after alternative non-pharmacological treatment. Future studies directed at understanding treatment patterns using patient medical records are warranted. As new treatments emerge, reassessment of treatment patterns may be needed.

e13072Background: The majority of NETs occur in the gastrointestinal (GI) and bronchopulmonary system. The overall incidence of GI NET in the US appears to have increased from the 1970s to 2004 wit...

e18250Background: In response to rising costs in oncology care, several organizations have developed frameworks to systematically assess the value of oncology drugs. The frameworks produce regimen-...

e13071Background: The majority of NETs occur in the gastrointestinal (GI) and bronchopulmonary system. The overall incidence of lung NET in the US appears to have increased from the 1970s to 2004 w...

450 Background: PNETs, rare neoplasms originating from secretory neuroendocrine cells, produce peptides and neuroamines causing hormonal syndromes, including carcinoid syndrome (CS). Somatostatin analogues (SSAs) are usual 1st-line treatment; everolimus and sunitinib are targeted therapies (TT) approved for progressive disease. Carcinoid or hormonal symptoms associated with CS may occur in pNET and are treated with SSAs. Information is limited on impact of these symptoms on health resource use in pNET. Methods: This retrospective cohort study merged 2 large HIPAA-compliant claims databases and identified pNET patients (ICD-9-CM 157.4) initiating treatment with TT or cytotoxic chemo (CC) from 2009-2012. Patients had no prior TT or CC for pNET for 1 year preindex and were continuously enrolled 1 year postindex. Descriptive statistics were reported for comorbidities and health care use, overall and by presence of symptoms defined to include: diarrhea, nausea/vomiting, flushing, cardiac palpitations, asthma, wheezing, fatigue, hypotension, or dizziness. Chi-square and t-tests were used for analyses of categorical and continuous variables, respectively. Results: This study included 338 newly treated pNET patients with 1 year followup; 237 1st-line CC and 101 TT (64 everolimus, 37 sunitinib). Mean age was 54.5; 46% female. Patients had mean of 5 (SD 2.0) chronic conditions and Charlson Index of 10.1 (SD 2.9). Utilization included 43.5% hospitalizations, 48.8% ED visits, and mean of 25.5 (SD 24.8) office visits. 212 (62.7%) patients had symptoms. Other conditions were infection (34.9%) and anemia (29.9%). PNET patients with symptoms had considerably higher rates of hospitalization, ED visits, and office visits than those without: 57.5% vs. 19.8%, 64.2% vs. 23.0%, mean 29.2 vs. 19.3, all P < 0.001. Conclusions: In this study, nearly 2/3 of treated pNET patients had carcinoid or hormonal symptoms; about 1/3 had infection and anemia. Symptoms may influence drug treatment patterns. PNET patients with symptoms were high utilizers of health resources; over half were treated as inpatients and more sought ED care. This increased use may result from a high burden of symptoms and comorbid conditions.

108 Background: When integrated into routine oncology care, palliative care can improve symptom burden, quality of life, and patient and caregiver satisfaction. However, not all oncology practices have access to subspecialty palliative medicine for referral. This project endeavored to define what constitutes high-quality primary palliative care as delivered by medical oncology practices. Methods: An expert steering committee developed a list of 966 palliative care service items, divided into nine domains, each describing an aspect of palliative care delivery for patients with advanced cancer. Using a modified Delphi methodology, 31 multi-disciplinary panelists ranked each service item according to importance, feasibility and scope. Results: Panelists endorsed the highest proportion of palliative care service items in the domains of End-of-Life Care (81%); Communication and Shared Decision-Making (79%); and Care Planning (78%). Lowest proportions were in: Spiritual and Cultural Assessment and Management (35%) and Psychosocial Assessment and Management (39%). In the largest domain, Symptom Assessment and Management, there was consensus that all symptoms should be assessed and managed at a basic level with more comprehensive management for common symptoms such as nausea, vomiting, diarrhea, dyspnea and pain. Under the domain of Appropriate Palliative Care and Hospice Referral, there was consensus that oncology practices should be able to describe the difference between palliative care and hospice to patients and refer patients with an expected survival of under 3 months or poor performance status (Zubrod 3-4) to hospice. Conclusions: This statement describes high-quality palliative care for patients with advanced cancer or high symptom burden, as delivered by oncology practices. Oncology providers wishing to enhance palliative care delivery may find the guidance useful to inform operational changes and quality improvement efforts.

