Additional Abstracts from the NCCN 21st Annual Conference: Advancing the Standard of Cancer Care™ (2024)

The following abstracts were accepted for presentation at the NCCN 21st Annual Conference: Advancing the Standard of Cancer Care™ General Poster Session at The Westin Diplomat in Hollywood, Florida, on March 31 and April 1, 2016. Additional abstracts presented at the conference were published in the print edition of JNCCN (2016;14:(5.5):604--614), and are available at JNCCN.org.

Best Practices in Implementation and Use of Clinical Practice Guidelines

AB2016-23. Efficacy of the Antiemetic Combination Agent, NEPA, in Patients With Head and Neck Cancers Receiving Cisplatin-Based Chemotherapy

Lee Schwartzberg, MD^a; Eric Roeland, MD^b; Paul J. Hesketh, MD^c; Marco Palmas, MD^d; and Elio Carreras^d

From ^aThe West Clinic, ^bUniversity of California San Diego Moores Cancer Center, ^cLahey Health Cancer Institute, and ^dHelsinn Healthcare, Switzerland

Background: First-line chemotherapy for head and neck (H&N) cancers generally includes cisplatin-based regimens. Despite cisplatin's high emetogenic potential, chemotherapy-induced nausea and vomiting (CINV) can be prevented in most patients with appropriate antiemetic regimens. Antiemetic guidelines recommend that patients treated with cisplatin should receive a combination of an NK₁ receptor antagonist (NK₁RA), 5-HT₃RA, and dexamethasone (DEX). Adherence to antiemetic guidelines is unacceptably low, with patients frequently not receiving recommended antiemetic combinations. NEPA, which is composed of the highly selective NK₁RA netupitant (300 mg) and the 5-HT₃RA palonosetron (PALO; 0.50 mg), is the first oral medication targeting 2 CINV pathways in a single fixed dose in patients receiving emetogenic chemotherapy. NEPA + DEX has demonstrated superior prevention of CINV compared with oral PALO + DEX. This retrospective analysis evaluated the efficacy of NEPA in a subset of H&N patients with cancer receiving cisplatin from a pivotal trial. Methods: In this randomized double-blind trial, patients received either oral NEPA or oral PALO on day 1 before a single cycle of cisplatin-based chemotherapy. 3 NEPA dose groups (netupitant, 100/200/300 mg + PALO, 0.50 mg) that show similar efficacy were pooled for this subset analysis. All patients received oral DEX on days 1–4. Concurrent radiotherapy was excluded. End points were complete response (CR = no emesis and no rescue medication), no emesis, and no significant nausea (maximum <25 mm on 100 mm visual analog scale) during the acute (0–24 hours), delayed (25–120 hours), and overall (0–120 hours) phases postchemotherapy. Results: 115 patients (86% men, 14% women) with H&N cancers received either NEPA (N=91) or PALO (N=24). CR, no emesis, and no significant nausea rates for NEPA were >90% across all phases, with an absolute incremental benefit compared with PALO of 12% for CR and no emesis and 15% for no significant nausea during the overall phase (Table 1). Overall response rates for NEPA in patients with H&N were numerically higher than those seen in patients with non–H&N cancers (N=316; CR, 87%, no emesis, 88%, no significant nausea, 84%). Conclusions: NEPA, the first oral antiemetic combination targeting 2 critical emesis pathways in a single dose, was more effective than PALO in preventing CINV in patients with H&N cancer receiving cisplatin-based highly emetogenic chemotherapy. Administered as a single oral capsule before chemotherapy, NEPA may offer a convenient and highly effective option for CINV guideline adherent treatment.

AB2016-23: Table 1.

NEPA for CINV in Head and Neck Cancer

AB2016-24. The Impact of Contouring the Axillary Lymph Nodal Levels on Achieving Adequate Doses During Tangential Chest Wall Irradiation

Diana Tuaño, MD; Gaudencio Vega, MD; and Jose Roel Resubal, MD

From St. Luke's Medical Center, Philippines

This abstract is available at Int J Radiat Oncol Biol Phys 2014;90:S261

AB2016-25. Cooperation of Clinical Pharmacists and Physicians in the Management of Cancer Pain: Results from a Retrospective Analysis in Chinese Population

