Additional Abstracts from the NCCN 20th Annual Conference: Advancing the Standard of Cancer Care™ (2024)

Outcomes and Health Services Research

AB2015-18. Racial Differences in Behavioral Patterns and Quality of Life from Multiple Myeloma Diagnosis and Treatment

Pooja Advani, MD; Radhika Ghosh, MD; David Hodge, MS; and Sikander Ailawadhi, MD

Mayo Clinic, Jacksonville, Florida

Background: Significant improvement in outcomes for multiple myeloma (MM) has occurred in the past decade; marked racial disparities in incidence and survival outcomes have also been reported, which may be attributed to differences in health care access and use. Efforts have been made to enhance patient education, improve quality of life (QoL), and promote survivorship. We undertook a large, international, patient-reported, questionnaire-based study to evaluate both the influence of race on patient knowledge and the psychophysical impact of MM diagnosis and treatment. Methods: An online questionnaire (61 questions) that involved patient demographics, QoL, treatment history, knowledge of MM and its treatment, and satisfaction with treatment and health care staff was administered in association with the International Myeloma Foundation. Some of the questions were mandatory and others were optional. Missing responses were excluded from the individual patient's final analysis. Chi-square and Kruskal-Wallis tests were used to compare categorical and continuous variables, respectively. SAS software (V 9.3) was used with a 2-sided significance level of 0.05. Results: Survey responses were available from 957 of 1324 patients (72%), with a completion rate of 62%. Average survey completion time was 16 minutes. US patients represented 89.5% of the responders, with the majority from New York state (14%). Race was reported by 77% of participants (white [W], 73%; non-white [NW], 4%]. Statistically significant differences in responses between W and NW patients are shown in Table 1. Although not statistically significant, a higher percent of NW patients reported lack of trouble sleeping, being satisfied with their QoL, coping with illness, and having expected treatment effectiveness compared with W patients. NW patients also reported lack of social support and participation in support groups in addition to decreased ability to work. Overall, no racial differences were noted in response to satisfaction with doctor/health care staff. Conclusions: We found that a majority of the patients who accessed the survey were W, which may be indicative of better access to education and technology. Differences were also observed in certain behavioral patterns between W and NW patients, which may be attributed to disparities in socioeconomic status, and knowledge and understanding of the disease. Strategies for improvement of patient awareness and better access and use of health care may help eliminate these differences.

Epidemiology

AB2015-19. Malignant Salivary Gland Tumors: A Single Institution Epidemiology Report

Bilal Ahmad, MD; Sobia Nabeel, MD; Haseeb Saeed, MD; Michael Machiorlatti, MS; Sara K. Vesely, PhD; Mohammed Muqeet Adnan, MD; Mohamad Khawandanah, MD; Wajeeha Razaq, MD; and Mohammad Razaq, MD

Oklahoma University Health Science Center

Background: Salivary gland tumors comprise a rare heterogeneous group of tumors, making up <1% of all cancers, occurring at a rate of about 1 case per 100,000 per year in the United States. These tumors represent 6%–8% of all head and neck cancers. We present a single-institution retrospective analysis of the epidemiology of salivary gland tumors from 1976–2013. Methods: We searched for all salivary gland tumors between June 1976 and Oct 2013. All benign tumors were excluded. Reviewed cases were either outpatient referrals or seen as a first consult at Oklahoma University Health Science Center (OUHSC), which is a major referral center for the state. We stratified patients in different groups according to age, sex, and ethnicity. We also evaluated location of primary tumor, histology, and grade of tumor among different groups. Simple descriptive statistics on counts and percentages for categorical covariates were created, and medians were calculated for the continuous variables. Fisher's exact tests were performed to examine the relationships between sex, age, race, histology, and grade of tumors due to low expected cell counts. SAS V9.2 was used for all analyses. Results: A total of 276 cases of malignant salivary gland tumors were identified. Cases included 117 patients aged <60 years and 159 aged ≥60 years. Median age at the time of diagnosis was 60 years, with 170 (62%) men and 106 (39%) women. Regarding ethnicity, 242 (88%; 95% CI, 84.2–92.0) patients were white, 26 (9%) were African American, 6 (2%) were Native American, and 2 (<1%) were Asian. Based on primary site of tumor location, 226 (82%) tumors were located in the parotid gland, 35 (12.4%) in the submandibular gland, 2 (<1%) in the sublingual gland, and 13 (5%) in salivary glands not otherwise specified (NOS). Based on histologic subtype, 68 (25%) were mucoepidermoid carcinoma, 57 (21%) were squamous cell carcinoma, 30 (11%) were adenoid cystic carcinoma, 26 (9%) were acinar cell carcinoma, 25 (9%) were adenocarcinoma, 20 (7%) were carcinoma NOS, and 50 (18%) were other less common carcinoma. These included 14 (5%) lymphoma, 4 (1%) large cell carcinoma, 2 (<1%) amelanotic melanoma, 2 (<1%) small cell carcinoma, 2 (<1%) sarcoma, 2 (<1%) neuroendocrine carcinoma, 1 (<1%) malignant fibrous histiocytoma, 2 (<1%) papillary cystadenocarcinoma, 2 (<1%) Merkel cell carcinoma, 2 (<1%), myoepithelial carcinoma, 2 (<1%) malignant mixed tumor, 1 (<1%)adenocarcinoma, and 15 (5%) unknown carcinoma. Tumor grade was also studied: 35 (13%) tumors were grade I or well differentiated, 45 (16%) were grade II or moderately differentiated, 85 (31%) were grade III or poorly differentiated, 11 (7%) were grade IV or undifferentiated, and in 100 (36%) cases, grade was not determined, including some lymphomas. Men and women differed statistically by grade and histology of tumors. (P=.0179 and .0017, respectively). Histology was statistically different in the age groups (<60 and ≥ 60 years) as well (P≤.0001). Ethnic groups differed statistically by site (P=.0386) and grade of tumor (P=.0556). In 2013, the US Census Bureau reported that75% of the Oklahoma population was white. This may explain the higher percentage of whites (88%) with salivary gland cancers than expected in our sample. Conclusions: Salivary gland tumors are a rare type of cancer. Based on our epidemiologic study, we concluded that the median age was 60 years at the time of diagnosis. Most cases were diagnosed at ≥60 years of age. Men were at 1.6 times higher risk than women. The most common histology among men was squamous cell carcinoma; it was mucoepidermoid carcinoma among women. The most common histology in ≥60 year age group was squamous cell carcinoma, and it was mucoepidermoid CA in the <60 age group. The parotid

gland was the most common site of primary tumor, with mucoepidermoid carcinoma the most common histology among these tumors. Grade III cancer with poorly differentiated histology was the most common of all tumors. This large retrospective study provides the epidemiology and biopathologic characteristics of salivary gland tumors.

Best Practices in Implementation and Use of Clinical Practice Guidelines

AB2015-20. Pilot Study Evaluating Users' Satisfaction with a Cancer Screening iPhone and iPad Application

Moein Alizadeh, MD; Houda Bahig, MD; and Jean-Paul Bahary, MD

Centre Hospitalier de l'Université de Montréal

Background: In recent years, patients' use of health information technology has increased and has been a catalyst for patient empowerment. The purpose of this project was to evaluate the use in a clinical setting of an iPhone application targeted to the general population that aims at enhancing users' responsibility towards their own health. Methods: International, national, and provincial official cancer screening recommendations were reviewed and combined to develop a comprehensive province-adapted screening program. NCCN Clinical Guidelines in Oncology (NCCN Guidelines) for detection, prevention, and risk reduction of breast, cervical, colorectal, lung, prostate, and genetic familial high-risk assessment for breast and ovarian cancer were collected. A short interactive questionnaire was developed in an application format for the iPhone and the iPad. The questionnaire includes assessment of users' risks factors such as gender, age, and personal and family medical history. An automatic computation of users' individual risks allows for instantaneous formulation of individualized cancer screening recommendations. A multiple-choice survey evaluating user-friendliness of the application, self-assessed impact on empowerment, anxiety induction, and conformity to recommendations was completed by 2 cohorts of users: patients in a radiation oncology follow-up clinic and health care workers at our institution. Results: Preliminary results from 17 patients and 21 health care workers are presented. Among patients, all participants reported strong satisfaction (41%) or satisfaction (58%) with the application and rated it as very user-friendly (88%) or user-friendly (12%). All participants strongly agreed (53%) or agreed (47%) that they felt the tool increased their feeling of empowerment toward their own health. No patient reported increased health-related anxiety. Among health care workers, all participants strongly agreed the tool was user-friendly; 88% reported strong satisfaction and 12% reported satisfaction with its use. Although 19% and 62% of participants strongly agreed or agreed (respectively) that the tool increased their feeling of empowerment toward their own health, 19% disagreed with this statement. One health care worker reported increased health-related anxiety. Conformity to recommendations was 86% and 76% among patients and health care professionals, respectively. Conclusions: This application was generally well received among our pilot group and has the potential to increase empowerment. Further data are currently being obtained. This product should be considered a starting point and not a substitution for a discussion with the treating physician. A large-scale diffusion of this application in the community is planned with governmental partners.

Clinical Oncology

AB2015-21. Standard of Care: 4 Doses of Dose-dense Pacl*taxel in the Adjuvant Treatment of Breast Cancer Has Equal Toxicity Profile to Weekly Pacl*taxel

Mohamed Alsharedi, MD; Jennifer Dotson, DO; Nabiha Elmsherghi, MD; Todd Gress, MD, MPH; and Mria Tirona, MD, FACP

Edwards Comprehensive Cancer Center, Marshall University, Joan C. Edwards School of Medicine

Background: Pacl*taxel-based adjuvant systemic chemotherapies are considered the standard of care in treatment of breast cancer; 2 commonly used pacl*taxel chemotherapy regimens are dose-dense pacl*taxel (ddP) and weekly pacl*taxel (wP). The standard of care ddP is administered on an every 2 week schedule, with a total of 4 doses of 175mg/m². wP is administered every week, total of 12 doses of 80mg/m². Pacl*taxel-induced toxicity might bias clinical oncologists to choose one regimen over the other, favoring wP to avoid toxicities because fewer toxicities were seen with wP in the S0221 trial. However, in S0221, wP was compared with 6 cycles of ddP instead of the standard of care 4 cycles ddP. To our knowledge to date, no data exist comparing toxicities and tolerability between these 2 commonly used standard of care pacl*taxel-based regimens. Methods: We performed a retrospective, single-institution, breast cancer chart review of 121 patients with newly diagnosed breast cancer who were treated with pacl*taxel-based adjuvant chemotherapy from January 2008 to June 2014 at Edwards Comprehensive Cancer Center/Marshall University. Results: Chart review showed that 76 patients were treated with ddP (Group A) and 45 with wP (Group B). We found that the variables for both groups were comparable with no significant difference in age, sex, race, comorbidities, biochemical receptor status, cancer stage, or grade.

We applied CTCAE criteria per NCI to grade most common pacl*taxel-related toxicities (neuropathy, hematologic toxicities, allergies, and musculoskeletal problems). We also compared the rate of treatment discontinuation and dose reduction between the 2 groups. Comparing main pacl*taxel-related toxicities between the groups showed that ddP does not have a more toxic profile than wP and both are equally tolerable (Table 1). Neuropathy, a common side effect that can become debilitating for some patients and require dose reduction or treatment discontinuation, was seen equally in both groups. Conclusions: Comparing the 2 commonly used schedules of pacl*taxel-based adjuvant chemotherapy in breast cancer showed that the standard ddP is comparable to wP regarding toxicities and tolerability. In contrast to the S0221 trial results, which showed more neurotoxicity with the nonstandard 6 cycles of ddP, this retrospective study showed no difference in toxicity profiles between the groups. Furthermore, severe grade (3 or 4) neuropathy was seen in the same percentage in of both groups. Further prospective trials are needed to validate these findings.

Clinical Oncology

AB2015-22. Comparison of EP2006, a Filgrastim Biosimilar, to the Originator: a Phase III, Randomized, Double-Blind Clinical Study in the Prevention of Severe Neutropenia in Patients with Breast Cancer Receiving Myelosuppressive Chemotherapy

Kimberly L Blackwell, MD;^a Vladimir Semiglazov, MD;^b Dmitry Krasnozhon;^c Irina Davidenko, MD, PhD;^d Lidia Nelyubina;^eRoumen Nakov, MD, PhD;^f Gabor Stiegler;^f Pritibha Singh;^f Arnd Schwebig;^f Stefan Kramer;^f and Nadia Harbeck, MD^g

^aDuke University School of Medicine; ^bRailway Clinical Hospital at OJSC; ^cLeningrad Regional Oncological Dispensary; ^dClinical Oncological Dispensary1, Healthcare Department of Krasnodar Territory; ^eRussian Oncology Research Center n.a. N.N.Blochin of RAMS; ^fHexal AG; ^gBreast Center, University of Munich

This abstract was published by Oxford University Press on behalf of the European Society for Medical Oncology. It is available at http://www.researchgate.net/profile/Arnd_Schwebig/publications. Annals of Oncology 06/2015; 26(9). DOI:10.1093/annonc/mdv281.

Clinical Oncology

AB2015-23. Development of a Proposed Biosimilar Product, Apo-Filgrastim, Based on the Demonstration of Physicochemical, Pharmacologic, and Clinical Similarity to Neupogen

Stephen Brokx, PhD; Kalpna Desai, PhD; Sanyukta Kher, MSc; Tina Catalano, MSc; Devanshi Maharja, MTech; Louise Scrocchi, PhD; and Jason Dowd, PhD

Apotex, Inc.

This abstract is available at J Clin Oncol 2015;33 (Suppl): Abstract e13528.

Quality Improvement

AB2015-24. Improving Efficiency and Capacity via Cancer Treatment Scheduling Standardization

Gloria G. Campos, MSIE; Lauren Gjolaj, MBA, BSN, RN; Angela Olier-Pino, DNP, MBA, RN; and Gustavo Fernandez, MD, MBA

Sylvester Comprehensive Cancer Center

Background: With increasing volumes and stagnant capacity, a busy academic cancer center's outpatient infusion area formed a multidisciplinary team to address infusion scheduling ensuring capacity matched scheduled treatment times with the goal of decreasing the variance between actual and scheduled chair times (the difference between the time the patient sat in the chair and the scheduled appointment length) and decreasing patient wait time from arrival to chair. Improving the scheduling of patient chair time allows for better resource management and use of capacity. Methods: Using a multidisciplinary approach, standardized scheduling guidelines were created to address approximate treatment chair time for patients (including time needed for mandatory nursing documentation, pharmacy turnaround time, drug infusion and post infusion monitoring) and nursing acuity. Specific information on

medication (including medications to notify pharmacy prior to preparation, medications with long reconstitution time, specific laboratory values or examinations needed before treatment, such as echocardiogram or multiple uptake gated acquisition scan), were also placed on the guidelines to make them a standardized tool for all disciplines in the treatment unit. Training was given to pharmacy, nursing, and scheduling staff to ensure a standardized approach was used. Master templates were developed to ensure maximum capacity was used with 4 templates (fast track, chairside, research, and bedside) hosting 30 minute slots with staggered start and end times. Nursing schedules were then updated to match maximum capacity in templates. Results: After implementation, a significant reduction was seen in the difference between the scheduled and actual treatment duration, both average and variance, showing a decrease in overestimated durations and better predictability. Additionally, the percentage of patients whose treatment was within 30 minutes of their scheduled duration increased by 50.1% from baseline, and patient arrival to chair time decreased by 33.3% from baseline (Table 1). Conclusions: Creating a multidisciplinary and standardized approach to infusion scheduling, using master templates and providing training to the key stakeholders in the process produced a decrease in difference between the scheduled and actual treatment times and wait times for patients receiving infusion services in a busy academic treatment center.

