The following abstracts were accepted for presentation at the NCCN 19th Annual Conference: Advancing the Standard of Cancer Care™ General Poster Session at the Westin Diplomat in Hollywood, Florida, on March 13-14, 2014. NCCN considered original abstracts from investigators in the oncology community related to Best Practices in Implementation and Use of Clinical Practice Guidelines, Quality Improvement, Outcomes and Health Services Research, and NCCN Oncology Research Program-funded Projects.
Best Practices in the Implementation and Use of Clinical Practice Guidelines
AB2014.12. Medicine for the Mind, Body, and Spirit: Integrating and Evaluating the Use of Distress Management Guidelines in Cancer Care
Samuel S. Gaster, MA; Julie A. Dixen, MSW, MBA; and Kristine I. Gaster, RN, MS, CNP
Avera Cancer Institute
Background: From soma to psyche, the effects of cancer extend well beyond abnormal, uncontrollable cells or invaded tissue to far within the deepest areas of a person’s mind, body, and spirit (van der Kolk, 1996; Institute of Medicine, 2008). Consequently, patients with cancer often experience a range of physical, psychological, and spiritual problems. Therefore, to provide the highest quality of care empirically-validated models should be adopted to provide routine psychosocial services that address these complex needs (IOM, 2008; IOM, 2013).
Recently, the Avera Cancer Institute in Sioux Falls, South Dakota, adopted the NCCN Guidelines for Distress Management. These guidelines were used to develop a protocol to assess the level and nature of patients’ distress using the NCCN Distress Thermometer and refer patients to support staff based on their responses to a 46-item adapted Problem List. Compliance was evaluated using the American Psychosocial Oncology Society’s quality indicators for psychosocial care.
Methods: Data include audits of 1544 office visits in the medical oncology clinic at the Avera Cancer Institute from October 7th to November 29th, 2013. For data analysis, charts were first reviewed for evidence that current distress was screened. Next, records were evaluated for evidence that problems were assessed in each office visit when moderate to severe distress (ie, >3) was reported. Third, charts were assessed for evidence that action was taken to address reported problems.
Results: Results indicate that current distress was assessed in 1488 (96.37%) of the 1544 office visits and that policy compliance was observed in 1522 office visits (98.58%). Of the 1488 office visits in which distress was screened, moderate to severe distress (ie, >3) was reported in 361 (24.26%), and of these, the nature of patients’ distress, or their problem areas, were assessed in 322 of the visits (89.20%). Lastly, evidence of action taken to address patients’ problems was available for 361 (100%) of the office visits in which patients expressed moderate to severe distress, with referrals documented in 129 (35.73%) of these visits and evidence of patient refusal in 232 (64.27%).
Conclusions: These results indicate a high compliance with guidelines for effective distress management and quality indicators for psychosocial care. This project also provides support for increased multidisciplinary and collaborative care to meet certain needs, including depression and traumatic stress. Thus, future efforts should be made to include a more formal model to address these needs. It is possible that existing models within primary care may prove useful to deliver such services.
AB2014.13. CAPTURE (Compliance and Preference for Tamoxifen Registry) Patient Survey Reveals Potential Strategies to Improve Long-Term Adherence to TAM Based on Choice: Results of a Large Internet-Based Survey
Stefan Glück, MDa; Jivesh Sharma, MDb; Peter C. Belafsky, MDc; Mary K. Delmedico, PhDd; Susan E. Spruill, MSe; Liz Rolisonf; David J. Drutz, MDd; on behalf of the CAPTURE Study Group
aSylvester Comprehensive Cancer Center; bNexGen Oncology; cUniversity of California, Davis School of Medicine; dDARA BioSciences, Inc.; eApplied Statistics and Consulting; and fBogier Clinical and IT Solutions, Inc.