Zeng Wang; Ling ya chen; Hai-ying Ding; Bin Cheng; Guo-nong Yang; Ping Huang

From Zhejiang Cancer Hospital, People's Republic of China

Background: Cancer pain management is insufficient in China, including inadequate dosing, nonstandardized implementation of opioids, etc. Recently, Chinese clinical pharmacists have begun to pay more attention to the use of opioids in patients with cancer, and have taken an active role in patient care by organizing clinical pharmacist–led guidance teams. Here, we evaluate the role of clinical pharmacists' cooperation with physicians in the management of cancer-related pain. Methods: Since 2011, several clinical pharmacists have been assigned to specific departments to collaborate with physicians to standardize treatment for cancer pain. According to the standardized guidelines for cancer pain treatment, a special review table was designed in collaboration, and the relative medical records of patients who used sustained-release opioids in the second quarter from 2012 to 2015 were collected and analyzed. Results: A total of 603 patients were enrolled: 161 in the clinical pharmacist cooperation group and 442 controls (Table 1). Of these groups, patients whose cancer pain management adhered to the standardized pain treatment were 134 (83.2%) and 264 (59.7%), respectively (P<.001). Standardization of opioid administration was improved significantly in the clinical pharmacist cooperation group, including more standardized dosing titration (P<.001) and less reports of breakout pain (P<.001). Moreover, the pain scores in the clinical pharmacist cooperation group were significantly improved compared with the control group (P<.01), as well as overall quality of life (P<.05). However, there was no significant difference in the incidences of constipation between both groups. Further, according to dosing titration methods, patients were divided into a morphine on-demand titration group and oxycodone extended-release tablet, 10-mg titration group. Results showed that in the latter group, patients whose cancer pain management adhered to the standardized pain treatment were 63 (94%) and 90 (75.6%) with or without clinical pharmacist cooperation, respectively (P=.002). Conclusions: In cooperation with physicians, clinical pharmacists improved standardization, efficiency, and efficacy of cancer pain management, and found a new mode for clinical pharmacy improvement.

AB2016-25: Table 1.

Evaluation of the Standardization of Sustained-Release Opioid Treatment

Clinical Oncology

AB2016-27. Patient-Reported Distress in Patients With Central Nervous System Tumors

William Carraher-Stross; Ajaykumar Patel, MD; Jacob Habboush; Megan Single; Laeticia Hollant; Katherine Gaines; Bridget Smart; Heather Biers; Atiya Day; Robert C. Miller, MD; Laura Vallow, MD; and Jennifer Peterson, MD

From Mayo Clinic, Jacksonville

Background: Distress screening is an important component of central nervous system (CNS) cancer management. Identifying psychosocial distress allows oncology teams to better care for patients. Patient-reported distress (PRD) screenings play a pivotal role in the personalized care of patients receiving radiation therapy (RT) for primary CNS tumors. We report on patterns of distress in this patient group. Methods: We reviewed records for 61 patients with primary CNS cancers who received external-beam RT (EBRT) at our institution from April 2012 to May 2015. Overall distress was measured using the NCCN Distress Thermometer (DT) scaling from 0–10. A more targeted 30-question PRD screening tool (Likert scale, 1–5) was used to identify specific areas of distress. Results: 27 men (44.3%) and 34 women (55.7%) completed the 30-question PRD surveys and NCCN DT; median age was 53 years (range, 20–84 years). 52 patients (85.3%) received curative EBRT and 9 (14.8%) received EBRT with palliative intent. Tumors treated included 19 glioblastoma multiforme (31.1%), 19 astrocytomas (31.1%), 13 meningiomas (21.3%), 6 oligodendrogliomas(9.8%), 1 optic pathway glioma (1.6%), 1 paraganglioma (1.6%), 1 chordoma(1.6%), and 1 gliosarcoma(1.6%). 45 patients (73.8%) underwent surgery and 41 (67.2%) received systemic chemotherapy. Mean NCCN DT score was 5.75, ranging from 0–10. 51 patients (83.61%) experienced some level of overall distress. Top concerns on the 30-questions PRD survey were “poor concentration or memory” (2.81), “fatigue” (2.59), “irritability” (2.54), “sleep difficulties” (2.52), and “how I will feel during my treatment” (2.44). Conversely, the least concerning points of distress were “completing a medical power of attorney or living will” (1.44), “questions about end of life” (1.52), “a loved one relying on me for their physical care” (1.57), “transportation” (1.59), and “housing” (1.60). Conclusions: Patients treated for primary CNS tumors in our clinic have a wide array of concerns and varying levels of distress. The NCCN DT showed that the vast majority of patients are dealing with some degree of distress during treatment. The 2 most commonly reported and most severe distress scores among this patient population were poor concentration and fatigue. Distress screens should be used in a multidisciplinary setting to identify possible prophylactic measures and interventions to minimize distress and improve patient quality of life.

AB2016-29. Treatment Outcomes in Patients With Metastatic Neuroendocrine Tumors: Retrospective Analysis of a Community Oncology Database

Maxine D. Fisher, PhD^a; Sonia Pulgar, MPH^b; Matthew H. Kulke, MD^c; Susan Pitman Lowenthal, MD, MPH^b; David Cox, PhD^b; Paul J. Miller, PhD^a; Mark S. Walker, PhD^a; and Lee S. Schwartzberg, MD^d

From ^aVector Oncology, ^bIpsen Biopharmaceuticals, Inc., ^cDana-Farber Cancer Institute, and ^dThe West Clinic