Quality Improvement

AB2015-25. Improving Quality Care for Cancer Patients in Spain Through the Development of Standards and Indicators

Carlos Camps, MD, PhD^ab; Eduardo Díaz-Rubio, MD, PhD^a,c; Antonio Antón, MD, PhD^a,d; Enrique Aranda, MD, PhD^a,e; Alfredo Carrato, MD, PhD^a,f; Pere Gascón, MD, PhD^a,g; Guillermo López-Vivanco, MD, PhD^a,h; Cristina Grávalos, MDⁱ; and Ana Lluch, MD, PhD^a

^aECO Foundation Board of Trustees; ^bConsorcio Hospital General Universitario; ^cHospital Clínico Universitario Clínico San Carlos; ^dHospital Universitario Miguel Servet; ^eHospital Universitario Reina Sofía; ^fHospital Universitario Ramón y Cajal; ^gHospital Clínico Universitario; ^hHospital Universitario de Cruces; ⁱHospital Universitario 12 de Octubre; and ^jHospital Clínico Universitario

Background: Care for patients with cancer is a complex and expensive process, with strong clinical and socioeconomic impacts. Health and medical organizations should define quality standards to improve structures, processes, and performance in oncology practices, helping to create a culture of self-evaluation and offering patients high-quality care. ECO Foundation is an alliance of oncologists from Spanish leading cancer centers with a clear mission of advancing excellence in patient care, research, and continuous medical education. Methods: The main objectives of this project are to reduce unwarranted clinical variability, to satisfy patient and professionals needs, and to promote a culture of continuous improvement among individuals and practices. To assess the quality of care of oncology practices in Spain, the project was developed in 4 stages. 1) Evidence and a regulated search of literature was performed to identify quality care criteria. 2) Expert review and validation was performed by a group of 9 senior and leading oncologists. 3) Professional consensus and endorsem*nt was obtained from a panel of 39 oncologists through Delphi methodology. 4) Quality indicators and standards for the selected criteria were developed. A total of 5 different areas were selected for the assessment: organization, palliative care, colorectal cancer, breast cancer, and lung cancer. Results: In a first approach, 9 oncology experts identified a total of 98 recommendations. After consensus was reached through the Delphi method, 56 quality items were selected nationwide (Table 1). Some of the principal topics were implementation of multidisciplinary committees, patient safety, diagnostic and therapeutic intervals and delays, anatomic pathology and biomarkers reports, patient information, and supportive care. Conclusions: Indicators are a reference model based on standards to measure quality care. The goal is to achieve a reference tool among Spanish oncology practices for quality self-assessment and, in a second phase, to offer a quality certification process. A public health service needs to be effective and efficient to obtain optimal health outcomes and thus ensure patients maximum objective quality.

Outcomes and Health Services Research

AB2015-26. Systematic Review and Meta-Analysis of Randomized Controlled Trials in Second-Line Treatment of Advanced Non–Small Cell Lung Cancer

Gebra Cuyun Carter, PhD, MPH^a; Annete Njue, PhD^b; Sarah Mitchell, MSc^b; Anne Heyes, MBA^b; Adrian Vickers, PhD^b; Robert Taaffe, BSc^b; Diego Novick, MD, PhD^a; Narayan Rajan, MA,MSc^a; Urpo Kiiskinen, PhD^a; Joseph Anthony Treat, MD^a; Faye W Chan-Diehl, RPh, MBA^a; Min-Hua Jen PhD^a; and Lisa M Hess, PhD^a

^aEli Lilly and Company and ^bRTI Health Solutions

Background: Non-small cell lung cancer (NSCLC) is the most common type of lung cancer, and more than half of patients present with advanced-stage disease. Despite the availability of multiple treatment options in the second-line setting, clinical outcomes remain poor. The goal of this study was to conduct a systematic literature review followed by direct and indirect comparisons of survival and toxicity data from randomized controlled trials (RCTs) investigating the use of second-line therapy in patients with advanced NSCLC. Methods: A systematic literature search was performed in MEDLINE, Embase, and the Cochrane Library for literature published between January 1, 1960, and June 26, 2014, according to a prespecified search strategy and protocol (PROSPERO 2014:CRD42014013780). Selected conferences were searched for abstracts published from January 2012 to June 2014. Bibliographic lists that included studies and recent systematic reviews were also searched. No geographic or language limitations were applied. Eligible studies included adult patients with advanced NSCLC (squamous or nonsquamous histologies) undergoing second-line treatment. Select comparisons included docetaxel, erlotinib, gemcitabine, nintedanib, pemetrexed, ramucirumab, and vinorelbine. Studies were required to include at least one outcome of overall survival (OS), progression-free survival (PFS), tumor response, toxicity, or quality of life (QOL). A Bayesian approach ensuring the preservation of randomization in the network will be used in the NMA, and density plots of the posterior samples will be compared. In addition, inconsistency and heterogeneity will be assessed, and sensitivity analysis conducted adjusting for covariates. Results: In total, 1400 citations were identified and screened for eligibility; 42 of these met eligibility criteria and were considered for the study. Thirteen trials (including the only 3 that contained gemcitabine and the only 4 that contained vinorelbine) did not connect to the evidence network. Most of the 29 trials investigated docetaxel (19 trials), pemetrexed (12 trials), or erlotinib (8 trials). There was limited evidence on ramucirumab (1 trial) and nintedanib (2 trials). OS was the most commonly reported outcome (27 trials), PFS was reported in 23 trials, and response data were reported in 18 trials. Although QOL was reported in some trials (n=12), heterogeneity between instruments and reporting methods hindered comparison. Toxicity was reported in 25 trials, but data were limited and inconsistent. Conclusions: Results presented at the NCCN Annual Conference in March 2015 reported the findings of the systematic literature review and highlights the challenges of conducting NMA's in the era of biomarkers and targeted therapies, in which there are often limited data available.

Correlative/Genomic

AB2015-27. A Clinically Validated Gene Expression Score Impacts Diagnosis and Management Recommendations of Melanocytic Lesions

Loren E. Clarke, MD; Emily Bess; Brent Evans, MS; John Kidd, MS; Kathryn A. Kolquist, MD; and Colleen Rock, PharmD, PhD

Myriad Genetic Laboratories, Inc.

Background: Many studies have documented suboptimal accuracy and reproducibility in the diagnosis of melanocytic lesions by histopathology, even by experienced dermatopathologists. Therefore, adjunctive methods that provide objective and reliable data have been sought. Recently, a 23-gene expression signature was clinically validated to differentiate benign nevi from malignant melanomas. The goal of this study was to quantify the impact of this test on diagnosis and management recommendations made by dermatopathologists. Methods: Representative sections of difficult-to-diagnose melanocytic lesions encountered during routine dermatopathology practice were submitted for gene expression testing. Samples were accompanied by a survey documenting the physician's pretest diagnosis, level of diagnostic confidence, further evaluation to be performed, and recommendations for management. After receiving the melanoma diagnostic score for the lesion, the survey was repeated. Changes between the pre- and post-test surveys were analyzed. Results: During a 10-month period, 1695 eligible cases were submitted by 79 dermatopathologists. Results revealed that 56.6% of cases initially diagnosed as indeterminate received a more definitive diagnosis of benign or malignant when the melanoma diagnostic score was available as part of the case review. Additionally, treatment recommendations were revised in 29.1% of all eligible cases. Conclusions: The melanoma diagnostic score impacts diagnosis and management recommendations by dermatopathologists confronted with diagnostically challenging melanocytic lesions. Integration of this assay into current practice has the potential to improve patient care by allowing more definitive diagnoses and management recommendations for melanocytic lesions.

Best Practices in Implementation and Use of Clinical Practice Guidelines

AB2015-28. Adherence to NCCN Survivorship Care Guidelines in Non-Small Cell Lung Cancer and Colorectal Cancer

Crystal Denlinger, MD; Kelly Filchner, MSN; Margaret O'Grady, MSN; Elaine Sein, BSN; Elyse Slater, MS; Tianyu Li, MS; Michael Slifker, MS; Susan Raysor, BA; Samuel Litwin, PHD; Hossein Borghaei, DO; and Steven J. Cohen, MD

Fox Chase Cancer Center

Background: NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for non-small cell lung cancer (NSCLC) and colorectal cancer (CRC) address survivorship care, but practical use is unknown. We assessed adherence to survivorship guidelines in an academic cancer center and associated community practice network before and after a provider-targeted educational initiative. Methods: After IRB approval at all sites, retrospective chart audits were performed to evaluate guideline adherence in 4 elements: surveillance, prevention, intervention, and care coordination. Patient charts were identified via Tumor Registry. Eligibility included patients with stage I to III CRC/NSCLC who completed curative therapy more than 6 months before audit, who had no metastatic disease at audit, and who had more than 1 surveillance visit at site between 2009 and 2013. Up to 20 charts/disease site were audited per practice. An educational initiative (lecture reviewing survivorship care guidelines, NCCN Guidelines, survivorship clinic visit note/care plan templates) was developed and disseminated to all sites after initial audit. Chart audits were performed before initiative (baseline), and at 3 and 6 months after the initiative to evaluate changes in adherence. Results: In all, 19 practices at 10 sites were audited. Baseline audit included 222 CRC/152 NSCLC; 3-month audit included 165 CRC/98 NSCLC; and the 6-month audit included 177 CRC/82 NSCLC cases. Demographic and treatment characteristics were similar across time points, although there were fewer current smokers (10% vs 20%; P=.04) and patients with stage 1 or 2 disease (90% vs 83%; P=.04) in the 3-month NSCLC audit than in the baseline. Guideline-adherent surveillance care and coordinated care was noted in 90% or more of patients in NSCLC/CRC at all time points. Cancer screening was documented in fewer than 50% of patients with CRC and NSCLC at all time points and did not improve over time. Health behavior screening documentation was similar across all time points in patients with CRC and those with NSLCL, although alcohol consumption screening increased in patients with NSCLC (P=.01 at 3 and P=.001 at 6 months vs baseline). Influenza vaccination documentation increased from 26% (baseline) to 40% (6 months) in patients with NSCLC (P=.04). Comprehensive assessment for treatment effects was not documented in most patients. Assessment of pain (P=.005/<.001), bowel dysfunction (P=.02/.01), and deconditioning (P=.03 at 6 months) declined; assessment of urinary dysfunction (P<.0001) and neuropathy increased (P=.01 at 6 months) at 3 and 6 months versus baseline in patients with CRC. Assessment of nephropathy increased at 3 and 6 months and of dyspnea at 6 months in patients with NSCLC (P<.0001). Conclusions: Patients with CRC or NSCLC receive coordinated surveillance care as per NCCN Guidelines for those diseases. However, comprehensive assessment of treatment effects, cancer screening, and health behaviors is not common. Provider-targeted education moderately improved guideline adherence. Evaluation of alternative provider initiatives to improve survivorship care is warranted.

Clinical Oncology

AB2015-29. Development of a Proposed Biosimilar Product, Pegylated Apo-Filgrastim, Based on the Demonstration of Physicochemical, Pharmacologic, and Clinical Similarity to Neulasta

Kalpna Desai, PhD; Stephen Brokx, PhD; Sanyukta Kher, MSc; Tina Catalano, MSc; Devanshi Maharaja, MTech; Louise Scrocchi, PhD; and Jason Dowd, PhD

Apotex, Inc.

This abstract is available at J Clin Oncol 2015;33 (Suppl): Abstract e13534.

Clinical Oncology

AB2015-30. Impact of Residual Disease After Chemoradiotherapy for Squamous Cell Carcinoma of the Head and Neck: A Retrospective Analysis of the Northwestern Experience

Michael D. DeVita, MD; Mitul D. Gandhi, MD; Harold J. Pelzer, DDS, MD; Bharat B. Mittal, MD; and Mark Agulnik, MD

Northwestern University, Feinberg School of Medicine, Robert H. Lurie Comprehensive Cancer Center

Background: Locally advanced squamous cell carcinoma of the head and neck (SCCHN) is commonly treated using initial chemoradiotherapy (CRT). Whether residual disease (RD) detected by CT after treatment has an impact on outcome remains uncertain. Similarly, the utility of neck dissection (ND) to resect and pathologically evaluate RD remains undefined. We performed a large, retrospective analysis to assess the frequency and impact of RD and the utility of ND in SCCHN. Methods: A query of the electronic medical record using ICD9 codes was performed to identify patients with the following inclusion criteria: age greater than 18 years, a histologically proven SCCHN diagnosed between 2003 and 2013, N1 to N3 disease, and initial treatment with CRT. Relevant clinical data were abstracted through chart review. Progression-free survival (PFS) and overall survival (OS) were estimated using Kaplan-Meier analysis, and comparisons were made with the log rank test. Chi square and receiver operating characteristics were used to assess size of RD as a predictive factor. Results: We found 123 patients fit criteria for analysis. The median age at diagnosis was 56 years. Eight percent had N1 disease, 86% had N2, and 5% had N3. After completion of CRT, 65 patients (53%) had RD. Of these, 47 patients underwent ND, and 9 (14%) were found to have residual malignancy. At a median follow up of 37 months, no difference was seen in PFS or OS between patients who had radiologic RD (P=.6; P=.65) or pathologically proven RD (P=.53; P=.65) compared with those who did not. Although RD greater than 1 cm was associated with pathological RD (P=.01), it lacked specificity (56%) and had a poor positive predictive value (14%). Size of RD was a fair predictor of residual cancer (area under the curve [AUC]=0.71). Of the 57 patients tested for p16, 74% had positive results by immunohistochemistry. Patients with p16 positive tumors appeared to have superior PFS (P=.06) and OS (P=.05), compared with those with p16 negative tumors. Conclusions: After CRT, although many patients have RD, only a small subset (14%) have residual active cancer. The presence of either does not impact PFS or OS. Moreover, using RD greater than 1 cm as a threshold to perform ND lacks specificity in identifying patients with pathologically positive residual cancer. These data suggest that ND may be safely avoided in most patients and will help inform a prospective trial using PET to identify the subset of patients most likely to benefit from ND.