Background: Despite the well-documented importance of adherence to 5 full years of TAM for long-term positive outcomes in women with hormone receptor-positive breast cancer, and despite the fact that no more than 50% of patients adhere to full length TAM therapy, few data exist on reasons for suboptimal patient adherence. Optimizing adherence includes understanding the possible role of swallowing difficulties and patient preference for an oral liquid formulation of TAM over tablets. The March 2013 update to the NCCN Guidelines for Breast Cancer includes considering up to 10 years of TAM therapy. To successfully implement 10 years of TAM therapy, it is important to identify strategies to increase patient adherence.
Methods: CAPTURE (Compliance and Preference for TAM Registry) was a 36-question Internet-based survey for patients with breast cancer who were taking TAM to quantify adherence to TAM, patient preference for an oral liquid version of TAM, the prevalence of swallowing difficulties in patients with breast cancer taking TAM, and the potential impact of formulation options on adherence. The CAPTURE survey was administered by health care professionals to patients in 17 United States-based sites and included a validated measure of swallowing difficulties: the EAT-10 tool. Women aged 18 years and older with a diagnosis of breast cancer who were currently taking either adjuvant or therapeutic TAM were included in the survey.
Results: The 626 survey respondents had a median age of 53 years, had been treated with surgery (97%) and radiation (58%) and/or chemotherapy (47%), and had been taking TAM for a median of 2 years. Fourteen percent (14%) of patients reported that they miss between 2 to 10 doses per month, and 12% of patients indicated that they had stopped taking TAM for periods longer than 2 weeks. Thirteen percent (13%) of patients indicated that they would prefer an oral liquid form of TAM versus a tablet, and 23% of patients were willing to try an oral liquid form. Additionally, 8% of patients showed evidence of swallowing difficulties, and this result was independent of their age. Seven percent (7%) of patients considered that an oral liquid form of TAM would help them improve their long-term adherence.
Conclusions: Within the limitations of self-reported data, these results provide an avenue for improving the adherence of patients to long-term TAM treatment by providing them a choice between a tablet and an oral liquid. The option of an oral liquid form of TAM could be used to help patients adhere to their TAM regimen, specifically for the 13% of patients who expressed a preference for a liquid formulation of TAM and the 8% of patients who displayed a need for a liquid formulation of TAM.
AB2014.14. Development of a Geriatrics Specific Curriculum Within a Hematology-Oncology Fellowship Program to Improve Training in the Treatment of Elderly Patients With Cancer
Thomas A. Giever, DO, MBA; Patrick C. Foy, MD; Gabriel Manzi, MD; Kathryn Denson, MD; Nicholas Dreger, BS; and Judi Rehm, BA
Medical College of Wisconsin
Background: The American Council of Graduate Medical Education mandates hematology-oncology (HO) fellowship programs educate fellows in the care of elderly patients. That education is typically undertaken as either case-specific teaching incorporated into routine clinical care or geriatrics-specific didactic lectures that emphasize learners’ knowledge rather than knowledge application.
Methods: A Geriatrics Educational Team (GET) was formed including an HO fellow, an associate HO program director, 2 geriatrics faculty, a medical student, and a medical educator. A Likert-type needs assessment administered by the GET identified the application of comprehensive geriatric assessment tools (eg, MiniCog, Montreal Cognitive Assessment, Barthel Index) as an educational gap. The curriculum included an initial didactic session introducing geriatric assessment tools followed a week later by an objective structured clinical examination (OSCE) and debrief session. The 2 OSCE cases used a standardized geriatric patient with lung cancer and a family member/caregiver presenting with a treatment dilemma. The cases were designed to elicit a history of either a physical or cognitive impairment to which geriatric assessment tools could be applied to better influence treatment recommendations.
Results: Ten fellows participated in the didactic session and OSCE. Postcurriculum evaluations using a 1 (low) to 5 (high) scale rating system reported that the OSCE was “an effective method for increasing confidence and skills in caring for elderly patients” (mean score, 4.6) and the ability of the fellow to “increase access and use of evidence-based geriatrics-specific point-of-care tools” (mean score, 4.3). The most common written comments (9 of 11 total) regarding changing practice indicated that the fellow will use a screening tool more often as a result of the OSCE. Ten fellows felt that the OSCE would enhance their ability to respectfully provide the highest level of interactions with patients, families, and caregivers (mean score, 4.1). The overall teaching effectiveness of the OSCE had a mean score of 4.4.