Background: To date, data describing the treatment experience of patients with metastatic neuroendocrine tumors (mNETs) from the community oncology setting have been limited. We recently described the demographic and clinical characteristics and overall treatment patterns of patients diagnosed with mNETs from a retrospective database analysis of community oncology practices. This report provides additional information on the treatment outcomes for this population. Methods: Clinical patient data from medical records of adults with mNETs were collected from the Vector Oncology Data Warehouse, a comprehensive database representing a network of community oncology practices in the United States. Patients treated between November 1996 and May 2015 were included in the analysis. Kaplan-Meier analysis was used to examine overall survival (OS) and progression-free survival (PFS). Results: 263 patients with mNETs were included in the analysis; median duration of follow-up was 22 months (range, 0.1–193.9). 30.4% of patients (80/263) had intestinal tumors, 11.0% had pancreatic tumors (29/263), and 59.7% had other/unknown tumors (157/263). Over the duration of study follow-up, 83.3% of patients (219/263) received ≥1 line of therapy, 53.2% (140/263) received ≥2 lines, 34.6% (91/263) received ≥3 lines, 18.6% (49/263) received ≥4 lines, and 11.4% (30/263) received ≥5 lines. Among patients with well-differentiated (G1) or moderately-differentiated (G2) tumors (n=118), median OS from initiation of first-line therapy was 83.4 months. For those whose tumor grade was undocumented (n=80), median OS from time of diagnosis was 71.5 months. For patients with G1 or G2 tumors, median OS from the time of diagnosis was 102.8 months for those with intestinal tumors and 56.9 months for those with other/unknown tumors. For patients with G1/G2 tumors, median PFS from initiation of therapy for those with intestinal (n=42), pancreatic (n=6), and other/unknown tumors (n=47) was 34.6, 10.4, and 8.1 months, respectively. Median PFS for patients with G1/G2 tumors by regimen-based line of therapy (1–5, respectively) was 12.1 (n=95), 6.7 (n=62), 8.7 (n=41), 5.4 (n=22), and 6.5 (n=12) months. 87.4% of patients (83/95) with G1/G2 tumors in first-line therapy received somatostatin analogue treatment. Conclusions: In this real-world community oncology retrospective study, patients with G1/G2 mNETs experienced a median OS similar to that which has been reported in other US studies. Those with intestinal tumors experienced longer OS rates than those with other/unknown tumors. Among this group, median PFS was longest for first-line therapy and decreased with subsequent regimens.

AB2016-30. Granzyme B Production as a Biomarker for the Immunomodulatory Activity of Sorafenib in Hepatocellular Carcinoma

Suresh Gopi Kalathil, PhD; Austin Miller, PhD; Renuka Iyer, MD; and Yasmin Thanavala, PhD

From Roswell Park Cancer Institute This study was approved and funded by the NCCN Oncology Research Program from general research support provided by Bayer HealthCare Pharmaceuticals, Inc.

Background: We hypothesized that patients with advanced-stage hepatocellular carcinoma (HCC) receiving sorafenib may exhibit a decrease in immunosuppressive cell subsets, which could translate to an increase in the antitumor cytotoxic activity of the patient's effector T cells. In this ongoing study we have addressed the question whether changes in the levels of immunosuppressive cells following sorafenib treatment and a concomitant improvement in immune function measured as an increase in IFN-γ and granzyme B production by cytotoxic T lymphocytes could be a functional biomarker of clinical improvement. This modulation of immunosuppressive cells was measured serially during sorafenib therapy to address our secondary scientific objective, which is to determine if improvement in immune function is sustained. This study has high translational potential from the perspective of identifying a novel immune biomarker of sorafenib efficacy. Methods: Our study is designed to accrue 30 patients with HCC for evaluation. The frequency of CD3⁺ CD4⁺ Foxp3⁺ regulatory T cells (Tregs) and HLADR⁻ CD14⁻ CD11b⁺ CD33⁺ myeloid-derived suppressor cells (MDSCs) was measured in the peripheral blood samples of patients with HCC and collected before initiation of drug treatment (pretreatment) and at 2 times posttreatment (28–35 days and 160 days). Peripheral blood mononuclear cells were stimulated with anti-CD3/CD28 antibody in vitro for 36 hours. T-cell function reflected by effector T-cell proliferation and IFN-γ and granzyme B production was evaluated by measuring Ki67⁺ IFN-γ⁺ CD4⁺ and Ki67⁺ IFN-γ⁺/granzyme B⁺ CD8⁺ T cells using 11 color flow cytometry. The data were analyzed using paired t-test, and P values <.05 were considered as statistically significant. Results: In an interim analysis of 10 patients, we observed a statistically significant reduction in the frequency of Tregs (P=.01) and MDSCs (P=.009) after sorafenib therapy. Concomitantly, a 3-fold increase in the number of proliferating CD4⁺ and CD8⁺ T effector cells, which produced IFN-γ (P=.001, for both cell types) and granzyme B (P=.002), was also observed after sorafenib treatment. This improved T effector cell function was found to be sustained in the analysis conducted 160 days after therapy initiation. Conclusions: Sorafenib treatment can result in reduced immunosuppressive Tregs and MDSC, which might contribute to its efficacy. Given the small number of patients included in this interim analysis, we have not correlated the changes in the magnitude of immune parameters with clinical benefit. These studies suggest that IFN-γ and granzyme B may be viable immune biomarkers for the treatment efficacy of sorafenib in patients with HCC.