Clinical Oncology

AB2015-31. Life After HIPEC; Measuring Quality of Life and Functional Status After Cytoreductive Surgery and Hyperthermic Intraperitoneal Chemotherapy

Jennifer Ford, MD, and Richard Berri, MD, FACS

St John Hospital and Medical Center

Background: Cytoreductive surgery (CRS) with hyperthermic intraperitoneal chemotherapy (HIPEC) for peritoneal carcinomatosis of gastrointestinal malignancies can improve long-term survival over conventional chemotherapy alone. However, postoperative morbidity can range between 15% and 33%. The few studies that have examined postoperative functional status report an inconsistent timeline with return to an acceptable functional status at 4 to 24 months. Patients can experience a delay in recovery for mental health, physical function, and emotional well-being. Whether patient age, peritoneal carcinomatosis index (PCI), length of hospital stay, or surgical time are significantly correlated to recovery and thus quality of life have yet to be determined. Methods: A HIPEC–specific questionnaire was developed based on the validated Functional Assessment of Cancer Therapy questionnaire. Each patient who received maximal CRS with HIPEC and met inclusion criteria was contacted by phone. Consent was obtained and a questionnaire completed. An averaged score was calculated for each section and stratified to an ECOG performance status (PS). Retrospective chart review was employed to gather demographic data. One way ANOVA analysis was used to ascertain if these characteristics showed significance relative to the patient's PS. Results: In this study, 63 patient records were evaluated for maximal CRS with HIPEC performed between Oct 2011 and July 2014. We found that 43 patients received HIPEC with maximal CRS;36 met inclusion criteria, and 33 consented. The average age was 54.9 years; 35% had colorectal and 47% appendiceal malignancy. Average scores were: physical well-being, 15.4 of 20; social well-being, 17.5 of 20; recovery, 15 of 20; mental well-being, 13.4 of 20; and functional well-being, 18.1 of 24. These correlated to an ECOG PS of 1, 0, 1, 1, and 1. Patient's age (P=.235), surgical length (P=.181), hospital duration (P=.43), complications or PCI (P=.815) showed no significance relative to postoperative PS. Conclusions: Patients can recover well from CRS and HIPEC. It is possible to return to an acceptable PS within 3 months postoperative. Age, surgical time, length of hospital stay, and PCI have no significant effects on long-term recovery.

Clinical Oncology

AB2015-32. A Retrospective Analysis of 25 Patients Diagnosed With Natural Killer T-Cell Lymphoma, Nasal Type, at a Single Academic Center

Ragisha Gopalakrishnan, MD, and Harris V. Naina, MD

UT Southwestern Medical Center

Background: Natural killer T-cell lymphoma, nasal-type (ENKT), is a rare type of non-Hodgkin lymphoma. Due to its rare prevalence, a standardized chemotherapy regimen and a prognostic model have yet to be established. Methods: In this retrospective study, we reviewed a total of 25 cases with a diagnosis of ENKT. Several variables were assessed, including international prognostic index (IPI), EGOG performance status (PS), regional lymph node involvement, bone marrow involvement, and proliferation index. A high proliferation index was defined as Ki67/MIB expression in more than 50% of neoplastic cells. We assessed clinical outcomes using various treatment modalities over the past 2 decades. Most of the patients in this study were Hispanic males (80%). The median patient age was 46 (20–72). Patients were segregated as follows: Ann Arbor stage I (n=12), stage II (n=5), stage III (n=3), and stage IV (n=5); IPI score less than 2 (n=17), IPI greater than 2 (n=8); ECOG less than 2 (n=19) and ECOG greater than 2 (n=6). Patients were treated with RCHOP plus radiation (XRT) (n=9); hyper-CVAD therapy plus XRT (n=10); or with a nonanthracycline–based therapy (VIPD + XRT, n=2; DEVIC + XRT, n=3; and SMILE + XRT, n=2). Results: The 5-year overall survival for this group was 38.4%. An IPI score greater than 2 was noted to have significant poor prognosis, with a median survival of 15 months compared with 22 months (P<.001) for IPI less than 2. Regional lymph node or bone marrow involvement at diagnosis was a poor prognostic indicator of overall survival. Patients who presented with regional lymph node or bone marrow involvement at diagnosis had a shorter median survival times than those who did not (12 vs 22 months respectively; P<.001). A high proliferation index was also associated with a poor prognosis. Patients who had a high proliferation index on diagnosis had a median survival of 15 months compared with 22 months for patients who did not (P<.001). Median survival in patients treated with RCHOP plus XRT was 12 months. For patients treated with HYPER-CVAD plus XRT, it was 24 months), and for those receiving the nonanthracycline approach, it was 28 months (P<.001). Patients treated with RCHOP plus XRT were noted to have increased incidence of disease progression compared with patients treated with HYPERCVAD plus XRT and nonanthracycline–based therapy plus XRT (P<.0001). Conclusions: This retrospective study suggests that Hyper-CVAD or a nonanthracycline–based chemotherapy with concomitant radiation are superior approaches to R-CHOP with concomitant radiation for treating ENKT. Furthermore, the results serve as a preliminary prognostic model for patients with ENKT diagnosed in the United States.

Clinical Oncology

AB2015-33. Cyclin D1 Expression in Triple-Negative Breast Cancer With New Treatment Implications

Paul Hartel, MD, and Donald Fleming, MD

Davis Medical Center

Background: Although targeted therapies are available for patients with breast cancer whose tumors express estrogen receptors (ER), progesterone receptors (PR), or show HER-2 overexpression, no such treatment exists for “triple-negative” cases. As immunohistochemistry (IHC) expression of cyclin D1 has been shown concordant with cyclin D1 gene amplification, we evaluated cyclin D1 IHC expression in 25 cases of triple-negative invasive ductal carcinoma to clarify its relationship with clinical and pathologic parameters. Methods: We reviewed 26 invasive carcinomas negative for ER, PR, and HER-2. One sarcomatoid carcinoma was omitted, leaving 25 invasive ductal carcinomas for this pilot study. Clinical information, including treatment response, was gleaned from patient records. Only one case was known to be BRCA-positive. Tumors were morphologically reviewed, and cyclin D1 immunohistochemistry was applied using BCL-1. Chi square and t-test statistical analyses were performed. Results: Patients were women who ranged in age from 32 to 88 years (m=59). Tumors were in left (n=16) and right breasts (n=9) with tumor size ranging from 0.4 to 7.2 cm (m=2.7 cm). Tumors were Nottingham grade 1 (n=1), 2 (n=10), and 3 (n=14). Eight cases had an in situ component, high grade (n=7) or low grade (n=1), ranging from 5% to 70% of tumor (m= 23%). All tumors showed cyclin D1 expression from light and focal staining (1+; n=6; Group 1) to focal intense (2+; n=9) or diffuse intense staining (3+; n=10). The latter 2 groups were combined for analyses (Group 2). Although patients were younger in Group 2, this was not statistically significant. Patients with tumors showing intense BCL-1 staining (Group 2) had larger tumors (P<.05) with more capillary/lymphatic invasion (P<.005) and lymph node metastases (P<.007). They were also less likely to respond to treatment (P=.01; Table 1). Conclusions: Cyclin D1 expression may serve as a marker for more biologically aggressive triple-negative breast cancer. These tumors may respond to targeted therapy that downregulates cyclin D1 amplification. Further research with larger sample sizes is needed.

Best Practices in Implementation and Use of Clinical Practice Guidelines

AB2015-34. Improving Level I Pathway Adherence to High-Value Chemotherapy Guidelines

Nicole Hartung, MD^a; Anna Briggs Kleckner, PhD, MPH^b; Rhonda Henschel, MBA^a; and Dean Gesme, MD^a

^aMN Oncology, Hematology; ^bAllina Health

Background: In an effort to apply the best science at the point of care, the NCCN and the US Oncology Network began collaborating in 2012 on Level I Chemotherapy Pathways, resulting in the creation of “Level I Value Pathways powered by NCCN.” These pathways are designed to address approximately 80% of patient care situations,¹ while recognizing that off-pathway regimens are acceptable in some situations. In 2013, MN Oncology and Allina Health established a comanagement agreement putting forth quality and performance improvement metrics accountable to the Oncology Clinical Service Line. One component of the agreement was an improvement of “on-pathway” chemotherapy treatment of patients with non-small cell lung cancer (NSCLC) and patients with colon cancer (CC). The objective of this research was to develop, implement, and evaluate the processes related to pathway adherence to high-value chemotherapy guidelines in patients with NSCLC and CC. Methods: Within MN Oncology, a multidisciplinary virtual pathway committee was formed consisting of local content expert physicians and supporting personnel. A company policy was created outlining guidelines for pathway adherence, including a process for off-pathway use at point of care. The off-pathway process required physicians to be notified of off-pathway ordering by either a lack of a pathway generation in the electronic medical record (EMR), or alert from the clinic patient financial counselor. The physician would then send an “off-pathway” request via email to the virtual pathway committee with details of the patient scenario, off-pathway regimen, and supporting evidence. Response with or without advice concerning the request was required from the committee within 48 business hours. To assess the evolution of on- versus off-pathway adherence, historical data were reviewed and trends in off-pathway care were noted. A chi-square test of independence with critical value of α=0.05 was performed to examine the relation between pre- and post-implementation of the process to pathway adherence. Results: Compared with baseline, physicians' adherence to the pathway improved for NSCLC from 64% to 82% and CC from 61% to 87%. Chi-square results indicate the process is likely to have improved adherence to chemotherapy pathways. (P<.001; Table 1). Conclusions: Using a physician-led virtual peer-reviewed off-pathway process, chemotherapy prescribing practices were altered to improve adherence to highvalue clinical practice guidelines.

Quality Improvement

AB2015-35. Improving Breast Cancer Care with a Rapid Diagnostic Clinic

Ruth Heisey,MD,CCFP,FCFP^a,b; John Semple, MD, FRCSC, FACS^a; David McCready, MD, MSc, FRCSC,FACS^b; Susan Hum, BSc; Bridgette Lord, APN^b; Christopher Lo, PhD^c; Tulin Cil, MD, FRCSC, MEd^a,b; and Jaime Escallon, MD, FRCSC, FACS^a

^aWomen's College Hospital, ^bPrincess Margaret Hospital, ^cToronto General Hospital

Background: Rapid diagnostic clinics (RDCs) provide imaging and biopsies (when indicated), with same or next day diagnosis. This approach may reduce women's anxiety and decrease wait times during the assessment period for a breast concern. The purpose of this study was to evaluate and compare an RDC program to usual care on women's stress and anxiety, satisfaction with care, and wait times during the diagnostic process. Methods: This was a prospective, double cohort study with 1 to 2 week pre- and postdiagnostic assessments. Women with suspicious symptoms or mammographic abnormalities suggestive of cancer were referred to Women's College Hospital's (WCH) Diagnostic Breast Clinic (usual care; UC) before an RDC was established at Princess Margaret Hospital (PMH) in April 2011. The UC group formed a “historical database” to compare with the RDC group. Prediagnostic surveys included a demographics form, and the GAD-7 and SASRQ, which assessed anxiety and stress, respectively. The latter surveys were repeated in the postdiagnostic assessment, and a modified CASC scale was performed to assess satisfaction with care. Wait time information was also collected. Results: The study included 139 women recruited in August 2009 through September 2013. Of these, 52% (18/34) of women treated through the RDC completed the pre- and postdiagnosis surveys as compared with 37% (39/105) of patients receiving UC. Half of RDC and 68% of UC diagnoses were benign. Pre- and postdiagnosis GAD-7 scores were not significantly different (NSD) in the RDC group, but UC post-GAD-7 scores were significantly decreased (P<.02). Pre- and postdiagnosis total and subscale SASRQ scores were NSD in the UC group; whereas total and subscale scores for dissociation and avoidance symptoms were significantly decreased in the RDC postdiagnosis period. Overall or subscale standardized CASC scores were NSD between cohorts. Wait times from referral to a clinic appointment, to diagnosis, or to surgical consult were 2 to 2.5 times shorter in the RDC versus The UC group; however, surgical wait time was NSD between cohorts (Table 1). Conclusions: Although clinic referral, diagnostic, and surgical consult wait times were improved in the RDC group, patients' surgical wait times were no better with RDC than with UC. However, rapid diagnosis did not adversely affect women's stress or anxiety levels. Shorter diagnostic wait times were associated with less traumatic stress from the uncertainty of a benign versus malignant diagnosis. Despite longer wait times in the UC group, satisfaction with care was essentially equal between both cohorts, thus suggesting that both models of care are meeting the needs of their patients during the diagnostic period.

Best Practices in Implementation and Use of Clinical Practice Guidelines

AB2015-36. NCCN Guidelines in Action: A Report from a Community Hospital Initiative

Karen Herold, DNP, WHCNP-BC, FNP-BC; Allyson Brooks, MD; Ahlam Jadalla, RN, PhD; and Sean Guerrero, MHA

Hoag Memorial Hospital Presbyterian

Background: According to established NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines), women at elevated risk of developing breast cancer can be identified and activated to use risk reduction strategies. Studies suggest that women's compliance with breast cancer prevention measures improves when they are informed of their elevated risk for developing breast cancer. A community hospital developed a pilot program using NCCN Guidelines to create an early risk assessment program using self-enrolled employees. We report on the lessons learned from this pilot. Methods: A retrospective descriptive analysis was conducted on data collected between April and September 2014 to assess the impact of using NCCN Guidelines on participants' practices relevant to breast cancer risk reduction. None of the participants had a history of breast cancer, and all self-enrolled into the program. Participants were sent written information about NCCN Guidelines specific to their assessed level of risk and were offered an individualized counseling session by a clinician using risk-specific recommendations based on NCCN Guidelines. Results: Of the 297 people who initially enrolled, 140 completed a 90-day post-program survey. Mean age was 45.2 years (standard deviation [SD]=6.3); 54% of participants were highly educated (bachelor's degree and higher) with a median annual income of $100,000. Of those initially enrolled, 14% were identified at elevated risk (Gail 5-year risk >1.67) compared with 11.4% of the 140 who completed a 90-day post-program survey. Remarkably, 94% of those at elevated risk opted to receive clinician counseling, compared with 46% of those at average risk. Using a 2-sample test for equality of proportions, those who received both written and clinician counseling were more likely to adopt lifestyle recommendations (Prop=0.123) to reduce their risk of breast cancer than those who received written information only (Prop=0.067); however, the difference was not statistically significant (P=.39). Conclusions: This program attracted a highly educated and fairly affluent sample. The percentage of women identified at elevated risk for developing breast cancer was close to that identified nationally, and most opted to take further steps (clinician counseling) to mitigate their risk. Given the socioeconomic status (SES) characteristics of participants, further research is needed to understand if women with an SES profile (education and income) closer to that of the national average would respond differently to being aware that they are at elevated risk.