Conclusions: The OSCE is an effective and well-received method to apply geriatric-specific education when combined with traditional didactics as part of an established HO fellowship program.
AB2014.15. The Athena Breast Health Network as an Effective Model for Identifying and Personalizing Care for Women at Elevated Risk for Breast Cancer
Paige Kendall, BSa; Timothy Henderson, BAa; Alexandra Solomon, BSa; Celia Kaplan, DrPha; Lauren Ryan, MS, LCGCa; Elizabeth Crawford, MS, LCGCa; Elissa Ozanne, PhDa; Laura Esserman, MD, MBAa; and Laura van’t Veer, PhDb
aUniversity of California, San Francisco and bDartmouth University
Background: NCCN Guidelines for Breast Cancer Screening include a risk assessment and intervention for breast cancer risk reduction, but risk assessment is not well incorporated in clinical practice. The Athena Breast Health Network is a University of California collaborative network developed to improve breast cancer prevention, screening, treatment, and survivorship through research and quality improvement efforts.
Methods: Athena implemented a personalized breast cancer risk assessment and follow-up intervention model for women having a screening mammogram at the University of California, San Francisco (UCSF). From June 2012 to June 2013, women aged 35 to 65 years with no previous diagnosis of breast cancer or ductal carcinoma in situ were enrolled, who then completed a form and consented to research. Answers to this Web-based intake form were fed through the Gail risk model to stratify participants based on risk. Women aged 35 to 54 years with a Gail score greater than 1.67 and in the top tenth percentile and aged 55 to 65 years in the top fifth percentile were identified as being at elevated risk. The Breast Health Specialist (BHS), a certified genetic counselor in a novel clinical role developed by Athena, sent risk information to the elevated risk cohort and then conducted a telephone consultation.
Results: A total of 982 eligible women completed the intake form and consented to research. The mean sample age was 50 years (SD, 7.8), and 60% were non-Latina White, 6% Latina, 4% African American, 21% Asian/Pacific Islander, and 9% Native American/other. A total of 6.5% of the women were classified as being at elevated risk for breast cancer, of which 78% received risk information and were contacted by the BHS for a follow-up consultation. Risk consultations were not completed for some women based on clinical judgment by the BHS. Of those contacted, 70% completed a telephone risk consultation, 52% were previously unaware of their increased risk status, 36% were referred to the UCSF Breast Cancer Prevention Program, and 18% were referred to genetic counseling. Results are presented in Table 1.
Conclusions: The Athena model effectively identifies women previously unaware of their elevated breast cancer risk status. The findings support the value and efficiency of a program like Athena’s for identifying and providing increased screening follow-up as recommended by NCCN Guidelines. The exclusion of women who were not able to complete a Web-based intake form limits this model.
AB2014.15. Table 1: Risk Assessment and Intervention Analysis and Intervention Analysis
AB2014.16. Improving Patient-Centric Quality Outcomes Through Real-Time Extraction of Data from EPIC Electronic Health Record
Julie Kuznetsov, MBA; Kathryn Gail Bailey, RN, OCN, MSN, MBA; Pacharintra Bombach, MSE; Xuemei Chen, MS; Terri Owen, CTR; Angela Primeau, MSN, RN, CNS-BC, OCN; and Douglas W. Blayney, MD
Stanford Cancer Institute
Background: In 2011, the clinical leadership at Stanford Cancer Institute (SCI) began process improvements to enhance patient satisfaction and quality of care. To measure impacts, unnecessary care variation, and outcomes, an informatics infrastructure was developed using data from SCI’s electronic health records (EHRs; Epic modules: BEACON, Cadence, OpTime, ADT). Initial efforts were successful, including real-time cohort identification (compared with “gold standard” registry data), improvements in time to treatment, and profiling variation in short- and medium-term morbidities based on variations in care.