AB2016-31. Clinical Values of Serum CA-125 and Human Epididymis Protein 4 in Predicting Optimal Cytoreduction After Neoadjuvant Chemotherapy in Advanced-Stage Ovarian Carcinoma

Shiv Rajan, MCh, MS^a; Sanjeev Misra, MCh, MS, FRCS^b; Vijay Kumar, MCh, MS^a; Sameer Gupta, MCh, MS^a; Naseem Akhtar, MCh, MS^a; Ravi Kant, MS, DNB, FRCS; Jeevan Ram Vishnoi, MCh, MS^a; and Prashant Dontula, MCh, MS^a

From ^aKing George's Medical University, and ^bAll India Institute of Medical Sciences, India

Background: Primary cytoreduction followed by adjuvant chemotherapy is the current standard of care for advanced ovarian cancers. Primary cytoreduction is often difficult in these cases due to its association with increased surgical morbidity. Fewer patients undergo successful optimal cytoreduction in the primary setting, which makes the rationale for a neoadjuvant chemotherapy (NACT) approach in these patients. Although recent trials have shown similar survivals with both approaches, optimal cytoreduction is the key independent prognostic factor in either approach. Currently, serum CA-125 and radiologic imaging is used to assess response after NACT. Serum human epididymis protein 4 (HE4) is a new marker approved to monitor recurrence or progression of ovarian cancer. This study was performed to determine the role of serum CA-125 & and HE4 levels in predicting optimal cytoreduction after NACT in these patients. Methods: This single institution prospective study was comprised of 112 patients with stage III (72 patients; 64.2%) and IV cancers (40 patients; 35.8%). All patients underwent 3–4 cycles of NACT (pacl*taxel and carboplatin) after diagnosis confirmation via cytology or biopsy. Serum CA-125 and HE4 levels and imaging was performed before initiation of NACT and before interval debulking surgery. Chemiluminescent microparticle immunoassay (CMIA) was used to measure HE4. A value of 35 U/mL was used as the CA-125 cutoff for all women. For HE4, values of 70 pmol/L and 140 pmol/L were used as cutoffs for premenopausal and postmenopausal women, respectively. Achievement of cytoreduction (optimal defined ≤1 cm macroscopic residual disease) was correlated with post-NACT CA-125 and HE4 levels obtained before surgery. Results: Among 112 patients, 100 had serous carcinoma and 12 had mucinous carcinoma. A total 68 of 112 (60.7%) underwent optimal cytoreduction. Mean changes of CA-125 and HE4 values before and after NACT were 1,190.12+/−371.10 and 1,140.30+/−379.62, respectively (P=.0001, both). Both the sensitivity and negative predictive value (NPV) of HE4 was 100%. The specificity and positive predictive value (PPV) of HE4 were 90.9%% and 94.4%, respectively. The sensitivity, specificity, PPV, and NPV of CA-125 in predicting optimal cytoreduction were 88.2%, 45.4%, 71.4%, and 71.4%, respectively. Area under curve for HE4 was 95.5% versus 66.8% for CA-125 (P=.0001). Overall accuracy of HE4 and CA-125 to predict optimal cytoreduction was 96.4% versus 71.4%. Among patients whose HE4 levels remained elevated after NACT, no optimal cytoreduction was achieved. Comparatively, approximately 29% of patients with elevated CA-125 levels post-NACT underwent optimal cytoreduction. Conclusions: Our results show that serum biomarkers can be used in predicting optimal cytoreduction after NACT in patients with ovarian cancer. In this setting, HE4 is a better marker compared with CA-125. This study is the first of its kind, and biomarkers should be studied further in the clinical trial setting for predicting optimal cytoreduction after NACT.

AB2016-32. Two-Year Advanced Melanoma Survival in Naïve Subjects Initiated With First-Line Ipilimumab 3mg/kg Therapy: Evidence From a US Multisite Retrospective Chart Review

Sumati Rao, PhD^a; Cynthia P. Macahilig, MA^b; Pete Wolthoff, BS^b; Qing Harshaw, MD, PhD^c; and Alicia Shillington, PhD^c