Human Subject Research Protection

AB2015-37. NCCN's Points to Consider on the Best Practices for Biorepositories, Registries, and Databases

Jan Hewett, BS, JD^a; Diane Paul, MS, RN^b; Donna Scharff^b; Fernando Moraleda, BS^b; and Gwenn Oki, MPH^c

aYale University, bNational Comprehensive Cancer Network, and cVan Andel Research Institute

Background: For research that involves the collection and analysis of biospecimens or data, the research investigator and those involved with the specific project are required to be knowledgeable on the federal and local regulations and guidelines that govern human subject research and protected health information. In some cases, for multicenter clinical trials, multiple institutional requirements relate to the collection, storage, access, distribution, and use of these biospecimens and data. These activities require the investigator to think strategically on the design of the research project, when participants are first enrolled through long-term protection and storage of specimens and data. Because of the many considerations and issues that encompass biorepositories, registries, and databases, the NCCN Institutional Review Board Directors Forum created a taskforce to develop a best practices tool and a resource for research investigators. Methods: The NCCN Task Force for Repositories, Registries, and Databases addressed a series of objectives as described below:

• Design an electronic web-based tool and resource for research investigators who plan to design, implement, and manage a research study that intends to deposit, transfer, and/or access specimens or data in a biorepository, registry, or database;

• Provide applicable principles, strategies, regulations, basic considerations, and best practices for research investigators to reference;

• Include examples and suggestions within each section of the resource to enable the research investigator to apply these to their own research project; and

• Develop template informed consent language addressing future use of biospecimens or data.

Results: Specific results on the number of users has not yet been collected as the NCCN Points to Consider on the Best Practices for Biorepositories, Registries, and Databases as the resource was launched in late October 2014. Preliminary assessment by individuals asked to review the sections, questions, and overall design of the points to consider format have been overwhelmingly positive to date.

Conclusions: While we are in the early stages of assessing the utility of this NCCN resource tool, we do plan to conduct a user satisfaction survey within the next 6 months. Ongoing assessment of any changes in the regulations, national trends, and inclusion of relevant subject matter will continue.

Outcomes and Health Services Research

AB2015-38. Impact of Genetic Mutations on Treatment Response and Costs Related to Tyrosine Kinase Inhibitor Treatment Among Patients With Chronic Myelogenous Leukemia

Elias Jabbour MD^a; Dinara Makenbaeva, MD, MBA^b; Melissa Lingohr-Smith, PhD^c; and Jay Lin, PhD, MBA^c

^aThe University of Texas MD Anderson Cancer Center, ^bBristol-Myers Squibb, and ^cNovosys Health

Background: The clinical response rate for patients with chronic myelogenous leukemia (CML) to second-generation tyrosine kinase inhibitors (2G-TKI), dasatinib and nilotinib, after failure of imatinib therapy, may depend on the emergence of genetic mutations, among other factors. Different mutations in the BCR-ABL gene confer different degrees of sensitivity to each of the 2G-TKIs. The objectives of this study were to assess the potential treatment responses and economic consequences of limiting access to 2G-TKIs, while accounting for frequencies of genetic mutations that impact patient responses to 2G-TKI therapy. Methods: A decision analytics model was developed to examine clinical and economic outcomes among patients treated with 2G-TKIs from a US payer perspective. The model was based on a hypothetical cohort of 1000 patients with CML who are treated with dasatinib or nilotinib after treatment failure with imatinib. The frequencies of BCR-ABL mutations that influence sensitivity to dasatinib and nilotinib and the impact of the mutations on responses to 2G-TKIs were obtained from the literature. Complete hematologic response (CHR) and major cytogenetic response (MCyR) were estimated for 12 months of treatment. Annual TKI drug costs per CHR and MCyR were then estimated. Three TKI access scenarios were compared: 1) open access to both 2G-TKIs; 2) access restricted to dasatinib only (DASA-Only); and 3) access restricted to nilotinib only (NILO-Only). Univariate sensitivity analyses were additionally conducted to evaluate the impact of variations in mutation frequency and treatment response rates on model outcomes. Results: Among the cohort, the percentage of patients with CHR and MCyR were greatest in the open access (CHR, 92.6%; MCyR, 56.4%), followed by DASA-Only (88.2%, 53.4%) and NILO-Only (66.7%, 46.9%). The incidence of mutations with insensitivity to nilotinib (19.1%) was higher than that seen with dasatinib (12.4%), which largely accounted for the differences in the proportion of patients responding to therapy. Compared with the TKI costs per CHR in open access ($120,706/CHR), the costs were 5% higher ($126,753/CHR) in DASA-Only and 41% higher ($169,990/CHR) in NILO-Only. Compared with the TKI costs per MCyR in open access ($198,284/MCyR), the costs were 6% higher ($209,259/MCyR) in DASA-Only and 22% higher ($241,515/MCyR) in NILO-Only. The sensitivity analyses showed that variations of frequency and response rates of different mutation classes in the model may lead to different levels of cost per CHR or MCyR. Conclusions: Open access to both 2G-TKIs is potentially associated with greater rates of CHR and MCyR and lower costs per response compared with restricted access to 2G-TKIs.

Correlative/Genomic

AB2015-39. Setting the Standard for Comprehensive Immune Receptor Profiling in Lymphocytes

Ilan Kirsch, MD, and Harlan Robins, PHD

Adaptive Biotechnologies, Inc.

Background: Hypervariability of the primary nucleotide sequence of immune receptors formed by site-specific V(D)J (variable, diversity, joining gene segments) recombination provides essentially a unique biomarker of the lymphocyte in which the recombination events have occurred and all of its clonal progeny. Multiplex polymerase chain reaction (PCR) and high throughput sequencing of these primary nucleotide sequences could provide biomarkers for the enumeration and quantification of T- or B-cells in a sample of interest if the amplification bias caused by differential hybridization kinetics of PCR could be controlled. Methods: Forward and reverse primers recognizing each and every V segment and each and every J segment for each immune receptor locus (TCRA/D, TCRB, TCRG, IgH, IgL[kappa and lambda]) were combined in a locus-specific fashion and subjected to multiplex PCR followed by high throughput sequencing on an Illumina sequencing platform. PCR amplification bias was controlled by independently synthesizing approximately 500 bp templates for each possible V-J pair for a given locus. Through iterative redesign and recalibration of the concentration of the primers, equivalent amplification of all V-J pairings was sought. The templates also served as in-line controls and standards for each reaction. Results: A robust, quantitative, highly accurate, standardized, and highly sensitive set of assays were developed for each of the human immune receptor loci. The assays were more accurate than flow cytometry within the limit of sensitivity of that technique and also at least 100 times more sensitive in terms of the ability to identify a particular lymphocyte clone within a complex mixture of other lymphoid and nonlymphoid cells (identification of one index cell within 1 million unrelated cells). The quantitative accuracy of the assays was demonstrated by spike-in experiments involving biological samples and actual patient samples, and by measurement of the template controls themselves. The establishment of the chemically synthesized templates allows for interlaboratory standardization and establishes a set of reagents for proficiency testing. The assays have been applied clinically to distinguish malignant from benign lymphoid proliferations and to monitor the response of leukemic cells to therapeutic intervention. Conclusions: A multiplex PCR and high throughput sequencing assay has been established that allows for the enumeration and quantification of each and every lymphocyte in a biologic sample of interest. The creation of sets of chemically synthesized locus-specific V-J templates has allowed for standardization of the assay and demonstration of interlaboratory reproducibility. The suite of assays has been made available as a Clinical Laboratory Improvement Amendments (CLIA)-qualified laboratory developed test and as a research use only kit.

Outcomes and Health Service Research

AB2015-40. Feasibility of Implementation of Bioimpedance (L-Dex) Measures into a Busy Multidisciplinary Clinic: Results of a Pilot Study

Louise Koelmeyer, BApp.Sc(OT)^a,b; Katrina de Carvalho^b; Frank Vicini, MD^c; and Chirag Shah, MD^d

^aMacquarie University, ^bWestmead Hospital, ^cMichigan Healthcare Professionals/21^st Century Oncology, and ^dCleveland Clinic, Taussig Cancer Institute, Department of Radiation Oncology

Background: Breast cancer related lymphedema (BCRL) remains a growing concern for women after treatment of their breast cancer. However, standard diagnostic modalities often fail to identify the condition until clinically apparent, requiring more aggressive interventions and impacting quality of life profoundly. New modalities, including bioimpedance (L-Dex) allow for the identification of subclinical BCRL, which allows for earlier detection and treatment with smaller arm volumes. The purpose of this analysis is to present the results of a pilot program that implemented L-Dex measurements for all women diagnosed with breast cancer before and after treatment in a multidisciplinary breast clinic. Methods: Between May 2008 and December 2010, a program evaluating the use of single-frequency bioimpedance measurements for BCRL assessment was undertaken. Women were measured before surgery and then postoperatively at 10 days, 6 months, and 12 months as part of their care at the breast clinic. Results: A total of 257 women underwent presurgical measurements in the breast clinic (L-Dex range, −8.9 to 13.4). Of these 84% (n=217) had a postoperative measurement, 62% (n=160) a 6-month measurement, and 44% (n=112) a 12-month measurement. After surgery, increases in L-Dex values were noted postoperatively (L-Dex range, −8.2 to 29.9) and at 6 (L-Dex range, −8.4 to 53.5) and 12 months (L-Dex range, −7.8 to 22.4) months. Thirty-nine women (15%) were referred for BCRL evaluation due to an L-Dex score increase greater than 10 from baseline or due to symptoms, with 24 receiving compression garments, 7 compression bandaging, and 3 complex decongestive physiotherapy. Patients referred because of elevated L-Dex scores were likely to have undergone more locoregional and systemic therapy; with 88% undergoing axillary lymph node dissection, 68% radiotherapy, and 59% taxane-based chemotherapy. Bioimpedance has been incorporated into daily practice for all women with limited increase in time requirements (5 minutes). The general profile of the clinic has increased and more lymphedema inquiries and early referrals to the clinic occur. Patient satisfaction has improved with regards to BCRL education. Conclusions: The results of this pilot program support the routine use of bioimpedance before and after treatment for their breast cancer. The use of bioimpedance was associated with a limited increase in visit duration; improved patient satisfaction and increased clinic referrals were seen, and women identified with increases in L-Dex scores postoperatively were sent for early intervention to the lymphedema clinic.

Best Practices in Implementation and Use of Clinical Practice Guidelines

AB2015-41. Compliance with NCCN Recommendations for Management of Axillary Lymph Nodes and Postmastectomy Radiation in the Neoadjuvant Setting at a Single Comprehensive Breast Cancer Center

Kavya Krishna, MD; Julie Stephens, MS; and Ewa Mrozek, MD

The Ohio State University Wexner Medical Center, The James Cancer Hospital

Background: Although the goal of neoadjuvant chemotherapy (NCT) is improvement in the rate of breast-conserving surgery (BCS), significant number of patients will have mastectomy after NCT. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for breast cancer (BC) recommend that postmastectomy radiation (PMRT) should be based on the worst pre- or posttreatment stage of BC and should be strongly considered for patients with any number of axillary lymph nodes (ALNs) involved. Since NCT can eradicate BC from ALNs, assessment of ALNs before NCT is essential for considering need for PMRT. NCCN Guidelines recommend that patients with clinically negative (CN) ALN should undergo axillary ultrasound followed by sampling of suspicious ALNs. Sentinel lymph node excision (SLNE) can be considered before NCT. Clinically positive (CP) ALN should be biopsied. In this single institution study, we investigate compliance with NCCN Guidelines recommendations for ALN assessment and PMRT. Methods: Retrospective review of 209 patients with stage II or III BC treated with NCT, surgery with or without radiation therapy from December 2009 to July 2013. Data on patient and BC characteristics, ALN assessment with imaging and biopsy, rate of SLNE, type of breast surgery (BS), and PMRT were collected. Results: The median age of patients receiving NCT was 50 years (range, 24–78). We found that 68 patients had CN ALNs but imaging assessment of ALN was performed in only 79% of them. In patients with CN ALN, only 9% had SLNE before NCT, 80% had SLNE after NCT and 11% had ALN dissection (ALND) during BS. Among 137 patients with CP ALN, 99% had imaging assessment of the axilla followed by ALN biopsy in 82%. Most (85%) patients with CP ALN had ALND during BS, the remaining 15% underwent SLNE. In this study, 161 (77%) patients treated with NCT underwent mastectomy; 140 patients met NCCN Guidelines criteria for PMRT and 84% of them received PMRT. Conclusions: The major deviation from NCCN Guidelines recommendations occurred in patients with CN ALN when imaging assessment of the ALN was performed in only 79% of patients; SLNE was rarely performed before NCT. In most patients, recommendation and referral for PMRT was based on the worst pre- or post-treatment stage of BC.

Clinical Oncology

AB2015-42. The Diagnostic Yield and Safety of Endobronchial Ultrasound Transbronchial Lung Biopsy with Guide-Sheath for Peripheral Lung Cancer

YuKun Kuang, MD; FengJia Chen, MD; and ZhiWen Zhu, MD, PhD

The First Affiliated Hospital of Sun Yat-Sen University

Background: Appropriate early diagnosis is one of the keys to lowering mortality from lung cancer, but traditional transbronchial lung biopsy offers lower diagnostic yield. Conversely, endobronchial ultrasound transbronchial lung biopsy with guide-sheath (EBUS-GS-TBLB) provides a higher diagnostic yield in peripheral lung cancer. In this study, we evaluated diagnostic yield and safety of EBUS-GS-TBLB in peripheral lung cancer (PLC). Methods: Between September 2012 and May 2014, EBUS-GS-TBLB was performed in the First Affiliated Hospital of Sun Yat-Sen University in patients with PLC suggested by chest computed tomography but whose tumors were inaccessible via conventional bronchoscopy. The diagnostic yield, safety, and associated factors were analyzed. Results: This study involves 35 patients (18 men and 17 women; mean age, 59±14) with a diagnosis of lung cancer confirmed by pathology and/or cytology. Of these, 29 cases were diagnosed as adenocarcinoma;2 were diagnosed as squamous carcinoma; and 4 were diagnosed as poorly differentiated carcinoma/anaplastic carcinoma. The mean procedure time for EBUS-GS-TBLB was 16.7 (±4.3) minutes. The average number of biopsy specimens obtained in each lesions was 3.8 (±1.5), and the mean diameter was 0.33 (±0.18) cm. A total of 36 lesions, with a mean diameter of 32.6 (±12.6) mm, were checked in 35 patients. All lesions could be displayed on EBUS images. Of the lesions studied, 25 were definitively diagnosed using EBUS-GS-TBLB, for a diagnostic yield of 79.4%. The diagnostic rate of pathology was 66.7% (24 of 36), and that of cytology was 50.5% (18 of 36). Some factors improved the diagnostic rate, such as lesion size greater than 25 mm in diameter, obtaining more than 5 biopsy specimens, and properly positioning the probe (Table 1). All patients tolerated the EBUS-GS-TBLB procedure well; no cases of pneumothorax, hemoptysis, or other serious complication occurred. Conclusions: EBUS-GS-TBLB demonstrated a high diagnostic yield and was minimally invasive with few complications. It was shown to be an effective and safe procedure for diagnosing peripheral lung cancer. Carefully choosing suitable cases may improve the diagnostic rate.