11050 Background: About 20% of breast cancers (BC) are DCIS. Local recurrence (LR) rates with surgery alone are 15-60%; about 50% are invasive. Radiation (XRT) reduces LR by 50% but has not impacte...

269 Background: Uncontrolled chemotherapy-induced nausea and vomiting (CINV) can lead to nutrient depletion, diminished function, disruption of chemotherapy, and increased costs. Standard antiemetic therapy includes 5-HT3RAs for CINV prophylaxis, with palonosetron recommended in National Comprehensive Cancer Network (NCCN), Multinational Association of Supportive Care in Cancer (MASCC), and ASCO guidelines as the preferred 5-HT3RA for CINV prophylaxis with MEC. There is evidence that using 5-HT3RAs can reduce costs but no comprehensive review of the evidence is available. Methods: We searched MEDLINE, National Institute for Health Research (NIHR), Centre for Reviews and Dissemination (CRD databases, 4 conferences (Academy of Managed Care Pharmacy, ASCO, International Society for Pharmacoeconomics and Outcomes Research, MASCC), and bibliographies of included articles. We queried Medical Subject Headings (MeSH) and key terms: “ondansetron,” “granisetron,” “palonosetron,” “dolasetron mesylate,” “costs,” “cost analysis,” and “economics.” Included records reported data on cost/utilization (rescue medication, outpatient/inpatient services) related to 5-HT3RA use for CINV in English, in human subjects, and published after 1997. Results: Of the 433 identified records, the 16 reporting utilization in the US were reviewed (excluded: 29 duplicates, 388 off-topic records). Studies varied significantly in designs, patients, 5-HT3RA regimens, and definition of outcomes. Twelve studies reported rescue medication use for CINV in patients using different 5-HT3RAs. In 5 studies, fewer patients treated with palonosetron required rescue medication versus ondansetron users (56% vs. 61%, 28% vs. 83%, 14% vs. 24%, 8% vs. 11%, 6% vs. 11%); 2 studies found palonosetron users had fewer outpatient services versus ondansetron users (5% vs. 10%, 8% vs. 10%). Four studies, with a variety of patients and outcomes, reported fewer patients treated with palonosetron versus ondansetron or other 5-HT3RAs used inpatient care (e.g., 0.2% vs. 0.4%, 16% vs. 23%, 7% vs. 10%, 0% vs. 5%), while 2 studies reported similar use (1% vs. 1%, 0% vs. 0%). Conclusions: CINV prophylaxis with palonosetron is generally associated with lower use of rescue medications, outpatient and inpatient services compared to ondansetron or other 5-HT3RAs. Use of palonosetron as a standard treatment may lead to reduced utilization of rescue medications and healthcare services for CINV and subsequent cost savings.

e22124 Background: Many stage II and most stage I non-small cell lung cancer (NSCLC) patients forgo adjuvant chemotherapy, despite the likely presence of occult metastasis in the 30-60% of these patients who die within 5 years of surgery, and despite demonstration that patients with undetected, i.e. very early, metastasis can achieve long-term survival with this intervention. A multi-gene assay using quantitative RT-PCR and run on formalin-fixed, paraffin-embedded samples (Pervenio Lung RS) has been independently validated in two large international studies to better differentiate high- and low-risk non-squamous NSCLC patients than either traditional risk assessment criteria or conventional staging. A survey of physicians who used this multi-gene assay probed whether better risk discrimination changed post-operative management. Methods: Physicians were queried via a web-based survey on practice patterns, patient characteristics and treatment recommendations. The survey was developed through cognitive interviews with oncologists and thoracic surgeons. Physicians were allowed to respond regarding up to 4 patients (stage I or II) for whom the multi-gene test was ordered. Results: Forty-one physicians (12 surgeons, 29 oncologists/internists) of 61 contacted (67%) provided responses on 77 patients. Community practices accounted for 83%, and the average years in practice was 15. Seventy-one percent provided responses on either 1 or 2 patients. Mean patient age was 66±10 years, with stage I and stage II comprising 73% and 27%, respectively. Prior to obtaining assay results, chemotherapy was recommended in 30% of patients, including 5 stage I patients. The assay characterized 5-year mortality risk as high, intermediate and low in 52%, 23% and 25% of patients, respectively. Treatment recommendations were changed post-assay in 36% of patients, including 73% of high-risk patients for whom observation was recommended initially. Change in treatment intensity correlated with the degree of assay-assessed risk. Conclusions: Refinement of risk assessment achieved via a multi-gene prognostic assay resulted in changes in treatment recommendations for a substantial number of early stage NSCLC patients.