From ^aBristol-Myers Squibb, ^bMedical Data Analytics, and ^cEpi-Q

Background: Ipilimumab (IPI; Yervoy) is FDA-approved for use in treatment-naïve or pretreated advanced (unresectable or metastatic) melanoma based on prolongation of overall survival (OS) in phase III trials. IPI-approved dosing is 3 mg/kg every 3 weeks for 4 doses. Published interim data in treatment-naïve patients with a minimum of 1 year follow-up indicates an OS of 14.5 months. This ongoing study describes disease characteristics, patterns of care, second-year outcomes, and safety in 273 treatment-naïve patients with advanced melanoma receiving IPI induction. Methods: This is a multicenter, observational cohort study with data derived from medical records. Included in this study are adults with a diagnosis of AJCC-defined unresectable stage III or IV melanoma of all primary types initiated with IPI 3 mg/kg monotherapy in the first line. Patients were enrolled and followed for a minimum of 2 years from initial IPI induction or until death. Excluded criteria included prior systemic treatment for advanced melanoma, those who received IPI for other cancers, and clinical trial and expanded access program participants. Results: Of 273 patients, 64.8% were male and median age at IPI initiation was 63 years (range, 26–91). Primary melanoma sites were cutaneous (88.5%), ocular (4.4%), mucosal (1.8%), anorectal (1.8%), nasal sinuses (0.7%), and not identifiable (2.9%). Most patients were stage M1 at advanced melanoma diagnosis (87.9%). Of these, 63.0% were M1C. Of the total, 12.1% had brain metastases. Of 227 tested before IPI initiation, 44.1% had LDH levels higher than the institutional upper limits of normal. Median number of IPI doses was 4, with 76.3% of patients completing all 4 doses; 5.1% experienced ≥1 dose interruptions during initial induction. Of the total, 9.9% were reinduced within 2 years of initial induction. Of 161 patients surviving 1-year, 42 (26.1%) received other anticancer therapies in year 2. From IPI initiation, median OS was 14.0 months (range, 0.1–41). The 2-year survival rate was 40.3% (95% CI, 34.4–46.4). Median OS was 17.6 months (range, 0.1–41) in those without brain metastases at advanced diagnosis and 2-year survival was 44.6% (95% CI, 38.2–51.1). Of 21 patients with IPI reinduction after the first full year of follow-up, the median number of doses at first reinduction was 4.0 (range, 1–5), and 1 patient received a third reinduction. Of these 21 patients, 38.1% experienced any drug-related adverse event (AE). The majority of AEs were cutaneous (76.2%), and no AEs were > grade 3. No deaths were attributable to drug-related AEs. Conclusions: 2-year OS in treatment-naïve patients remains high with additional patients and accumulated follow-up. IPI is being administered in year 2 at the same dose as at initial induction and is well-tolerated.

Epidemiology/Risks

AB2016-33. Incidence, Clinical Findings, and Survival of Hepatosplenic T-Cell Lymphoma in the United States

Urshila Durani, MD, MPH; Adam C. Bartley, MS; and Ronald S. Go, MD

From Mayo Clinic

Background: Hepatosplenic t-cell lymphoma (HSTCL) is a rare subset of peripheral T-cell lymphoma associated with chronic immunosuppression. <100 cases have been reported in the literature. Using 2 national cancer datasets, we describe its incidence over time, clinical findings, and overall survival (OS). Methods: Incidence rates (case/100,000,000) were calculated using 2000–2012 data from the SEER Program and age-adjusted to the US population at 2000. Patient data were obtained from the National Cancer Data Base (NCDB) and analyzed for demographic patterns and OS (1998–2011). SEER collects cancer incidence and survival data from population-based cancer registries covering approximately 28% of the population. The NCDB is a national database from more than 1,500 US cancer centers capturing approximately 70% of all newly diagnosed cancer cases. Results: The overall incidence of HSTCL calculated from SEER has increased from 1.8 in 2000 to 15.2 in 2012 (P<.01). Blacks have a significantly higher incidence rate (18.9; P<.01) compared with whites (6.1) or Asians/Pacific Islanders (11.4). There were 185 patients with HSTCL registered in the NCDB. Median age at diagnosis was 47 years (range, 30–62) and 69% were males. A majority of patients (62%) received chemotherapy alone and 12% received stem cell transplant. 51% of patients had B-cell lymphoma symptoms at diagnosis. Only 1% of patients had confirmed HIV. Median OS was 8 months. On stratification, median OS was significantly better among women (17 vs 7 months; P<.01) and those who received upfront transplant (33 vs 7 months; P<.01) but was unaffected by age, race, or time period (Table 1). Conclusions: The incidence of HSTCL has increased over time and is more common among men and blacks. Overall prognosis is poor and has not changed in the past decade. Further research is warranted to determine optimal treatment for these patients and to examine causes of disparity in incidence and survival.

AB2016-33: Table 1.

Univariate Cox Regression for Mortality

AB2016-34. Trends in Incidence and Factors Affecting Survival of Cholangiocarcinoma in the United States

Shiva Kumar R. Mukkamalla, MD, MPH; Hussain M. Naseri, MD; Byung M. Kim, MD; Gregory Haidemenos, MD; Steven C. Katz, MD; and Vincent Armenio, MD