Clinical Oncology

AB2015-43. Organ-Preserving Surgery Using Endografts From Superelastic Titanium-Nickelid-Based Alloy Tissue-Tissues for Patients With Laryngeal Cancer

Denis E. Kulbakin, MD,^a,b; Marat R. Muhamedov, MD^a,c; Evgeny L. Choinzonov, MD^a,c; Alexandr A. Zheravin, MD^a; and Viktor E. Gunther, PHD^d

^aCancer Research Institute of RAMS, ^bTomsk State University, ^cSiberian State Medical University, and ^dResearch Institute of Medical Materials and Implants with Shape Memory

Background: The need for reconstruction of lost structures of the larynx as a result of extensive resection of the larynx for locally advanced carcinoma of the larynx is still current. The success of these surgeries depends on the choice of methods and material used for reconstruction. Materials: In this study, 18 patients with morphologically verified cancer of a larynx, T₂ to T₃ stages, were included. All patients were subjected to different types of partial laryngectomy (from the anterolateral to the subtotal.) In all cases, for the purpose of reconstruction, endografts from superelastic titanium-nickelid-based alloy were used. In 12 cases, these reconstructions were performed on patients with recurrence of laryngeal cancer after of radiotherapy (rT_2, 4; rT₃, 8). The formation of the inner lining of the reconstructed larynx was performed by means of the displaced flap of subcutaneous muscle of the neck (in 9 cases) and preliminary prefabrication of the flap with the inclusion of the endografts from superelastic titanium-nickelid-based alloy. The stage of flap prefabrication in 5 cases was performed in the rectus abdominis muscle, followed by microsurgical reconstruction. In 4 cases, prefabrication area was located in the subclavial area. Results: The internal surface of the larynx was consistent in all cases. In all cases (100%) the correct position of an implant, without shift, was noted. Average time of decannulation was 30 days after surgery. In 16 patients (89%) sufficient lumen of the larynx was recovered. Individual feeding was fully restored on day 5 to 7 after surgery. Voice function was preserved in 15 patients (83%). In 2 (11%) of 18 patient, local recurrence occurred. These patients were successfully treated with total laryngectomy. Conclusions: Endografts from superelastic titanium-nickelid-based alloy were used to adequately recover the frame form of the larynx. Using a prefabrication flap is preferred for reconstruction of extensive defects of the larynx during organ-preserving surgical treatment in patients with locally advanced cancer of the larynx, and especially in cases of recurrence after radiotherapy.

Outcomes and Health Services Research

AB2015-44. Cancer Cachexia and Survival in Veterans with Lung Cancer by Histology: Who Is Most at Risk?

Sheetal Malhotra, MD^a,b; Kan Huang, MD, PhD^c; Ravi Vinnakota, MD^a; Kosana Simeunovic, MD^a; Tatjana Gavrancic, MD^a; Nicolas Villanueva, MD^c; Donald Dietz, MD^c; Kathleen Fenn, MD^c; Anup Biswas, PhD^d; Kristen Beck^e; Courtney co*ker^d, Yeun-Hee A Park, MD^a; and Swarnali Acharyya, PhD^d

^aJames J. Peters Bronx Veterans Affairs Hospital, ^bNorth Central Bronx Hospital, ^cColumbia University Medical Center, ^dInstitute For Cancer Genetics at Columbia University, and ^eCollege of Physician and Surgeons at Columbia University.

Background: Cancer cachexia is a syndrome characterized by wasting of lean body mass and fat, resulting in decreased treatment tolerance and quality of life. Approximately 50% to 60% of patients with lung cancer develop cancer cachexia. Knowledge of cancer cachexia trends and outcomes in different types of lung cancers can help target patients most at risk of decreased survival. In our study, we examined weight loss trends in patients with lung cancer and survival according to cancer histology. Methods: We conducted a chart review of veterans diagnosed with lung cancer from 2007 to 2013 at a Veterans Affairs Medical Center in a large metropolitan area. Data collected included demographics, date of diagnosis, treatment received, type of cancer by histology, presence of metastasis, preexisting conditions, and other medications used. Body mass index (BMI) and weight before, during, and after treatments were recorded, and percentage weight loss and survival days were calculated. Patients were also categorized according to the extent of weight loss (<5%, 5%–10%, >10%) through their disease course. Descriptive statistics, ANOVA, T-tests, and regression analysis were used for data analysis to assess differences and relationships between percentage weight loss and survival for all groups. Results: Data from 197 patients were available for the study. All patients were men with an average age of 68.6 years (SD±8.84 years); 55% of the patients were white, 35% were African-American. Of the total patients, 47.7% had adenocarcinoma, 8.1% had small cell lung cancer (SCLC), and 44.2% had squamous cell cancer (SCC). Percentage weight loss trends for patients with adenocarcinoma showed a 9% loss from baseline till time of diagnosis and a total of 15% loss after treatments; patients with SCLC similarly showed a decline from 13% to 22%; and patients with SCC, 14% to 17% weight loss. Of all patients with > 10% weight loss over time, 48% had adenocarcinoma, compared with 10% with SCLC and 42% with SCC. Average survival for patients with adenocarcinoma was 420 days; for patients with SCLC, 321 days; and 492 days for patients with SCC (P=.07). Weight loss percentage was significantly related to survival in all patients (r, −0.19; P<.01) and patients with adenocarcinoma (r, −0.25; P<.01) but not in patients with SCC or with SCLC. Conclusions: Cancer cachexia trends in veterans with lung cancer are similar for patients with adenocarcinoma, SCLC, and SCC; however, percentage weight loss may adversely impact survivalespecially in patients with adenocarcinoma.

Best Practices in Implementation and Use of Clinical Practice Guidelines

AB2015-45. Interdisciplinary Management of Head and Neck Cancer: Putting NCCN Guidelines into Practice to Improve Patient Quality of Life

Joel D Marcus, PsyD; Brian Moore, MD; and Christian Hasney, MD

Ochsner Cancer Institute

Background: Well-documented studies indicate that one third of patients with cancer have distress associated with their diagnosis.¹ There is a high level of pain and psychosocial or emotional distress, and few patients with head and neck cancers are referred to psychosocial care.^2–4 Thus, among patients with head and neck cancers, there is an overwhelming unmet psychological need. ⁵ Oncologists underestimate patients' distress and need for support. This underestimation was most apparent in patients with head and neck and lung cancers.⁶ Methods: Our policy was developed in concordance with NCCN standards. Each new patient seen in the Head & Neck Surgical Oncology clinic is given the NCCN Distress Management Assessment tool. Those data are entered into the electronic medical record (EMR). In our program, we have established 7 or above on the Distress Thermometer (DT) as generating a best practice advisory, asking the physician to make a referral to the psychosocial oncologist or social worker. Results: Use of the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for distress signified that referrals to our psychosocial oncologist and social worker were indicated. This generated a Best Practice Advisory in our EMR. We obtained data on 477 distinct patients with a head and neck cancer diagnosis (dx 140-149.99). Conclusions: Using best clinical practices initiated by the use of the NCCN Guidelines for distress can be useful in the triage of patients with head and neck cancers. We discuss the benefits and limitations of starting distress screening in a multidisciplinary outpatient surgical head and neck clinic. We describe treatment of in concordance with published data.

Clinical Oncology

AB2015-46. PhaseI/II Study With Fotemustine in Recurrent Unmethylated MGMT Glioblastoma

Alfredo Marinelli, MD^a,b; Giuseppe Lamberti, MD^a; Carlo Buonerba, MD^a; Luigi Cerbone, MD^a; Nadia Cordua^a; and Gianfranco Peluso, MD^c

^aUniversity of Naples Federico II ^bIRCCS Istituto Neurologico Mediterraneo Neuromed, and ^cInstitute of Biomedicine and Bioresources

Background: Fotemustine (FM) is a third-generation nitrosourea used in Europe for the treatment of recurrent glioblastoma (rGBM). According to a registration schedule (1987), FM is given intravenously at 100 mg/m² weekly for 3 weeks, followed by a 5-week pause and then at 100 mg/m² every 3 weeks. This schedule causes high adverse events (AE) rates. Our goal was to prove the safety and efficacy of a higher dose intensity biweekly FM schedule. Methods: This is a phase I-II trial on FM in patients affected by histology-determined rGBM with unmethylated MGMT promoter. In Phase I, FM is administered at a starting dose level (DL) of 120 mg/m² biweekly for at least 5 cycles.¹ FM dose is increased at 140 mg/m², 150 mg/m² and 155 mg/m² after a standard phase I 3 plus 3 design. The goal of Phase I is to determine the maximum tolerated dose (MTD), the dose to be used in phase II (P2D), and to assess toxicity. In phase II, FM is given with the same schedule at P2D dose level. MRI scans with spectroscopy are performed every 6 to 8 weeks and evaluated according to Macdonald criteria. The primary endpoint in phase II is progression-free survival rate at 24 weeks (PFS-24). Secondary endpoints are overall survival from diagnosis (OS-D) and from start of FM (OS-F), disease rate control, toxicity assessment, and schedule feasibility assessment by using the real dose density (DD). Results: In this study, 37 consecutive patients with rGBM were enrolled. All of the patients received Stupp protocol as first-line treatment. Phase I, included 9 patients; 3 were treated at the first DL (120 mg/m²). No AE higher than grade 2 were seen in this cohort. Three patients were enrolled at the second DL (140 mg/m²). In this cohort, 1 grade 4 thrombocytopenia was noted. Three more patients were enrolled at the same DL, with 3 grade 4 AE (thrombocytopenia) noted. Then 140 mg/m² was established to be the MTD and 120 mg/m² the P2D. Phase II included 31 patients, including the 3 treated at 120 mg/m² in phase I. The PFS-24 was 19.4% (6/31). Median OS-D and median OS-F were 13.97 months and 4.37 months, respectively. According to Macdonald criteria, 15 of 31 patients experienced stable disease at the end of the schedule (disease control rate, 48.37%). We noted that 3 of 31 patients had grade 3 to 4 hematologic toxicity (9.7%): 1 grade 3 and 1 grade 4 thrombocytopenia; 1 grade 4 anemia. The grade 1 to 2 hematologic toxicity rate was 29% (9/31). The real DD was 56.56 mg/m² weekly (of a maximum of 60 mg/m² weekly as expected with the experimental schedule) against 37.5 mg/m² weekly maximum expected for the registration schedule. Conclusions: FM biweekly administration schedule allows us to deliver higher doses of FM in the same period of time as registration schedule, resulting in a significantly higher DD, in a safe way. Moreover, our study highlights the need to explore and rearrange old chemotherapy schedules to enhance DD.

Quality Improvement

AB2015-47. An Exploration of Why Oncology Patients Do Not Adhere to Oral Chemotherapy Regimes

Kelly A. McCue, DNP, MSN, AOCNS, RN, CHPN

Northeastern University

Background: Over the past 20 years, oral chemotherapeutic agents have become a viable option in cancer treatment, offering vulnerable oncology patients fewer and less-severe adverse effects than intravenous chemotherapy and better quality lives. However, this oral chemotherapeutic agent delivery trend has begun to show adherence barriers for patients who are now enabled to self-medicate oral chemotherapy, and it poses challenges for oncology nurses. When oral chemotherapy is administered outside of a controlled clinical setting, barriers to adherence can be tenfold as compared with chemotherapy in an observed clinical setting. It can involve patient and nurse behaviors that are concordant or discordant with those associated with therapeutic cancer treatment, significantly contributing to the success or failure of the treatment. Methods: The purpose of this study was to determine factors that present barriers and challenges to oral chemotherapy adherence with a goal improving quality of care. The purposive sampling allowed recruitment of 2 groups in an oncology department in a rural community hospital: all 4 staff nurses and a convenience sample of 12 existing and new oncology patients who met the standard eligibility criteria for oral chemotherapy self-administration. A multiperspective qualitative study was completed using investigator-developed surveys specific to each group with followup taped audio discussions. Results: Predominant themes emerged from 2 perspectives into provision of patient care related to oral chemotherapeutic regimen adherence, providing useful insights. Patients reported that they want to be informed about adverse effects to allay fears that they might have. They reported such barriers to oral chemotherapy adherence as adverse effects, drug distributor or manufacture issues, forgetting to take drug(s), and insurance issues. Barriers to oral chemotherapy adherence reported by nurses were cost, regimen confusion, drug-distributor delay, forgetting to take drug(s), nurse role confusion, documentation and provider communication problems, oncology nurse competency, adverse effects, and fear. Nurses professed a commitment to oral chemotherapy barrier assessment and provision of interventions to overcome such barriers. Conclusions: Results reveal that barriers to adherence can be paramount in preventing patients from attaining positive therapeutic outcomes. A better understanding of oral chemotherapy nonadherence and the barriers and challenges inherent in the process help guide patients to success and ultimately improve quality of life. Future plans include the development of assessment and education tools designed for use by nurses to assist patients to navigate through barriers, and contribute to improved, safe oral chemotherapy adherence and treatment outcomes.