e15104 Background: Molecular classification of tumors has been incorporated into clinical algorithms for patients with metastatic cancers of uncertain origin. This study is a comprehensive cost analysis—characterizing standard of care for metastatic disease, cost implications of correct vs incorrect diagnosis, and cost effectiveness of implementing molecular classification into the diagnostic paradigm. Methods: A deterministic model was developed to project the cost and effectiveness of standard care for metastatic tumors of uncertain origin, compared with incorporating use of the 92-gene assay at 3 points: before immunohistochemical (IHC) analysis, or after 1, or 2 sets of IHC stains. For each strategy, total and incremental costs, life-years (LYs), quality-adjusted life years (QALYs), and incremental cost effectiveness ratios (ICERs) were projected, as well as the % of patients treated correctly vs incorrectly. Model inputs and costs were based on published literature, analyses of the SEER database, the Physicians’ Fee and Coding Guide 2012, and interviews with disease experts. Sensitivity analyses were performed on key inputs. Results: When standard care was compared with each 92-gene assay strategy, costs increased between $4,804 and $6,038 and QA-survival by 1.03-1.15 months. Using the 92-gene assay after 1 set of IHC stains was the most cost effective strategy ($50,273/QALY). Overall, the 92-gene assay increased the % of patients treated correctly and decreased the % of patients treated with empiric therapy compared with standard care. When compared with incorrect treatment, correct treatment provided an additional 5.4 QA-months of survival. Sensitivity analyses identified that costs of empiric treatment, the 92-gene assay, and lung cancer treatment had the greatest effect on outcomes. At a societal willingness-to-pay (WTP) threshold of $100,000, the 92-gene assay was cost effective 94% to 99% of the time. Conclusions: Use of the 92-gene assay is considered cost effective using current societal WTP thresholds, represents an efficient use of economic resources that improves patients' survival and quality of life, and could contribute to standardization of the costs of diagnosis.

6057 Background: Expensive treatments and a growing number of cancer patients have resulted in increased spending on cancer care. Within a framework we developed for measuring the value of quality improvement (QI), we describe the cost-effectiveness of improving compliance with cancer screening measures compared to other quality measures. Methods: We used our framework to examine 18 Healthcare Effectiveness Data and Information Set (HEDIS) 2010 quality measures, synthesize related cost-effectiveness (CE) data and describe measure-specific QI-adjusted incremental cost-effectiveness ratios (ICERs). For each measure we: 1) quantified current compliance; 2) reviewed literature for ICERs; 3) estimated per-person steady state cost and quality-adjusted life years (QALYs); 4) estimated affected population size; 5) estimated the cost of QI; and 6) calculated QI-adjusted ICERs at full compliance, defined as 95%. We assumed per-person QI costs did not change with compliance and varied this in sensitivity analyses. We compared QI-adjusted ICERs for 3 cancer screening measures to the remaining measures. Results: Published ICERs for the cancer screening measures were $43,180/QALY (breast), $5,102/QALY (cervix) and $15,173/QALY (colon) and for other measures from $195/QALY (drug treatment) to $35,616/QALY (flu shots). Incorporating QI costs for cancer screening measures gave QI-adjusted ICERs of $64,549/QALY (breast), $15,463/QALY (cervix) and $22,991/QALY (colon), respectively. Incorporating QI costs for all 18 measures resulted in QI-adjusted ICERs from $195/QALY (drug treatment) to $9,075,868/QALY(antidepressant management), with a median of $15,463/QALY. Reaching 95% compliance with the 3 cancer measures would cost $5.1 billion and add 160,000 QALYs ($32,640/QALY) and with all 18 measures would cost $13.4 billion and add 5.8 million QALYs ($2,313/QALY). Conclusions: Improving compliance with cancer screening may be cost-effective at a threshold of $50k/QALY, although improving care on all HEDIS measures may be even more cost-effective. Accurate assessment of the cost of increasing cancer screening requires integration of both the cost-effectiveness of the screening tests and the cost of the QI programs needed to change practice.