Methods: A cohort of 2985 patients was defined by 3 criteria: newly evaluated at SCI, and received treatment (surgery, chemotherapy, or radiation) at SCI, and had a breast, thoracic, or gastrointestinal cancer-related ICD-9 code. Data were analyzed by fiscal year (FY), starting September 2011 (FY11) to August 2013 (FY13). Time to treatment was measured as the interval between first EHR time stamp for SCI patient contact and treatment date. BEACON protocols were used to identify chemotherapy regimens. The cohorts were compared with SCI tumor registry data, and presented in a dynamic QlikView dashboard.
Results: A total of 98% of the EHR-defined cohort matched a similarly defined tumor registry cohort. Use of a dynamic dashboard facilitated detection of the effect of process improvements (eg, scheduling a visit on first contact, coordinating among surgical specialties, and having outside records available previsit), which resulted in a significantly accelerated time to treatment even while occurring during a period of 5% compound annual growth rate of volume. The methodology is scalable and has been successfully replicated across different disease cohorts. Data extracted from the EHR showed that more than 80% of that patient cohort adhered to guidelines-consistent chemotherapy regimens (Table 1). Furthermore, intermediate clinical outcomes (eg, unplanned emergency visits and hospitalizations) were analyzed and a profile was provided of increased or decreased use of emergency services and hospitalizations associated with chemotherapy regimens in use with this patient cohort.
Conclusions: This work demonstrates the utility of EHR data for tracking patient-centric process improvements and intermediate clinical outcomes associated with specific patient populations and/or chemotherapy regimens. Future work will include use of these data and cost data to develop and inform care pathway choice.
AB2014.16. Table 1 Improvements in Time to Treatment
AB2014.17. Palonosetron for Chemotherapy-Induced Nausea and Vomiting: Description and Report of a Quality Management Program
Edward Li, PharmD, BCOP; Cecilia Tran, PharmD, BCOP; Barry Peterson, PharmD, BCOP, BCPS; Michael Sturgill; and Stanley Forston Jr, MD, MPH
New Century Health
Background: Palonosetron has demonstrated superior efficacy compared with other 5-HT3 antagonists in preventing emesis for both high- and moderate-emetic risk categories. However, no consensus regarding the best intervention exists for low emetogenic categories. Palonosetron is preferred by guidelines for patients undergoing high-/moderate-emetic chemotherapy, but other treatment options (other SHT3 antagonists) may be more cost-effective in low-emetogenic regimens. This study examines a cohort of denied chemotherapy treatment authorization requests containing palonosetron to determine a trend in demographics, authorization request patterns, and cost.
Methods: This is a retrospective analysis of denied authorization requests containing palonosetron reviewed by New Century Health’s oncology quality management program from January 2013 to October 2013. Descriptive statistics identify cohort demographics, reasons for withdrawal or adverse determination, and follow-up to authorization withdrawals (ie, changes in chemotherapy or supportive care regimen). Lastly, cost impact was calculated based on the difference between the approved authorization request and the original, withdrawn request.
Results: A total of 378 cases were identified; the cohort median age was 68 years, with 42% men and 58% women who were mostly from the Southern (60%) and Midwest (33%) United States and had public insurance (79%). The most common tumor types were lung (24%), breast (21%), colon/rectal/anal (10%), pancreatic (7%), and ovarian (7%). Hematologic malignancies constituted 11% of the cohort. Single-agent chemotherapy regimens were most common (233 requests), with docetaxel, paclitaxel, gemcitabine, pemetrexed, and fluorouracil totaling 125 requests. Combination regimens were less common (108 requests), with gemcitabine/nab-paclitaxel being the most common regimen (18 requests). Reasons for denial include no compendia support (81%), availability of therapeutic alternative (42%), clinical criteria not met (15%), and dosing/frequency errors (10%). The program prompted chemotherapy regimen changes in 156 cases and supportive care treatment changes in 222 cases; palonosetron was removed in 31% and 51% of the cases, respectively. The total net cost savings based on all the changes within the cohort was $1,087,073. Evidence-based changes in therapy that resulted in enhanced care quality and increased cost totaled $757,473.