Preclinical Oncology

AB2015-48. Stabilization of Cyclin D1 With Pyrrolidine Dithiocarbamate Leads to MDA-MB 231 Cell Death

Inderjit Mehmi, MD;* and Travis Salibury, PhD

ECCC/JCESOM, Marshall University

Background: Triple-negative breast cancer (TNBC) a heterogeneous group of breast cancers defined by lack of estrogen receptor (ER), progesterone receptor (PR), or HER-2/neu receptors, represents about 20% of breast cancer. It tends to occur in younger women and carries a less favorable prognosis (local recurrence, disease-free survival) when compared with ER-, PR-, or HER-2/neu receptor–positive breast cancers. There is an active effort to find more effective and less toxic therapeutic options for TNBC. Pyrrolidine dithiocarbamate (PDTC) is a NFkappa B inhibitor that leads to down regulation of RelA (a subunit of NFkappa B). This study reports PDTC-mediated cell death of MDA-MB 231 (a representative of TNBC cell lines) via a previously unknown mechanism. Methods: For cytotoxicity assay, We plated 50,000 cells/well in 6-well plates using fetal bovine supplemented DMEM media. After 24 hours, the cells were treated with dimethylsulfoxide (DMSO; control), and PDTC at concentrations of 100, 10, 1, and 0.1 μM. After 24 hours of treatment, the cells were collected and counted. For western blot analysis, to analyze the mechanism of injury, we plated 150,000 cells/well in a 6-well plate and treated with PDTC 100 μM for 24 hours. These cells were then lysed in sample loading buffer; 12.5 μL of this was used to analyze with western blotting technique for cyclin D1 and phosphorylated retinoblastoma (pRb). Also the cell lysate was fractionated into whole cell, cytosolic, and nuclear fraction and analyzed for cyclin D1. Results: In cytotoxicity assay, after treating the MDA-MB 231 cells with PDTC for 24 hours, we counted the cells. Our data indicates that PDTC at 10 and 100 μM lead to a significant cytotoxicity with less than 50% of the MDA-MB 231 viable cells remaining. With 1 μM PDTC, about 75% of the cells were viable after 24 hours treatment. No significant difference was seen in 0.1 μM and DMSO treated cells. In western blot analysis for cyclin D1, the data showed a striking increase in levels of cyclin D1 after treatment with PDTC, compared with DMSO controls. Furthermore, when these treated cells were fractionated into whole cell, cytosolic, and nuclear fractions, we found that most of the cyclin D1 was localized to the nucleus. Rb, a target of cyclin D1 for phosphorylation that allows the cell cycle to proceed once phosphorylated, was also analyzed using western blotting technique. We found PDTC-treated cells had decreased phosphorylated Rb compared with DMSO (control) treated cells. Conclusions: Our data clearly show that PDTC treatment leads to cells death. After 24 hours, only half of the plated cells remained viable when treated with PDTC. We also were able to see a dose response toward this cytotoxicity. We examined this cytotoxicity starting at the cell cycle proteins level. To our surprise, cyclin D1 was increased to very high levels in PDTC–treated cells; however, this increase in levels of cyclin D1 appears not to hold its normal cellular function. Our data indicate that PDTC–treated cells had decreased pRb. Under normal cellular conditions, cyclin D1 leads to phosphorylation of Rb and continuation of cell cycle to the next phase. However, according to our data, increased cyclin D1 is unable to carry out phosphorylation of Rb, shown by lower levels of pRb in PDTC–treated cells. Furthermore, we localized cyclin D1 largely to the nucleus. When we prepared whole cell, cytoplasmic, and nuclear fraction of PDTC–treated cells, most of the cyclin D1 was present in the nucleus, when compared with DMSO–treated cells. It appears that PDTC treatment of MDA-MB 231 cells leads to stabilization of cyclin D1. However, this stabilization does not lead to intact function of cyclin D1 to help MDA-MB231 cells progress through the cell cycle, but rather cell death. Recent reports have linked cyclin D1 activity and autophagy. Decreased activity of cyclin D1 has especially been linked to autophagy.¹ Data also exist that link Rb-E2F interaction leading to autophagy.^2–4 Our data indicate that after PDTC treatment of MDA-MB 231 treated cells, cyclin D1, though increased in levels, can't perform its normal function of phosphorylating Rb. We propose that lack of activity of cyclin D1 is leading to autophagy and cell death.

Outcomes and Health Services Research

AB2015-49. Costs and Health Care Utilization of a Large Insured Population With Stage IV Metastatic Breast Cancer by Receipt of HER2-Targeted Agents

Nicole Meyer, MA^a; Yanni Hao, PhD^b; Paul Juneau, PhD^a; Nianwen Shi, PhD^a; Xue Song, PhD^a; and Jaqueline Willemann Rogerio, MD^b

^aTruven Health Analytics and ^bNovartis Pharmaceuticals

Background: Treatment costs of HER2 postivie metastatic breast cancer (mBC) are substantial. Little is known on treatment cost variations and health care utilization (HCU) based on HER2 positive status. This retrospective study evaluated HCU and costs of treating mBC by receipt of HER2-targeted agents (H2T). Methods: Women 18 years or older, diagnosed with mBC were selected from the 2008 to 2012 MarketScan Commercial and Medicare databases; date of first metastasis was the index date. Patients were followed up until the end of enrollment or inpatient death. Patients with less than 12 months' enrollment, non-BC cancers before index, HIV, or pregnancy were excluded. Receipt of H2T was defined as 1 or more claim for trastuzumab, lapatinib, or pertuzumab in the study period. Study cohorts were women treated with H2T, women treated with antineoplastic agents but not treated with H2T (non-H2T), and patients with no treatment. Per patient per month (PPPM) services and costs were calculated for all-cause (ALL) and BC-related (BCR) services. Costs were inflation-adjusted to 2012 US dollars and represent total payments to all service providers. Results: Of 18,059 eligible women, 14.6% received H2T, 71.1% had non-H2T treatment, and 14.3% had no treatment; median length of followup ranged from 325 to 549 days. H2T patients were younger than non-H2T and no treatment patients (H2T, 56; non-H2T, 60; no treatment, 62 years). No treatment patients had significantly more ALL and BCR inpatient admissions, total inpatient days, and ALL emergency room visits than H2T and non-H2T patients. H2T users had greater ALL and BCR office visits, laboratory and diagnostic radiology, radiation treatments, other outpatient services, and prescription antineoplastic use than non-H2T and no treatment patients (all P<.001). ALL prescription fills were highest for non-H2T users. After adjusting for patient characteristics, ALL and BCR total costs were both significantly higher for H2T than non-H2T users and no treatment patients (ALL mean [standard error; SE] was $11,107 [373] for H2T; $8,458 [98]for non-H2T; and $8,284 [183] no treatment) and (BCR mean [SE] was $6,215 [245] H2T; $3,908 [47] non-H2T; $2,858 [74] no treatment). Conclusions: Among patients with mBC, receipt of H2T was associated with greater levels of ALL and BCR HCU and total costs; untreated patients had the highest ALL emergency room use and ALL and BCR inpatient costs.

Outcomes and Health Services Research

AB2015-50. Healthcare Disparities According to Insurance Status Among Patients With Salivary Gland Tumors

Sobia Nabeel, MD; Bilal Ahmad, MD, Haseeb Saeed, MD; Michael Machiorlatti, MS5; Sara K. Vesely, PhD; Mohammed Muqeet Adnan, MD; Mohamad Khawandanah, MD; Wajeeha Razaq, MD; and Mohammad Razaq, MD

Oklahoma University Health Science Center OKC

Background: Insurance status appears to be an independent factor in health care utilization in common cancers according to some studies. No data are available in this regard for rare tumors such as salivary gland tumors. The University of Oklahoma Health Sciences Center (OUHSC) is the main tertiary hospital in Oklahoma, treating patients with rare tumors like salivary gland tumors. These tumors comprise a group of heterogeneous cancers that encompass than 1% of all cancers. As health care reforms and the Affordable Care Act (ACA) proceeds in the United States, aiming to cover a large number of the uninsured population, the trends and outcomes in the setting of cancer, especially in rare tumors, is important to know. We describe a retrospective analysis of insurance status and distance to travel for treatment of patients with malignant salivary gland tumors at our institution, evaluating the impact on clinical outcomes. Methods: From June 1976 to October 2013, we identified a total of 276 patients who met the inclusion criteria for the study. Of these patients, 117 were younger than 60 years and 159 were 60 years or older. Median age at the time of diagnosis was 60 years, with 170 (62%) men and 106 (39%) women. Patients were divided into 2 groups on the basis of insurance status: Group 1 included patients with Medicare, Medicaid, private and military insurance (n=184; 67%); Group 2 included uninsured patients (n=92; 33%). Distance to travel for treatment was assessed in 276 patients using a continuous variable, dividing into 4 groups: 0 to 40 miles (82; 30%); 41 to 75 miles (107; 39%); more than 75 miles (84; 31%); and 3 patients whose residential address was not available. These patients were excluded. A chi-square test of independence was performed on insurance status and distance groups. A Kaplan-Meier method was used to estimate survival rates for categoric variables. A Cox proportional hazards model on overall survival (OS) was implemented to assess continuous covariates as well as create a final multivariable model. Interaction for age, distance, insurance, and gender was assessed with all covariates. All statistical analysis was performed using SAS 9.2 software (SAS Institute Inc). Results: No association was found between insurance status and distance groupings (log rank, P=.8740). We found no statistical difference in the survival rates (log-rank, P =.5905) in the overall survival (years) between the insured and uninsured groups. No statistical difference was found among groups with varying distance to travel for treatment (log-rank, P =.4497). In the presence of other variables and subgroup analysis, using the Cox proportional hazards model, the variables that were influential and statistically significant included age 60 years or older (hazard ratio [HR], 1.86; 95% CI, 1.25–2.80) and female gender (HR, 0.58; 95% CI, 0.40–0.86). Of the 276 total reviewed cases, about 272 (99%) were from Oklahoma and 4 (1%) were from neighboring states. Conclusions: Based on our retrospective analysis, insurance status did not impact survival outcomes. Similarly, distance travelled to receive treatment did not alter survival outcomes. However, the survival analysis showed a statistically significant OS benefit in patients with age less than 60 years and female gender. Further studies are warranted for a better understanding of health care disparities and subsequently proper implementation of the ACA for rare tumors like salivary gland tumors.

Outcomes and Health Services Research

AB2015-51. United States Treatment Patterns After Ipilimumab Use in Patients With Advanced Melanoma

Tony Okoro, PharmD^a; Elisabetta Malangone-Monaco,MS^b; Beata Korytowsky, MA^a; Amy Stanford, PharmD^a; Katy Pan, BS^b; William Johnson, BS^b; and Sigrun Hallmeyer, MD^c

^aBristol-Myers Squibb, ^bTruven Health, and ^cAdvocate Lutheran General Hospital

Background: Melanoma is the fifth leading cancer among adults in the United States, resulting in an estimated 9,710 deaths in 2014. Historical survival rates for advanced melanoma have been poor, but with the approval of the first immunotherapy, recent ipilimumab (IPI) data have shown survival rates of 46% and 24% at 1 and 2 years, respectively. Although newer treatment options are available, including targeted therapies and immunotherapies, for first-line treatment of advanced melanoma, there is currently no standard of care after progression with first-line therapy. The objective of this study was to examine real world treatment patterns after IPI use in patients diagnosed with advanced melanoma using a claims dataset. Methods: Adult patients with diagnosed stage III or stage IV melanoma treated with IPI were selected between April 1, 2011, and September 30, 2013 from the commercial and Medicare Supplemental MarketScan claims databases. Patients were evaluated for treatment with chemotherapy, targeted therapy, or immunotherapy after IPI (defined as index date). Baseline characteristics and treatment patterns post-index date were examined. Results: Among the 361 eligible patients, most had stage IV (74%) disease and were men (62%). Before the start of IPI, approximately 30% of patients received a treatment recommended per 2013 NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines), including non-IPI immunotherapy in 13%, chemotherapy in 12%, and targeted therapy in 11% of patients. After IPI use, 31% of patients received an additional line of therapy, with an average duration of 5.6 months for the first subsequent line of therapy post-IPI. The most common treatments after IPI use were vemurafenib (32.4%), pacl*taxel (28.8%), and temozolomide (20.7%). Only 4 patients received further treatment with a second subsequent line of therapy post-IPI. Conclusions: The treatment of advanced melanoma was considerably altered with the approval of IPI, which has become a mainstream first-line therapy. Although recent approvals in advanced melanoma have provided additional benefit, many physicians revert back to treating patients with traditional chemotherapies. These data highlight the limited use and lack of a standard of care in the post-IPI setting and demonstrate a degree of unmet need that continues to exist for patients who exhaust treatment options. These data should be revisited when longer term data are available to provide further evidence of the real world treatment use after IPI.

Quality Improvement

AB2015-52. Decreasing Diagnosis to Treatment Times through a Multidisciplinary Lung Cancer Program

Kimberly Byrwa-Neff, RN, BA, CPHQ, and Rhondalyn S. Prince RN, BSN, MBA

Oncology Hematology Care (OHC) Inc.

Background: Lung cancer is the leading cause of cancer death among both men and women. Each year, more people die of lung cancer than of colon, breast, and prostate cancers combined. The evaluation and treatment of lung cancer is a complicated process and the possibility of an extended interval from diagnosis to treatment adds to patient anxiety and may lead to gaps in care. Importantly, the time from diagnosis to treatment may be a reflection of the availability of resources at a hospital or physician office or the efficiency of the overall system of care. Methods: Baseline data were collected by the hospital cancer registrar and then reported quarterly. On review, it was noted that in most instances, more than 50% of newly diagnosed patients with lung cancer were waiting longer than 30 days from diagnosis by pathology to the beginning of treatment. Results: Before the development of the multidisciplinary lung cancer program, data showed the percentage of patients with 30 calendar days or fewer from the pathologic diagnosis of lung cancer to treatment time were:

• Quarter 1 2013, 51%

• Quarter 2 2013, 45%

• Quarter 3 2013, 25%

The multidisciplinary lung cancer team began meeting in November 2013. One of its primary objectives was to reduce the number of calendar days from the pathologic diagnosis of lung cancer to treatment time. Treatment was defined as whatever therapeutic intervention was appropriate for the patient's specific diagnosis, but in general meant initiation of chemotherapy, radiation therapy, or surgery. Followup data for the percentage of patients with 30 days or fewer from pathologic diagnosis to treatment time were:

• Quarter 4 2013, 86%

• Quarter 1 2014, 86%

• Quarter 2 2014, 100%

• Quarter 3 2014, 100%

Conclusions: With the development of a multidisciplinary lung cancer program, time from a new diagnosis by pathology of lung cancer to treatment time was improved by 35% in 1 year. The multidisciplinary program components included 1) establishment of a multidisciplinary lung cancer team including oncologists, pulmonologists, radiologists, pathologists, and surgeons; 2) development of a data collection tool and review of patient data and root cause analysis as to why there were delays from diagnosis to treatment; 3) deployment of a lung cancer screening program; 4) development and deployment of a flow sheet to standardize the care process; 5) employment of nurse navigators; 6) education of primary care physicians and specialty physicians on current guidelines, patient risk factors, and processes to standardize care; and 7) provision of patient education.

Epidemiology

AB2015-53. Are We Ageist? Outcomes of Major Colonic Resections in Septuagenarian and Octogenarians Versus Younger Patients

Ilayaraja Rajendran, MBBS, Hrishikesh Shenoy, and Palanichamy Chandran

Wrexham Maelor Hospital

Background: We compared postoperative outcomes after major colonic surgery in patients younger and older than 70 years of age. We assessed postoperative inpatient stay and major complications in the 2 groups. In contrast to the burgeoning elderly population in the United Kingdom, studies comparing the outcome following major colonic resection in this group are few. Methods: This is a single center retrospective study of prospectively collected data (cancer database) of all colonic cancer resections performed between January 1, 2011 and December 31, 2013. We excluded emergent colonic resections, colonic stent insertions, surgery solely involving stoma formation, and advanced unresectable cancers. We compared outcomes for patients aged 70 or older with those of the control group, which included patients aged 69 or younger. We used the Z test and Mann-Whitney U test to calculate P value; P value less than .05 was considered significant. Results: Of the total 412 patients, 364 were included in the study. Of these, 178 (49.7%) were in the study group and 186 (50.3%) were in the control group. We found that 92.2% (165/178) of the study group were treated with curative intent, compared with 94% (175/186) of the control group .The postoperative length of stay was significantly longer in the study population (P=.014; Mann Whitney U test; range, 3–46 days; median, 7 days) than in the control population (range, 2–36 days; median, 7 days). We found no significant difference (P=.4009, Z test) with regards to mortality. The 30-day mortality was 5.6% (10/178) in the study group versus 3.7% (7/186) in the control group. Similarly, no significant difference ((P=.645, Z test) was noted related to anastomotic leaks (3.37% [6/178] in the study group; 4.3% [8/186] in the control group). Conclusions: This study suggests that septuagenarian and octogenarian patients have a similar complication rate after colonic surgery as younger patients. We found a statistically significant difference in inpatient stay; however, social circ*mstances play a role beside physiologic status in delaying discharges. In our study, 50% of the population was 70 or older; we suggest that geriatricians should be a part of a multidisciplinary cancer care team with members of the social care professions. This will expedite the discharge of patients, providing the earliest opportunity for patients to recover in home at their own pace. General surgeons with special interest in elderly populations may become the standard in the future.