e14610 Background: Neuroendocrine tumors (NETs) comprise mostly carcinoid or pancreatic NETs and are rare with symptoms that may be difficult to control. Current treatment guidelines lack some specificity. We summarize an expert panel consensus on medical treatment of well-differentiated unresectable midgut NETs. Methods: Consensus statements were developed via RAND/UCLA Delphi process, which involved a diverse group of physician experts (e.g., by specialty, geography, practice) developing comprehensive clinical patient scenarios and rating the scenarios on the appropriateness of various medical therapies before and after a face-to-face meeting. Experts and moderator were blinded to funding source. Scenarios were rated on a 1-9 scale and were labeled as appropriate, inappropriate, or uncertain. Scenarios with >2 ratings in 1-3 and >2 in 7-9 range were considered to have disagreement and were not assigned an appropriateness rating. Results: Panelists (age: 38-63 years) were from the northeast, midwest, south, and west regions. Specialties represented were medical and surgical oncology, interventional radiology, and gastroenterology. Panelists had practiced for a mean 15.5 years (range: 6-33). Panelists rated 202 scenarios. The proportion for which there was disagreement decreased from 11.7% (23 scenarios) before the meeting to 4.5% (9) after. Post-meeting, 49% (99 scenarios) were rated inappropriate, 29.7% (60) were uncertain, and 16.8% (34) were appropriate. Consensus statements from the scenarios included: 1) it is appropriate to use somatostatin analogs (SA) as 1st-line therapy in all patients, 2) it is appropriate to increase the dose/frequency of octreotide-LAR as 2nd-line therapy in patients with uncontrolled symptoms up to 60 mg every 4 weeks or up to 40 mg every 3 or 4 weeks for refractory carcinoid syndrome. Other treatment options may also be appropriate in 2nd-line. Conclusions: Treatment consensus obtained in this study is concordant with NCCN recommendations. The Delphi process allowed quantification of ratings in a systematic and reliable way while improving consensus in a group of physicians on the appropriateness of medical therapies in midgut NETs.

3626 Background: The Oncotype DX Colon Cancer Recurrence Score (RS) has been clinically validated as an independent predictor of individual recurrence risk in stage II colon cancer patients following surgery. As a result, physicians have been ordering the Oncotype DX assay for stage II colon cancer patients since January 2010, yet no data exist on the assay’s impact on adjuvant treatment planning in clinical practice. We performed a survey to characterize the impact of the assay on adjuvant treatment recommendations in stage II colon cancer. Methods: U.S. medical oncologists (n=346) who ordered Oncotype DX for ≥3 patients with stage II colon cancer were contacted and asked to complete a web-based survey regarding the single most recent stage II colon cancer patient for whom the assay was ordered. The survey was developed through cognitive interviews with four medical oncologists, and the protocol was institutional review board approved. Results: We analyzed survey results from 116 eligible physicians. Physicians were more often in community (86%) than academic or other practices, and had a median of 14.5 years (range, 2-40) of practice experience. The median patient age was 62 years (range, 32–85). Most patients (81%) had T3 disease. Patients had ≤8, 9-11 and ≥12 nodes examined 3%, 13% and 84% of the time and 38% had comorbidities. Of the 76 patients tested for MMR/MSI, 13 (17%) were MMR-D or MSI-H and 46 (61%) were MMR-P or MSI-L; 17 (22%) unknown. Median RS was 20 (range, 1-77). Before obtaining the RS, chemotherapy was planned in 52 (45%) patients, observation in 40 (34%), and there was no recommendation in 24 (21%). For the 92 patients with pre-assay recommendations, recommended treatment changed after obtaining the RS for 27 patients (29%). Treatment intensity decreased for 18 (67%) and increased for 9 (33%) of 27 changed recommendations. A significant trend of decreasing treatment intensity with lower RS values was observed (p=.0035). Conclusions: These findings indicate that for stage II colon cancer patients, treatment recommendations were changed by RS results in 29% of patients. Use of the Oncotype DX assay may lead to reductions in treatment intensity, contributing to the assay’s cost effectiveness.