Conclusions: Palonosetron, although effective, may not be the most cost-effective option for low-emetic regimens considering other lower-cost SHT3 antagonists are available. The investigators were able to direct use toward the lower-cost antiemetic medications and to influence changes in evidenced-based chemotherapy prescribing choices.
AB2014.18. Process for Identification of Risk Factors for Chemotherapy-Related 30-Day Readmissions
Jeremy Lund, PharmD; Georgia Keriazes, PharmD, BCOP, BCPS; and Angela Pearson, PharmD, BCPS
Lakeland Regional Medical Center
Background: At the 2013 NCCN Policy Summit on measuring quality in oncology, during a roundtable discussion focused on Perspectives of Quality After a Cancer Diagnosis, it was determined that significant opportunity exists for improvement in measuring quality care in oncology. Improvements are needed in several areas, including developing better measures to gauge the patient experience. In response to this call for better measures, a process for identifying risk factors for chemotherapy-related 30-day readmissions was developed at Lakeland Regional Medical Center, a tertiary care hospital.
With the recent changes to repayment structures by the Centers for Medicare & Medicaid Services (CMS) for selected diagnoses and the prospect of that list to grow, it is desirable for institutions to take measures to identify reasons for 30-day unplanned readmissions in the oncology patient population. Although the repayment structure as it relates to the oncology population has not been targeted by CMS to date, this does not preclude medical professionals from intervening in meaningful ways to prevent readmissions and performing research to improve patient experience.
Methods: The MIDAS DataVision Readmission Tool Pack (Version 2.x) decision support application was used to identify and provide standard baseline characteristics for patients who received inpatient chemotherapy between April 1, 2010, and May 31, 2013. Patients undergoing surgery during the primary admission, other than surgery for the routine diagnosis of cancer, were excluded. Thirty-day readmissions in this patient population were then identified.
In addition to the standard data provided by the decision support application, additional baseline characteristics specific to the oncology population were evaluated through manual chart review. Readmissions were reviewed by a multidisciplinary panel of oncology health care professionals to identify those deemed to be chemotherapy-related.
Results: Through using a combination of a standard readmission decision support application and development of a streamlined approach to manual chart reviews, 618 inpatient chemotherapy encounters and risk factors for chemotherapy-related readmissions were evaluated. Preliminary results show that the overall 30-day readmission rate seen in patients receiving chemotherapy was 37.2%. The average length of stay in the overall population was 4.49 days, whereas the average length of primary stay in the readmitted population was 5.82 days. Patients older than 64 years constituted 30.7% in the overall population and 34.3% in the readmitted population. Chemotherapy-related readmissions and further results and analysis are pending.
Conclusions: Ongoing results from this quality improvement process will be used by the Oncology Service Committee to develop strategies to reduce 30-day chemotherapy-related readmissions and improve the patient experience.
AB2014.19. A Matrix Solution to the Problem of False-Positives in Next-Generation Sequencing Data
Carl Morrison, MD, DVM1; Mia A. Levy, MD, PhD2; William Pao, MD, PhD2; Manuel Glynias3; Song Liu, PhD1; Lei Wei, PhD1; Jianmin Wang, PhD1; Mary Nesline1; and Jeffrey M. Conroy1
1Roswell Park Cancer Institute, 2University of Vanderbilt, and 3GenomeOncology
Background: The association of next-generation sequencing (NGS) results with knowledge resources to define actionable mutations is at the forefront of the delivery of personalized medicine for oncology. In the authors’ Clinical Laboratory Improvement Amendments-certified laboratory, a targeted sequencing NGS assay was designed that includes all genes for testing in My Cancer Genome. To address the inherent problem of minimizing false-positive calls while optimizing sensitivity, 2 different NGS methodologies were used for parallel confirmation: lon Torrent and IIlumina.