From Roger Williams Medical Center/Boston University

Background: Cholangiocarcinoma (CC) includes cancers arising from the intrahepatic and extrahepatic bile ducts. Previous studies examinded SEER data until year 2000 and reported increasing incidence of intrahepatic cholangiocarcinoma (ICC) in the United States, while noticing a decrease in incidence of extrahepatic cholangiocarcinoma (ECC). In this study, we show timeline trends in incidence of CC using SEER data from 1973 to 2009, and we investigate various demographic, clinical, and treatment-related factors affecting survival. Methods: Incidence rates were age-adjusted to the 2000 United States standard population. A retrospective cohort of 29,228 patients with CC diagnosed from 1973 to 2009 was identified from the SEER database. Minimum follow-up time was 5 years. Descriptive statistics were calculated using chi-square test. 5-year overall survival (OS) and disease-specific survival (DSS) characteristics were calculated using Cox proportional hazards model. Results: Incidence (per 100,000) of CC increased by 65% from 1.7 in 1973–1975 to 2.8 in 2011–2012. The incidence of ICC increased dramatically by 350% from 0.2 in 1973–1975 to 0.9 in 2011–2012, although the incidence of ECC increased by only 19%. Men always had a higher incidence compared with women, with a 65% and 53% increase in incidence rates, respectively. Caucasians continued to experience a higher incidence compared with African Americans (AA), but the percent increase was 59% in AA versus 100% in Caucasians. Mean age of the study population was 69.9±13 years. Of the total cohort, 70.67% were Caucasians, 10.67% Hispanics, 7.01% AA, and 11.65% from other ethnicities. Variables including age, race, sex, anatomic site, stage, histologic grade, surgery, radiation, first primary status, insurance, marital status, smoking, language isolation, income, and education were evaluated to assess for any influence on OS and DSS. Patients with ECC were found to experience better DSS outcomes compared with those with ICC (hazard ratio, [HR], 0.885; P<.0001). HRs of all-cause and disease-specific mortality increased significantly with age, stage, grade, and smoking rate (P<.0001). Disease-specific mortality was found to be higher in AA compared with Caucasians (HR, 1.089; P<.005). Men had an increased risk for all-cause (HR, 1.071; P<.0001) and disease-specific mortality (HR, 1.067; P<.0001). The married patients were determined to experience better survival compared with single/divorced/widowed patients (P<0.0001). As expected, patients who did not undergo surgery experienced poor survival outcomes versus those who did (HR, 2.302; P<.0001). Similar outcomes were also noted in the radiation arm (HR, 1.274; P<.0001). Interestingly, patients with CC as the first primary had poor OS (HR, 1.046; P<.05) but better DSS (HR, 0.951; P<.05). Conclusions: There has been a marked increase in the incidence of ICC from 1973 to 2012, thereby contributing largely to the overall increase in incidence of CC. The cause of increase in incidence rates, especially for ICC, is unclear because the etiopathogenesis for most patients with CC remains obscure. This study also demonstrates the influence of various demographic, clinical, and treatment variables affecting the survival of patients with CC.

AB2016-35. Systematic Review of Baroreceptor Failure in Cancer Survivors

Jennifer A. Park, MD^a; Marisol Hernandez, MLS, MA^b; and Shrujal S. Baxi, MD, MPH^b

From ^aSUNY Downstate Medical Center, and ^bMemorial Sloan Kettering Cancer Center

Background: Baroreceptor failure, also known as autonomic failure or baroreceptor dysfunction, can present with volatile blood pressures or hypertensive crisis and less commonly with orthostatic hypotension, orthostatic tachycardia, or malignant vagotonia. Baroreflex failure is a possible long-term complication resulting from cancer treatment due to iatrogenic injuries to the baroreceptor reflex arc from surgery, radiation therapy (RT), or chemotherapy. We completed a systematic review on chronic baroreceptor failure in long-term patients with cancer. Methods: A comprehensive online literature search of publications was conducted through PubMed, Cochrane, Embase, and CINAHL from 1975 to present. Search terms we used in place of baroreceptor toxicity included: bradycardia/orthostatic hypotension, dizzy spells, hypertension, severe tachycardia/orthostatic hypotension, orthostasis, lightheadedness, volatile hypertension, autonomic neuropathy, postural hypotension, orthostasis, and supine hypertension. Results: Of 7,423 references screened, 15 met the inclusion criteria. There were 12 case reports, 2 case series, and 4 prospective studies with 21 individual cases of Hodgkin's lymphoma or head and neck (H&N) cancers that resulted in treatment-related baroreflex failure. Baroreceptor failure followed RT alone (15), RT and surgery (5), chemotherapy and RT, (5), surgery alone (3), and chemotherapy alone (2). There were 4 prospective studies that tested autonomic function in mainly patients with H&N cancers (Table 1). Patients were found to have baroreceptor dysfunction mostly after neck RT, and those who underwent chemotherapy presented with severe orthostatic hypotension consistent with autonomic failure. No prospective studies evaluated management strategies. Conclusions: Chronic baroreceptor failure after cancer treatment is an underreported complication. Its clinically significant prevalence is confirmed through case reports and 5 prospective studies that detected abnormal baroreceptor function in mainly patients with H&N cancers after treatment with RT, chemotherapy, and surgery. Baroreceptor failure was found to occur most often after RT, but can be a late effect of any treatment modality. The diagnosis of baroreceptor failure should be considered when labile hypertension is seen after cancer treatment, especially to the neck.

AB 2016-35: Table 1.