Best Practices

AB2015-54. Analysis of Clinicopathologic Presentation of Breast Cancer in Sri Lankan Women: Single Unit Experiences During 2 Years' Followup

Thilakshi U Subasinghe, MBBS^a; Abdul J Hilmi, MD^b; Niroshan S Atulugama, MD, MRCS^c; Wasanthi Subasinghe, PhD^d; Vipuli J Kobbegala, MBBS^d; Prasad Pitigalaarachchi, MS, FRCS^e; and Padmasiri K.A. Siddhisena, PhD^b,e

^aNational Hospital of Sri Lanka, ^bTeaching Hospital Kandy, ^cNational Cancer Institute of Maharagama, ^dUniversity of Kelaniya, ^eUniversity of Colombo

Background: Breast cancer is the most common cancer among women worldwide, including in Sri Lanka. Global incidence of breast cancer is rising, with marked geographic variation. Incidence varies from 19.3 per 100,000 women in eastern Africa to 89.7 per 100,000 women in Western Europe.¹ According to data published by the National Cancer Institute of Sri Lanka, crude cancer incidence rate and age-standardized rate among women were 19 and 18.8 per 100,000 population, respectively, in 2007.² The rate has increased in in the past 20 years. A woman's lifetime risk of developing breast cancer is 2 per 100 women in Sri Lanka.² Most affected women are diagnosed in the late stage of the disease. Several clinical and morphologic parameters, such as histologic type of tumor, grade, stage, axillary lymph node status, closed or involved margins, and age younger than 50 years have been established as the risk factors for local recurrence in breast cancer.³ The goal of our study was to describe certain clinicopathologic features of breast cancer in patients visiting a tertiary level hospital in the central province of Sri Lanka covering approximately 7,032,000 people. The data have been analyzed with regard to different parameters, including demographic data (age and gender), morphologic patterns of tumor, such as size, grade, lymphatic-vascular invasion, TNM stage, hormone receptor status, local recurrence, presence of distant metastasis, and survival. Methods: We retrospectively reviewed clinical records for 101 patients with breast cancer treated at Teaching Hospital Kandy, Sri Lanka, between January 2012 and January 2103 at the Oncology clinic. Ethical clearance for accessing patients' medical records was obtained from the Director, Teaching Hospital, Kandy. All pathologic specimens were analyzed by 2 consultant pathologists working at the same institution. We used American Joint Committee on Cancer and the TNM system for staging. For grading, the Nottingham grading system was used. Data were analyzed using SPSS 20 statistical software. We evaluated mean, standard deviations (SD), and odds ratios (OR), along with 95% confidence interval (CI). Results: The mean patient age at presentation was 55 years, ranging from 31 to 88 years. The highest incidence was in the age group of 51 to 60 (31%) years of age. Of 101 patients, 97 were women and 4 were men. Right (51%) and left (49%) breast were equally involved. Size of the tumor at presentation was analyzed. TNM classifications of T1, T2, T3, and T4 were identified in 28%, 54%, 7%, and 11%, respectively, in the study population. More than half of the patients presented with a maximum tumor diameter of T2 (2–5 cm). Different morphologic patterns of histology were seen. The most common histologic type was invasive ductal carcinoma not otherwise specified (86%). This was followed by, in decreasing order, ductal carcinoma in situ (4%), invasive lobular carcinoma (3%), invasive papillary (2%), mucinous carcinoma (2%), mixed invasive ductal and lobular carcinoma (1%), invasive squamous cell carcinoma (1%), and undifferentiated (1%). Of the patients already diagnosed with invasive ductal carcinoma, 22% also had multiple foci of ductal carcinoma in situ in dissected histology specimens. The resection margins were free of tumor cells in 88% of patients and from them only 3% had a resection margin of less than 1 mm. Vascular invasion, neural invasion, and necrosis were found in 31%, 2%, and 21% of histology specimens, respectively. Skin infiltration and muscle invasive disease were found in 10% and 5% of patients at presentation. Four percent of patients had Paget's disease of the nipple with invasive ductal carcinoma. Histologic grading of tumor cells was seen in the following frequencies: grade 1, 12%; grade 2, 47%; and grade 3, 28%; 13% of patients had undetermined grade (GX). Irrespective of tumor size or lymph node status, 94% of patients had undergone modified radical mastectomy with level 2 axillary clearances. Level 3 axillary clearances had been performed in 4% of patients due to macroscopic positivity of level 2 lymph nodes. Only 1% had undergone breast-conserving therapy (wide local excision and axillary clearance followed by radiotherapy). None of the patients in the sample underwent sentinel lymph node biopsy before axillary clearance. Of patients who underwent modified radical mastectomy and level 2 axillary clearances, 45% had negative axillary lymph nodes in their histology specimen analysis. Two patients (2%) who underwent level 2 axillary clearance had positive level 2 nodes with level 1 lymph nodes being spared. Therefore 57% of patients had positive axilla. None of the patients had positive level 3 nodes. Twenty seven percent of patient were diagnosed with a TNM stage of 2A (Table 1). The next highest frequency was stage 2B (22%). Five percent of patients had distant metastasis at presentation. All 5 patients had liver metastasis, and 2 others had both brain and lung metastasis in addition to liver metastasis. Eight patients were diagnosed with new liver, brain or lung metastasis within 18 to 24 months. Regarding hormone receptor status, 42% of patients had positive estrogen receptors. All of these patients were treated with tamoxifen. HER2/neu receptors were positive in 13% of patients, and 29% were diagnosed with triple-negative disease. Hormone receptor status was either not performed or reports were not available in the clinic records for 20 patients. Regarding local recurrence, 3% of patients were noted to have local recurrence, and 87% showed no clinical or mammographic evidence of recurrence. However, unfortunately, 10% of patients were lost to followup by 24 months. Thirteen percent of patients had distant metastasis. We found a statistically significant association between local recurrence and size of tumor (P<.05) in this study population. No statically significant association was found between local recurrence and histology type, grade, resection margin of less than 1 mm, lymph node status, or muscle invasion. Conclusions: Most of the Sri Lankan women in our breast cancer study presented between 50 and 70 years of age (71%). In our study population, 54% of women had T2 tumors. The most common histology type was invasive ductal carcinoma, which is comparable to the findings in the rest of the world. Only 28% had high-grade tumors, and 84% of the patients had a TNM stage less than T3N1M0. Therefore, most the women in our study were diagnosed with early-stage breast cancer. We also found that 11% presented with locally advanced breast cancer and only 5% presented with metastatic disease. As the field of breast cancer and its management are evolving rapidly with evidence-based data, we need to focus more on performing sentinel lymph node biopsy and breast conservation therapy, because most patients are diagnosed at early stages. Early detection to improve breast cancer outcome and survival remain the cornerstone of breast cancer control worldwide. It is important to emphasize the need of continuous evaluation of breast screening programs and to promote self-breast examination amongst high-risk groups.

Best Practices in Implementation and Use of Clinical Practice Guidelines

AB2015-55. Update on NCCN Guideline Development and Public Transparency Process for 2014

Iris Tam, PharmD, and Jordana Wollmann, PharmDGenentech

Background: The NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) are developed through a rigorous review process that includes public posting of panel meeting minutes for transparency. According to the Institute of Medicine (IOM), transparency is one of 8 key standards for developing trustworthy guidelines. We previously reported trends from minutes posted from April 2010 through September 2013. This report is an update through 2014. Methods: NCCN panel meeting minutes were reviewed and the following data were extracted: date, panel, request and source of request, decision, and references (categorized by level of evidence). Results: In total, 753 panel decisions were made; of these, 92% resulted in an update and 8% resulted in no update to the NCCN Guidelines. NCCN classifies the source of requests as “internal”—via NCCN member institutions, “external”—via external parties (eg, pharmaceutical companies), or “internal/external”—where both made the same request. Most requests were internal (77%), but external requests are increasing each year. Internal requests resulted in an update more often than external requests (99% vs 64%). The levels of evidence cited were similar between update and no-update decisions with the exception of systematic reviews and meta-analyses, which were only cited in no-update decisions (Table 1). A full publication was more often cited in the update compared to the no-update decision group (76% vs 62%, respectively). A subset of external requests was analyzed. Phase III or II data was cited for 92% of updates compared with 71% of no-update decisions. Interestingly, there was no difference in citing abstracts/posters in the update and no update groups (35% vs 36%) whereas a difference was seen when all requests were considered (23% vs 35%). For external requests, 82% with FDA approval for the specific request resulted in an update while 59% without FDA approval resulted in an update. Conclusions: Nearly all internal requests resulted in an update to the guidelines. The level of evidence cited most often was phase III or II data regardless of decision and source of request. For external submissions, FDA approval for the specific request more often resulted in an update but was not essential for an update. The NCCN panel meeting minutes provide key insights and improved transparency into the NCCN guideline development process in accordance with IOM standards.

Outcomes and Health Services Research

AB2015-56. Predictors of Healthcare Costs During Ipilimumab Therapy for Advanced Melanoma: Results from a Retrospective Chart Review Study

Ahmad Tarhini, MD^a; Sumati Rao, PhD^b; Xiang Ji, MS^c; Shelby Corman, PharmD, MS^c; Sonam Mehta, MS^c; and Marc Botteman, MA, MSc^c

^aUniversity of Pittsburgh Cancer Institute, ^bBristol-Myers Squibb, and ^cPharmerit International

Background: Ipilimumab has been shown to prolong survival in patients with advanced melanoma (AM). Although the drug costs associated with ipilimumab are well understood, other, nondrug-related healthcare costs have not been evaluated among patients receiving ipilimumab in clinical (“real-world”) settings. This analysis aimed to identify demographic and clinical characteristics that are predictors of health care costs during ipilimumab therapy for AM in the US community setting. Methods: We analyzed data from a retrospective chart review of patients with unresectable stage III or IV melanoma treated with ipilimumab 3 mg/kg as first-line monotherapy between April 2011 and September 2012. Healthcare resource use data included hospitalizations, emergency department visits, laboratory tests, radiation, surgeries, specialist visits, nursing home use, and hospice visits. Publically available US unit costs were applied to each type of resource used to estimate total costs. The unit of analysis of health care costs was a dose interval, defined as the period between consecutive ipilimumab doses. Generalized linear mixed models with random intercepts were used to explore potential predictors of costs per dose interval. Fixed effects included occurrence of a grade 3 or 4 adverse event (AE) during the dose interval, ECOG status before administration of the dose, site of melanoma, presence of brain metastases, and stage at AM diagnosis. A miniscule number (1 e-05) was added to zero-cost dose intervals to avoid being excluded from regression. Results: Data were abstracted from 273 patient charts at 34 sites. Most patients had cutaneous melanoma (88.3%), stage M1 at AM diagnosis (87.9%), and had no prior brain metastases (87.9%). Baseline ECOG status was 0 in 38.1% of patients, 1 in 42.5%, 2 or greater in 7.0%, and unknown in the remaining 12.5%. Patients received a mean of 3.6 doses of ipilimumab (range, 1–5 doses), resulting in 988 analyzable dose intervals. A grade 3 or 4 AE occurred during 77 dose intervals (7.8%). Total health care costs, excluding ipilimumab drug costs, were 27 times higher (95% confidence interval [CI], 12.8–59.7) during dose intervals with a grade 3 or 4 AE (P<.0001). ECOG status 2 or higher was associated with a 4.5-fold increase in costs (95% CI, 1.94–10.32) compared with ECOG status of 0 (P=.0005), and cost per dose interval was 2.3-fold higher (95% CI, 1.06–4.80) in patients with cutaneous compared with noncutaneous melanoma (P=.0355). Health care costs were also higher in patients with stage M1 disease and brain metastases at initial diagnosis of AM, but these differences were not statistically significant. Conclusions: Grade 3 or 4 AEs were associated with a marked increase in health care costs but occurred in a small proportion of dose intervals. Higher ECOG status and cutaneous melanoma were also associated with significant increases in cost per dose interval.

Clinical Oncology

AB2015-57. Isolated Limb Perfusion, an Alternative Treatment in Poor Prognosis Sarcomas: First Latin-American Experience

Luis Alberto Tavares-de la Paz, MD; David Felipe Navarro-Barquín, MD; Gerardo Chávez-Pérez, MD; and Carlos Eleazar López-López, MD

Hospital Regional de Alta Especialidad del Bajío

Background: Soft tissue sarcomas (STS) are a rare group of malignancies; most occur on the limbs. At diagnosis, most of the tumors are large with complex involvement of neurovascular, osseous, and articular structures. In Mexico, most reach a national oncology referral center at a very late stage, and they are usually unresectable at that time. Thus the traditional therapeutic approach is mutilating functional surgery or loss of the limb. Isolated perfusion of the limb (IPL) has shown to be a safe and effective therapeutic option in the treatment STS with poor prognostis, with consistent limb salvage at approximately 80%. ILP is a one-event strategy that fits patient needs well in Mexico, where low-income patients often need transportation to high-volume hospitals to receive treatment. Methods: We reviewed 17 cases with at least 1-year follow-up between June 2010 and October 2014; all patients were candidates for amputation for locally advanced limb STS without the presence of systemic metastasis at perfusion. IPL was performed under mild hyperthermic conditions with 1 to 2 mg of tumor necrosis factor (TNF) and 50 to 150 mg of melphalan. The statistical analysis was done using SPSS 20 for a MacIntosh computer. Results: Of 17 patients, the average age was 37.4 (±15.5) years; 41.2 % were male and 58.8% female. The diameter of the tumor was 61.12 (±58.8) mm; 73.4% were in the lower limbs and 26.6% in the upper limbs; 88.2% were multicompartmental and 82% also HAD extracompartmental invasion; 64.7% had neurovascular and 41.2% had articular or bone invasion; 11.8% were multifocal. In our study, 58% were recurrent, and 76.4% of patients had received previous diverse treatments. The ILP was done with TNFα doses of 1.4 (±0.5) mg and 81.1 (±24.2) mg of melphalan and with a perfusion time of 87.6 minutes. The average outflow was 2.08%. Resection after ILP was feasible in 94%. The followup median time was 25 months, with a salvage rate of 94%. We documented 1 case of severe local toxicity grade V in the Wieberdink scale; no grade IV toxicity; 29.4% grade III; 52.9% grade II; and 11.8% grade I. Systemic toxicity was noted in 11.8%, which was hypotension in the postsurgical period. In our study, 23% of patients required adjuvant radiotherapy because of close margins or poor histologic response. We noted complete responses in 29.4% of patients. Synchronic local and systemic recurrence developed in 11.8%, systemic recurrence in 17.8%, and isolated local recurrence in 23.5%, with an overall survival in our followup period of 70.6% and a global recurrence of 52%. Conclusions: In our experience, ILP is an effective alternative in high-grade, advanced, and complex STS. In the setting where geographical distances are an issue for low-income patients needing treatment in high-volume centers, ILP offers an effective, safe, and 1-event strategy for limb sparing. However, our sample size and followup are small; further research is needed to confirm these results.