Methods: A parallel system of using the Iliumina MiSeq and lon Torrent Personal Genome Machine platforms to sequence all specimens within an IT systems control and a “Classify Calls” matrix solution for mutational analysis was designed. The “matrix” is a unique solution to the problem of false-positives in NGS, and requires a sophisticated set of engine rules and unique IT solutions to simultaneously manage the VCF and BAM files.
Results: Within the matrix, using both the lon Torrent Suite and MiSeq Reporter platforms, the authors were able to achieve 99% sensitivity and a 99% positive predictive value (PPV). Sensitivity as a single platform was greater than 95% for both PGM and MiSeq, but could not be maximally optimized because of the presence of false-positive results, with PPV ranging from 10% to 80% for formalin-fixed paraffin-embedded and fresh frozen specimens, respectively. In addition to random false-positive results, both platforms have systematic errors that are unique and gene-specific. As a single platform, the MiSeq was plagued by systematic false-positive results at high levels of coverage that could not readily be removed by quality filters.
Conclusions: Delivering only the highest quality NGS results to the oncologist for therapeutic decision-making requires near-perfect sensitivity and PPV. In addition, a high confidence of the results requires that all variants be confirmed with independent technologies. The lon Torrent and IIlumina technologies are examples of independent technologies and have less in common than, for example, Sanger Sequencing does with the latter of these 2 technologies. Through designing a matrix solution to parallel sequencing on the PGM and MiSeq, the authors developed a unique solution to this problem with near-perfect sensitivity and PPV that cannot be achieved with either platform alone. The only state in the United States that is currently asking for independent confirmation of NGS results is New York.
AB2014.20. Risk Factors For Postoperative Cognitive Change in Elderly Patients Undergoing Gastrointestinal Surgery
Jackson Ng, MD; Christopher Tan, MD; Rajkumar Jeganathan, MD; Fernando Kawai, MD; Simcha Pollock, MD; H. Leon Pachter, MD; Marcovalerio Melis, MD; James Turner, MD; Steven Cohen, MD; Cynthia X. Pan, MD; and Mitchell Chorost, MD
New York Hospital Queens
Background: Postoperative cognitive dysfunction (POCD) in the elderly is an increasing concern. The Cambridge Neuropsychological Test Automated Battery (CANTAB) was used to assess cognitive changes in an objective and sensitive manner. The goal of this study was to determine if the changes in CANTAB scores have a significant correlation to established risk factors (Charlson Comorbidity Index score, Physical Self Maintenance Scale score, polypharmacy, psychoactive drug use, nonpsychoactive drug use, preoperative frailty, and surgical technique used).
Methods: A prospective, nonrandomized, multidisciplinary clinical study, was conducted at an urban tertiary care hospital on patients aged 65 years and older, without documented dementia, who were undergoing elective colon surgery. Subjects underwent pre- and early postoperative CANTAB testing. Worsened individual scores from Paired Associated Learning (PAL) and Spatial Working Memory (SWM) portions of the CANTAB determined the presence of POCD.
Results: Review of the difference between most risk factors among improved and worsened CANTAB groups showed no statistically significant correlation. A significant difference in Charlson Comorbidity Index scores was observed with the change in PAL scores (P=.04) although it was paradoxically higher in the improved PAL group versus the worsened PAL group. Subjects with worsened CANTAB scores exhibited longer hospitalization (8.7 vs 7.33 days for PAL and 22.9 vs 8.1 days for SWM scores), despite being statistically insignificant.
Conclusions: This study showed that monitoring the pre- and postoperative changes in CANTAB scores provided a more sensitive and objective measurement of postoperative cognitive changes. Increasing the sample size would have demonstrated the effect of the studied postoperative cognitive changes.