Prospective Studies Identified

Epidemiology/Risks, Outcomes, and Health Services Research

AB2016-38. Feasibility Assessment for Identifying Patients With Smoldering Multiple Myeloma in an Electronic Medical Records Database

Matthew C. Cheung, PharmD^a; Safiya Abouzaid, PharmD, MPH^b; Quanhong Ni, MS^b; Gail Larkins, RN, MBA^b; Elizabeth Dawn Flick, RPh, MPH, ScD^b; and Kejal Parikh, BPharm, MS^b

From ^aRutgers University, Ernest Mario School of Pharmacy, and ^bCelgene Corporation

Background: Smoldering multiple myeloma (SMM) is an asymptomatic clonal plasma cell disorder described as an early precursor to multiple myeloma (MM). Diagnostic criteria include a serum monoclonal protein (M-protein) of >3 g/dL and/or 10%–60% clonal bone marrow plasma cells. In the absence of an ICD-9 code specific to SMM, it may be difficult to identify these patients in electronic medical records (EMRs) for research purposes. The primary objective of this study was to assess the availability of the clinical data necessary to identify patients with SMM in EMRs using diagnostic criteria. The secondary objective was to evaluate differences in baseline and follow-up characteristics between patients with and without a SMM mention (SMM vs no-SMM) in the open-text field to inform the potential development of an identification algorithm. Methods: A retrospective analysis of patients with ≥2 diagnoses of MM between July 2011 and August 2015 was conducted using the Navigating Cancer EMR database. Eligible patients were aged ≥18 years at the index date (defined as first MM diagnosis) and had ≤1 follow-up visit subsequent to the index date. Patients were excluded if they participated in any interventional clinical trials during the study period. Baseline demographics, clinical characteristics, laboratory results, and health care utilization after the index date were assessed for the total study population and compared between cohorts (SMM vs no-SMM). Results: 6,235 patients with a mean age of 69 years were included. Approximately 20% of patients had available laboratory results on serum M-protein and 40% on renal, hepatic, and iron status. In the EMR open-text field, there were 543 patients with a mention of SMM and 5,692 without. A higher percentage of patients with SMM (10.1%) had lower cancer staging (stage 1) versus the no-SMM group (2.7%). ECOG performance status (PS) of those in the SMM group was reported to be numerically lower (ECOG 0, 20.4% vs 10.8%), indicating better PS. Lower health care resource use, including number of visits (51 vs 59) and use of MM treatment within 30 days after index (23.8% vs 44.8%) was also observed in the SMM group compared with the no-SMM group, respectively. Conclusions: Using diagnostic criteria to identify patients with SMM in EMRs is not feasible given the low proportion of populated laboratory fields and the scarcity of data available in the notes. It may be possible to develop an algorithm to identify patients with SMM using baseline and follow-up characteristics identified in this study; however, further data exploration and prospective data validation would be needed.

Outcomes and Health Services Research

AB2016-39. Stereotactic Radiosurgery for Spinal Tumors: A Single Institution Experience

J. Richelcyn M. Baclay, MD; Angela D. Gaerlan, MD; Ibet Marie Y. Sih, MD; Roy Allan Dominique G. Torcuator, MD; Kathleen Jane U. Cortez; Julius Cezar P. Rojales; Nonette A. Cupino, MD; Ma Conchitina M. Mendoza, MD; Juan Martin J. Magsanoc, MD; and Miriam Joy C. Calaguas, MD

From St. Luke's Medical Center, Philippines

Background: Stereotactic radiosurgery (SRS) involves the delivery of a single or a few fractions of high-dose radiation treatment with high degree of accuracy and precision. A high potent biological dose of radiation is delivered to the tumor, with relative sparing of nearby structures through a steep dose gradient outside the treatment volume. A retrospective review was performed to determine the rate of pain relief, toxicity, and neurologic outcomes of patients treated with spine SRS in this institution. Methods: From August 2012 to April 2015, 24 patients who underwent spine SRS in this institution were retrospectively reviewed. Pain outcome was measured using the Numerical Pain Rating Scale. Neurologic examination was performed by the attending neurosurgeon and radiation oncologist. Acute effects were scored according to the Common Toxicity Criteria (CTC), version 3. Mean patient age was 58 ± 16.06 years. Radiation was delivered via intensity-modulated radiation therapy (IMRT) and prescribed to cover ≥80% of the planning target volume (PTV). Patients were treated with a median dose of 16 Gy (range, 12–25Gy), given in 1–5 fractions. Results: Spine SRS was performed in a total of 30 spinal tumors from 24 patients (7 primary spinal tumors, 23 metastatic). The most common origins of the metastatic tumors were prostate (4/23), renal cell (2/23), lung (2/23), and breast (2/23) cancers. The most common treated primary spinal tumor was schwannoma (4/7). A total of 4, 20, 1, and 5 lesions were treated in the cervical, thoracic, lumbosacral, and lumbar spine, respectively. None of the patients received prior irradiation to the spine. Pain was present in 18 patients pre-SRS, with 83% documenting complete pain relief 1 month after SRS (Table 1). 4 of the treated patients only experienced partial pain response. Treatment was well tolerated, with no acute toxicities (ie, nausea, vomiting, and headache reported). 1 patient experienced pain flare 1 day after stereotactic body radiation therapy which resolved with steroids, and 1 patient developed vertebral compression fracture 2 months after treatment. Of the 10 patients who had motor deficits pre-SRS, 3 achieved complete recovery of motor function and 7 achieved partial improvement. Conclusions: SRS for spinal tumors is a well-tolerated, safe, and effective treatment option. Results from our institution appear comparable to other institution reports investigating outcomes of spine SRS.