398 Background: The Oncotype DX Colon Cancer Recurrence Score (RS) has been clinically validated as an independent predictor of individual recurrence risk in stage II colon cancer patients following surgery. As a result, physicians have been ordering the Oncotype DX assay for stage II colon cancer patients since January 2010, yet no data exist on the assay’s impact on adjuvant treatment planning in practice. We performed a survey to characterize the impact of the assay on adjuvant treatment recommendations in stage II colon cancer. Methods: U.S. medical oncologists (N=277) who ordered Oncotype DX for ≥3 patients with stage II colon cancer were contacted and asked to complete a web-based survey regarding the single most recent stage II colon cancer patient for whom the assay was ordered. The survey was developed through cognitive interviews with four medical oncologists, and the protocol was institutional review board approved. Results: As a planned preliminary analysis, we analyzed surveys from 92 eligible physicians. Physicians were more often in community (85%) than academic or other practices, and had a median of 14.5 years (range, 2-40) of practice experience. The median patient age was 62 years (range, 34–81). 84% of patients had T3 disease. Patients had ≤8, 9-11 and ≥12 nodes examined 2%, 14% and 84% of the time and 36% had comorbidities. Of the 60 patients tested for MMR/MSI, 9 (15%) were MMR-D or MSI-high and 37 (62%) were MMR-P or MSI-low; 14 (23%) unknown. Median RS was 20 (range, 1-77). Before obtaining the RS, chemotherapy was planned in 38 (41%) patients, observation in 35 (38%), and there was no recommendation in 19 (21%). For the 73 patients with pre-assay recommendations, recommended treatment changed after obtaining the RS for 23 patients (32%), including changes from chemotherapy to observation and vice-versa. Conclusions: These preliminary findings indicate that for stage II colon cancer patients, treatment recommendations were changed by RS results approximately one-third of the time. Final results will be reported to include accrual through December 2011.

6552 Background: Myelodysplastic syndrome (MDS) affects about 1 in 10,000 individuals. After diagnosis, patients are treated with hypomethylating agents (HMA) and thalidomide analogues (TA) or rece...

6560 Background: Myelodysplastic syndrome (MDS) is rare in people under 50. Little is known about the disease in this group, particularly among those patients who receive supportive care only. Methods: This was a descriptive cohort study using a large commercial claims database. The study included patients with an initial MDS claim (ICD-9-CM 238.72-238.75) between 2/1/2007 and 7/31/2008, who were continuously enrolled for 6 months prior to and 12 months following the index claim. Patients were excluded it they were treated with either hypomethylating agents (HMA) or thalidomide analogues (TA). Once identified, patients were stratified into two group, those aged

641 Background: The similarity of the prognostic performance of different gene expression sets is thought to result from the identification of common tumor phenotypes. Knowledge of the components of composite expression scores provides insight into the relative contributions of biological processes involved and may better inform therapeutic decisions. In this study, we assessed the correlation of a constructed composite Oncotype Dx Recurrence Score (cRS) using the published algorithm with a reported 14-gene Metastasis Score (MS), individual gene and group components of the cRS, as well as a composite score of MS and progesterone receptor (PGR). Methods: We examined the RNA expression profile of 114 formalin-fixed, paraffin-embedded sections from ER-positive breast cancer cases from University of Muenster. The individual genes of the signatures reported by Paik et al (21-gene recurrence score) and Tutt et al (14-gene metastasis score) were profiled using RT-PCR. The changes in gene expression relative to r...

e11077 Background: Results of recent Oncotype DX studies indicate that the recurrence score (RS) can identify patients with node-positive, estrogen receptor-positive (N+/ER+) breast cancer who may ...

21009 Background: We sought to validate a previously developed 14-gene prognostic signature and a metastasis score (MS) that predicted distant metastasis in N-, ER+ breast cancer patients in an independent sample set of patients without systemic treatment. The genes consisted primarily of proliferation genes involved in p53 and TNF signaling pathways. Methods: : A cohort of 294 N-, ER+ breast cancer patients from Guy's Hospitals, London, UK without systemic therapy were tested. The cohort had a mean age of 55.5 yrs with 49% > 55 yrs, mean tumor size of 1.93 (max. 3) cm, and a median follow up of 14.3 yrs. The primary endpoint was distant metastasis free survival. RT-PCR was carried out on fixed sections. The MS was calculated. Results: The mean MS (SD) was 0.44 (0.59) with a range of -1.31 to 2.0. Hazard of distant metastasis increased 3.02 fold (95% CI 1.91–4.76, p No significant financial relationships to disclose.

3651 Background: Molecular markers that predict inherent resistance or sensitivity to chemotherapy would greatly assist oncologists in selecting drugs with the best therapeutic index. Genetic alter...