Outcomes and Health Services Research
AB2014.21. Socioeconomic and Ethnic Variations in Survival of Patients With Non-Hodgkin’s Lymphoma in the United States
Shiva Kumar R. Mukkamalla, MD, MPH; Rajshekhar Chakraborty, MD; and Barbara Seligman MD
Icahn School of Medicine at Mount Sinai/Queens Hospital Center
Background: The incidence of non-Hodgkin’s lymphoma (NHL) in the United States has been increasing for the past several decades. However, mortality from NHL has seen a downtrend in recent years. As per the SEER factsheet, the annual percent change for NHL mortality from 1997 to 2009 was -3.01. This could be attributed to the use of the cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) regimen followed by the addition of rituximab. Research has shown that socioeconomic status influences diagnosis, treatment, and survival of patients with NHL. Several studies have shown ethnic disparities in past. However, some of them examined the elderly population alone, whereas the remaining failed to take into consideration ethnicities other than Hispanics, African Americans, and Caucasians. This study investigated socioeconomic and ethnic disparities in the survival of patients with NHL belonging to all age groups and all ethnicities.
Methods: A retrospective cohort of 22,419 patients with NHL diagnosed from 2004 to 2005, belonging to all age groups and ethnicities, was identified from the SEER database. Median follow-up was 4.5 years. Descriptive statistics (chi-square test) were used to analyze the differences in distribution of baseline characteristics by race. The Cox proportional model was used to compare survival among different ethnic groups using SAS 9.2.
Results: Of the cohort, 85.2% were Caucasians, 8.32% were African Americans, and 6.48% were from other ethnicities. The mean age of the study population was 62.2 ± 17.3 years. A large proportion of other ethnicities (non-black and non-white) were in the highest quartile of education (32.48%) and economic status (35.86%). Hazard ratio (HR) of all-cause and NHL mortality increased significantly with age, stage, unemployment rate, smoking rate, and language isolation rate, and decreased with higher household income (P<.05). The married population was found to have better survival than the single and divorced/widowed populations (P<.0001). After confounding for all other variables, including subtypes and stage at presentation, NHL mortality was found to be higher in African Americans (HR, 1.055; 95% CI, 1.003-1.109) and other ethnicities (HR, 1.113; 95% CI, 1.053-1.175) compared with Caucasians. No significant difference in NHL mortality was seen between women and men (P=.13).
Conclusions: This study shows a strong correlation between social deprivation and decreased survival in patients with NHL belonging to all age groups. Also, Caucasians seem to have higher survival compared with African Americans and other ethnicities. However additional studies, including all other possible prognostic variables, are needed to substantiate this finding.
Clinical Oncology Research Trials
AB2014.22. Predictive Markers of Response in a Phase II Trial of Axitinib in Advanced Melanoma
Nikhil I. Khushalani, MD; Austin Miller, PhD; Araba Adjei, PhD; Dan Iancu, MD, MSc; Peter Loud, MD; Holly Meyers, BS, CCRP; and Dominick Lamonica, MD
Roswell Park Cancer Institute
This study was approved and funded by NCCN from general research support provided by Pfizer.
Background: Therapy in metastatic melanoma remains largely palliative despite the recent approval of several new drugs. Angiogenesis is crucial for progression in melanoma. Increased circulating levels of proangiogenic factors have been linked to poorer prognosis. Axitinib is an oral, potent, and selective inhibitor of vascular endothelial growth factor receptor- (VEGFRs) 1, VEGFR2, and VEGFR3. This phase II trial examined the efficacy of axitinib in advanced melanoma and sought to identify biomarkers of response/toxicity as secondary correlates. In addition, 18F-fluorothymidine (FLT) was explored as an alternate imaging probe for cellular proliferation in a subset of patients.