Quality Improvement

AB2015-58. Creation of a Comprehensive Oncology Fall Prevention Program

Tahitia Timmons, MSN; and Nicole Worthington, MBA, CPHQ

Cancer Treatment Centers of America at Eastern Regional Medical Center

Background: Health promotion programs seek to address disparities and create equalities by improving access for vulnerable populations. Oncology patients are a vulnerable population, as defined in the literature, due to risk for poor physical, psychologic, or social health. Oncology health promotion has traditionally focused on prevention of risk factors such as smoking and on early screening to accomplish this. However, falls are of particular concern because they can have a devastating outcome due to underlying pathology and they can impact outcomes. Creating tailored fall-prevention health promotion programs can promote continuity of care and help patients maintain function. Methods: Multifactorial fall prevention focuses on reducing environmental factors, anticipating patient needs, using screening tools, and increasing the culture of safety. These programs have reduced falls; however, special populations present challenges. A literature search revealed that most of these programs are nursing driven; few are interdisciplinary, and even fewer involved non-clinical staff. Another missing component was patient and caregiver perspectives on fall prevention. The program we developed used clinical and nonclinical staff to examine, design, and implement interventions for preventing falls in oncology patients. This program was based on information gained from the literature search and historical chart reviews. A fall-prevention committee composed of 6 teams facilitated the process. The teams focused on patient education, stakeholder education, environmental safety, pharmacy/rounds, informatics, and holism. Each team was interdisciplinary and reported to the committee. The patient and caregiver perspectives were incorporated through the use of a post-fall “patient huddle.” The underlying goal was patient-centered health promotion to ensure that patients maintained access to their oncology regimens. Results: Fall rates began to trend down after initiation of the program. This is evidenced by a reduction of the inpatient fall rate by 17% from calendar year 2014 quarter 1 and quarter 2 to calendar year 2014 quarter 3. The falls with injury rate per 1000 days was 0 for all 3 quarters. The post-fall patient huddle allowed themes to be trended within the patient population regarding behaviors and learning. Trends that emerged were incorporated into educational activities focused on health promotion. Conclusions: Successful health promotion must incorporate multiple factors, including clinical, nonclinical, patient perspectives, and the environment of influence. These must be included as part of a patient-centered culture change, as ways to address potential weakness and create sustainability. The oncology patient has multiple challenges; the hardship of a preventable fall should not be one of them.

Clinical Oncology

AB2015-59. Narrow-Band-Imaging: A New Tool in the Intraoperative Definition of Surgical Resection Margins in Oral and Oropharyngeal Carcinoma

Giancarlo Tirelli, MD; Marco Piovesana, MD; Annalisa Gatto, MD; Pierluigi Bonini, MD;and Francesca Boscolo Nata, MD

Azienda Ospedaliero-Universitaria Ospedali Riuniti Trieste

Background: A 3-dimensional surgical resection 1.5 to 2 cm from the gross tumor edge is currently considered appropriate, and the histologic status of resection margins (clear/close/involved) is the most reliable indicator of oncological radicality. Awareness of the “field cancerization” phenomenon calls for a reevaluation of the benchmarks of oncologic surgery; its identification is not simple because the dysplastic areas may be far from the main lesion and difficult to recognize macroscopically. The systematic use of new technologies such as narrow band imaging (NBI) could improve the detection of neoplastic and preneoplastic areas, ensuring more precise resections. The main purpose of this pilot study was to investigate the value of NBI technique in detecting precancerous (dysplasias) and cancerous areas around the tumor bulk intraoperatively, to achieve adequate resection of entire local tumor and maximizing the number of free resection margins at definitive histologic examination. Methods: In this pilot study, 15 previously untreated patients with primary oral cavity (n=7) or oropharyngeal (n=8) biopsy-proven carcinoma, staged with CT and/or MRI and undergoing surgical treatment with curative intent were considered. The surgical resection margins were first drawn, maintaining 1.5 cm from the macroscopic lesion boundaries defined visually and by palpation and then redefined using NBI. The 2 templates were compared and the distance between them measured; resections were performed following the second drawing. The number of clear margins, both on extemporaneous and definitive histologic examinations, and the presence of dysplasia and/or cancer in the NBI-guided resection enlargement were evaluated. Results: The use of NBI resulted in a resection enlargement from 8 to 12 mm on average. Among the 15 patients treated, the resection margins were free of dysplasia or cancer at extemporaneous examination; on definitive histology, one patient (6.7%) with oropharyngeal carcinoma had a margin positive for cancer. The NBI-defined enlargement was positive for dysplasia in 4 of 15 patients (26.7%) and for cancer in 11 of 15 (73.3%). Despite the larger resection, no increase in postoperative morbidity was seen with regard to either swallowing or sound articulation. Conclusions: The method we proposed could be useful for obtaining free surgical margins, and reducing the potential development of tumor foci resulting from incomplete resection.

Clinical Oncology

AB2015-60. Intraperitoneal Chemotherapy, a Less-Preferred Option in Clinical Practice

Yazhini Vallatharasu, MD; Sigurdur Bodvarrson, MD; Angela Smith, MA; and Andrew Borgert, PhD

Gundersen Health System

Background: After the results of the GOG 172 study published in 2006 reporting significant survival advantage with intraperitoneal chemotherapy (IP), the National Cancer Institute recommended that IP be considered for patients with optimally debulked stage III ovarian cancer. There have been reports of IP being used less frequently in clinical practice and low completion rates outside clinical trials. Therefore, we decided to study the frequency of IP use, survival differences, and completion rates in our community hospital. Methods: Patients with a stage IIIC ovarian cancer diagnosis between 2008 and 2013 at Gundersen Health System were identified. A total of 32 charts were reviewed, of which 22 met inclusion criteria. Patients who did not receive optimal debulking (6), were treated at an outside facility (3), and were participating in a clinical trial (1) were excluded for analysis. Kaplan-Meier curves were generated for overall and disease-free survival and compared via the log-rank test. For overall survival, surviving patients were censored at the date of last contact. For disease-free survival, patients still in remission were censored at the date of last contact or at the date of death. Results: Of the 22 included patients, discussion about IP therapy took place with 11; one declined and one did not qualify due to chronic kidney disease. Therefore, 9 patients agreed to undergo IP chemotherapy and 5 completed the proposed number of cycles (55%). Abdominal pain caused 2 patients to discontinue IP early, 1 developed pneumatosis intestinalis, and 1 stopped due to treatment failure. Patients receiving IP treatment were significantly younger, with a mean age of 55 versus 74 years of those given intravenous therapy only (P=.005). No significant difference was seen in overall and disease-free survival between the IP/intravenous and intravenous-only groups (P=0.65 and 0.18). Conclusions: Despite the survival advantage and tolerability of IP chemotherapy in patients with stage III ovarian cancer demonstrated by the GOG 172 trial (with fewer number of cycles), IP chemotherapy is not as frequently presented as a treatment option in clinical practice. In our institution, only 50% of eligible patients were offered IP chemotherapy and 90% of those patients proceeded with it. We also found that the patients offered IP were significantly younger. Our IP completion rate was 55%, which is higher than the 42% reported in the GOG 172 trial (with 6 cycles). However, our institution used a modified regimen with a lower dose (75% of carboplatin used in GOG 172) and in some cases fewer treatment cycles (4 cycles instead of 6). The most common reason for noncompletion in our study was abdominal pain, compared with catheter-related complications in GOG 172. No difference in overall survival was seen between the IP and intravenous therapy groups, but it should be noted that this was a retrospective study with a limited patient population.

Epidemiology/Risks

AB2015-61. Prevalence of Tumor Subtypes Among Long-Term Survivors of Brain Metastatic Breast Cancer

Vipin Villgran, MD, MMS^a; Aju Mathew, MD, MPhil^a; Margaret Rosenzweig, PhD, CRNP-C, AOCN^a; Shira Abberbock, MS^b; Rohan Naik, MD^a; Mini Jacob, MD^b; and Adam Brufsky, MD, PhD, FACP^a

^aUniversity of Pittsburgh Medical Center, and ^bUniversity of Pittsburgh

Background: Brain metastases occur in 20% to 30% of patients with metastatic breast cancer and it portends a dismal prognosis. Approximately 80% of patients die within a year of diagnosis of brain disease. In a single-institution cohort of patients with breast cancer with brain metastasis, we aimed to describe the proportion of breast cancer tumor subtypes among long-term survivors of brain metastatic breast cancer. Methods: Patients with breast cancer with brain metastasis treated at a university-based academic practice from 1997 through 2013 were identified. Clinicopathologic characteristics and survival information were prospectively collected. Hormone receptor (HR) status was assessed using immunohistochemistry (IHC), and HER2 status was determined using IHC or fluorescence in situ hybridization (FISH). Tumor subtype category was defined as HR+/HER2−, HER2+, and triple-negative (TN). Long-term survivorship from brain metastatic breast cancer was defined as survival more than 2 years from diagnosis of brain metastatic disease. Results: We identified 62 long-term survivors (25%) in our cohort of 243 patients. Only 10% of patients with TN disease survived for more than 2 years as compared with 38% of those with HER2+ disease. Of 85 patients with HR+/HER2− disease, 21% were long-term survivors. Among the long-term survivors of brain metastatic breast cancer, 61% had HER2+ disease. TN subtype represented only 10% of these patients. We noted an association between tumor subtype and long-term survival (Chi-square P-value = .001; Table 1). Conclusions: One in 4 patients with brain metastatic breast cancer is alive after 2 years from the diagnosis of metastatic brain disease. In our cohort, HER2+- disease was more commonly seen among long-term survivors than the TN subtype.

Outcomes and Health Services Research

AB2015-62. Acceptability and Effectiveness of a Novel Internet-Based Decision-Making Aid Based on the NCCN Non-Small Cell Lung Cancer Patient Guidelines

Sue S. Yom, MD, PhD; Matthew A. Gubens, MD, MS; Taylor Dunbar, BA; and Jeffrey Belkora, PhD

University of California, San Francisco

Background: Many physicians are reluctant to disclose specific details to cancer patients. Discussions are detailed and prolonged, leading to information loss and misunderstandings. Recently, decision aids have been developed to help patients by engaging them more directly in the decision process. This study tests an interactive decision aid designed to facilitate improved physician-patient communication, hypothesized to yield benefits in knowledge, decision certainty, and satisfaction, and to improve NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines)–guided decision making. Methods: We are field-testing a decision aid based on the NCCN Guidelines for Patients for non-small cell lung cancer (NSCLC). After initial prototyping, a pre/post design is used to estimate the effect size of the intervention with respect to disseminating NCCN-guided treatment choices. A database is used to characterize existing treatment decisions. After implementation of the aid, we establish its acceptability, feasibility, and effectiveness and evaluate changes in practice patterns before and after implementation. Quality of life, decision making practices, decisional conflict, and satisfaction with health care decision making are queried as secondary endpoints. Results: The decision aid programming and pretesting is complete (Figure 1). We have begun to enroll patients into the prospective trial. The primary endpoint of the trial compares baseline data to results obtained after implementation of the decision aid. Five patient-reported instruments will be used to score overall and lung cancer-specific quality of life, decision-making preferences, decisional conflict, and satisfaction with decision. In 12 months, we plan to enroll 360 patients. Anticipating a 30% decline, dropout, or ineligibility rate, at least 250 patients should be evaluable. For each disease stage, a test of at least a 25% improvement using the decision aid compared with the preintervention results in the proportion of patients satisfying the stage-specific criterion listed above will be performed. In addition, the entire study sample will be evaluated to test for a change in smoking cessation plans. Patients will be included in up to 3 individual analyses. The maximum sample size for any one test is based on assuming a null hypothesis of 50% in the preintervention cohort. Based on the 1-sample binary exact test with power set at 90% and a 2-sided type I error of 0.017 (adjusted for up to 3 tests for all patients), 53 patients would be required for each test. Conclusions: Translation of the NCCN Guidelines for Patients into a user-friendly, widely accessible Internet-based interface is feasible. Pretesting demonstrates that the aid is acceptable to providers, the research team, and laypersons. Enrollment has begun into a clinical trial to assess the value of this intervention

in influencing existing practice patterns based on increased involvement of patients in decision-making about their lung cancer treatments.

Correlative/Genomic

AB2015-63. FasL -844T/C and Fas -1377G/A Mutations of Pulmonary Adenocarcinoma in South China and Clinical Significance

Hongguang Zhao, MD^a; Wenhu Chen, MS^b; Peng Du, PhD^b; Aihua Sun, MD^b; Chenyu Zhuang^b; Jiali Tong^b; and Lifang Wang, MD^b

^aZhejiang Cancer Hospital and ^bZhejiang Medicine College

Background: Apoptosis is an important mechanism of malignant tumor formation and progression. Single nucleotide polymorphisms (SNPs) located within cell death genes may influence cancer risk. We explored the relationship between FasL -844T/C and/or Fas -1377G/A SNPs and pulmonary adenocarcinoma (AD). Methods: We randomly selected 275 patients with pulmonary AD in south China admitted into Zhejiang Cancer Hospital from July 2007 to October 2011 and collected clinicopathologic data at the same time. In this study, 297 healthy individuals were also selected as a control group. FasL-844T/C and Fas -1377G/A SNPs were detected using polymerase chain reaction–restriction random fragment length polymorphism (PCR-RFLP) technique to evaluate the relationships between these 2 SNPs and pulmonary AD. Results: Age, FasL -844 and Fas -1377 SNPs were associated with increased risk of pulmonary AD susceptibility in main effect analysis. FasL -844 CC and Fas -1377 AA were associated with an increased risk for the development of pulmonary AD in people younger than 60 years, but not in those 60 years or older. FasL -844 CC genotype was associated with an increased risk for pulmonary AD (adjusted odds ratio [OR], 2.010; 95% CI, 1.196–3.379; P=.008) compared with TT genotype. However, Fas -1377 AA was a risk factor only when FasL -844 genotype was CC. Fas -1377 genotypes showed significant effect modification of pulmonary AD risk by FasL -844 genotype with a test of the interaction term adjusting for age, gender, FasL -844 SNP. Fas -1377G/A was not associated with clinicopathologic factors, and FasL -844 C/T was associated with tumor stage and lymph node metastasis in patients aged 60 years and above, and with tumor stage in patients younger than 60 years. Conclusions: FasL -844 SNP is associated with susceptibility to pulmonary AD in patients younger than age 60 years. Fas -1377 SNP may modify the association of FasL -844 SNP with the risk of pulmonary AD. FasL -844 genotype plays an important role in the occurrence and progression of pulmonary AD.