Methods: Eligible patients must have unresectable or metastatic melanoma; had 0 to 1 previous systemic treatment regimens; adequate blood and organ function; ECOG performance status less than 2; urine protein/creatinine ratio less than 2; and baseline blood pressure less than 140/90 mm Hg. Salient exclusion criteria include prior antiangiogenic therapy, recent significant bleeding or arterial thrombosis, and a non-functional gastrointestinal tract. The starting dose of axitinib is 5 mg orally twice daily continuous, with planned escalation to 10 mg orally twice daily if parameters are met. Correlates include (1) blood at baseline for genomic DNA extraction to study single nucleotide polymorphisms in axitinib-related genes (VEGFR1, VEGFR2, and VEGFR3) using the Mass-ARRAY Compact System (Sequenom, San Diego, CA); (2) blood at baseline, 4 weeks, and 8 weeks for circulating melanoma cells using the CELLTRACKS circulating melanoma cell kit (Veridex, LLC, Raritan, NJ); and (3) FLT-PET at baseline and 4 weeks for the first 15 patients. In an exact 2-stage design for futility, greater than 2 responses in the first 20 patients permitted continuation to an accrual of 40 patients, with an 80% power to detect an overall response rate of 0.25 with axitinib. Secondary objectives include survival estimates and examining the correlative parameters for efficacy/toxicity.
Results: The study met its predefined efficacy criterion at interim analysis to proceed to full accrual. To date, 23 patients with a median age of 73 years (range, 40-88 years; 13 men) have been accrued. A total of 61% (14/23) have received prior therapy for metastatic melanoma. The most common toxicities (n>5) are fatigue, hoarseness, hypertension, diarrhea, mucositis, anorexia, dysgeusia, increased aspartate transaminase, and increased thyroid-stimulating hormone. These findings are commensurate with labeling information.
Conclusions: Trial accrual is ongoing. Efficacy of axitinib in the treatment of melanoma and its effects on survival will be determined after trial completion. FLT-PET data are being analyzed, and correlations between this modality and standard imaging-based criteria (RECIST) will be available shortly.
AB2014.23. Treatment Patterns and Health Care Costs of Advanced Gastric Cancer Among Elderly Patients in the United States
Maria Lorenzo, MS1; Astra M. Liepa, PharmD, BS1; Lisa M. Hess, PhD1; Sudeep Karve, PhD2; Brian Calingaert, MS2; and James A. Kaye, MD, DrPH2
1Eli Lilly and Co. and 2RTI Health Solutions
Background: The goal of this study is to describe treatment patterns and health care costs after first-line therapy among elderly patients diagnosed with gastric cancer (GC) in the United States.
Methods: A retrospective cohort analysis was conducted of the SEER-Medicare linked database. Patients aged 65 years or older with metastatic and/or unresectable GC (including gastroesophageal junction cancer) diagnosed according to ICD-O-3, from July 1, 2000 to December 31, 2007, were eligible if they had received a platinum and/or a fluoropyrimidine agent as first-line therapy. Continuous enrollment in Medicare Part A and B was required for 6 months or more before the index GC diagnosis. Patients were classified as additionally treated (any chemotherapy, biologic therapy, or radiotherapy received after first-line) or supportive care only (all other patients). Summary statistics are presented for demographic characteristics, treatments, and GC-specific resource use and heath care costs (2012 value).
Results: Of the 2583 eligible patients, 1752 (68%) were men. Mean age was 74.8 ± 6.0 years. At diagnosis, most patients had distant (1386 [54%]) or regional (541 [21%]) disease. Of all eligible patients, 1168 (45%) received supportive care only, whereas 1415 (55%) were additionally treated. A total of 1000 patients received second-line chemotherapy or biologic therapy. Of these, 358 (14%) of patients received single-agent therapy, 403 (16%) received doublet regimens, and 205 (8%) received triplet regimens. The most common regimens were docetaxel (8%), fluorouracil/leucovorin (7%), paclitaxel (6%), and carboplatin/paclitaxel (6%). Further results are shown in Table 1.
Conclusions: With high death rates and no consistent treatment strategy, patients with advanced GC have high unmet needs. Inpatient care was the largest component of health care-related costs. Results are limited to patients with SEER-Medicare linkage, but follow-up is relatively complete. Further research is needed for patients younger than 65 years and to identify optimal treatment practice.
AB2014.23. Table 1: Follow-up Duration (days) and Select GC-Related Total Costs Overall (US $ median, Q1, Q3) and by Cohort