AB2016-39: Table.

Clinical Outcomes

AB2016-40. Outcomes of Patients With EGFR-Positive Non–Small Cell Lung Cancer Brain Metastases

Suresh Kumar Balasubramanian, MD; Samuel Chao, MD; Alireza Mohammad Mohammadi, MD; Lilyana Angelov, MD; Paul Elson, ScD; and Manmeet Singh Ahluwalia, MD

From Taussig Cancer Institute and Brain Tumor Institute, Cleveland Clinic Foundation

Background: Brain metastasis (BM) is a serious complication of systemic malignancies and is common in patients with non–small cell lung cancer (NSCLC). Activating mutations in EGFR constitute 10%–15% of NSCLC. There is limited information about the outcomes of EGFR-positive patients with NSCLC BM. Methods: With IRB approval, the Cleveland Clinic Brain Tumor and Neuro-Oncology Center's database was used to identify patients treated between 2000 and 2014 for NSCLC BM with activating EGFR mutations. Overall survival (OS) from the diagnosis of EGFR-positive NSCLC BM was the primary end point. Cox proportional hazards model was used for data analysis. Results: Data from 55 patients with EGFR-positive NSCLC BM (60% female) with a median age 60 years (range, 29–82) were analyzed. BM was concurrent with the diagnosis of lung cancer in 55% (30 patients). 76% (42 patients) had extra cranial metastases. 36% (20 patients) had a single BM; 29% (16 patients) had 2–3 BM and 35% (19 patients) had >3 BM. Most patients achieved good performance status (PS) at the time of BM (KPS ≥80 in 78% [43 patients]). Initial treatment of the BM included whole-brain radiation (WBRT) and surgery in 47% (26 patients), stereotactic radiosurgery (SRS) and surgery in 27% (15 patients), SRS + WBRT in 20% (11 patients), and 5% (3 patients) underwent medical therapy only, surgery only, or observation only. 7 patients (13%) received erlotinib before the BM diagnosis, 32 (58%) at BM diagnosis (before progression), 10 (18%) after first progression of BM, and 5 (9%) received erlotinib before and after BM diagnosis. Median OS was 19.9 months (95% CI, 10.2–23.7). PS (P<.0001) and location of disease (P=.02) were independent factors that impacted OS (Table 1). Conclusions: Median OS of 19.9 months in this small subset of patients with EGFR-positive NSCLC BM appears to be better than the traditionally reported OS of 6–12 months in patients with NSCLC BM.

AB2016-40: Table 1.

Multivariable Analysis

AB2016-41. Comparative Effectiveness of Early-Line Nab-Pacl*taxel versus Eribulin in Patients With Hormone Receptor–Positive/HER2-Negative or Triple-Negative Metastatic Breast Cancer: A Community-Based Real-World Analysis

Fadi Braiteh, MD^a,b; Monika Parisi, MPH^c; Quanhong Ni, MS ^c; Si Yeon Park, BS ^d; and Claudio Faria, PharmD, MPH ^c

From ^aUniversity of Nevada School of Medicine, ^bComprehensive Cancer Centers of Nevada, ^cCelgene Corporation, and ^dThe Ohio State University

Background: Chemotherapy-based regimens are recommended for the treatment of hormone receptor–positive (HR+)/HER2-negative (HER2−) and triple-negative (TN) metastatic breast cancer (MBC). This analysis evaluated the comparative effectiveness of nab-pacl*taxel (nab-P) and eribulin in HR+/HER2− or TN MBC in a community-based setting. Methods: A retrospective cohort study was performed using fully deidentified data from a US electronic medical record platform of 1,300 community oncology physicians. Women aged ≥18 years with HR+/HER2− or TN MBC were eligible if nab-P or eribulin monotherapy was used as first-line or second-line treatment from December 4, 2010 to October 6, 2014 (≥1 cycle of nab-P or eribulin required). The primary objective was to examine duration of treatment (including time to treatment discontinuation [TTD]; censored if last dose >30 days from the last date in the database) and time to next treatment (TTNT; day 1 from drug 1 to drug 2). Adverse events (AEs) and supportive care were also examined. Results: nab-P was associated with significantly longer TTD than eribulin in HR+/HER2− MBC (Table 1). Granulocyte colony-stimulating factor (G-CSF) and antiemetics were used less with nab-P in both HR+/HER2− and TN MBC. Steroids were used less with eribulin in patients with HER2− disease. In patients with HR+/HER2− MBC, pain and thrombocytopenia were less common in the nab-P group (1% vs 10%; and 7% vs 17%, respectively), whereas nausea/vomiting and dehydration were less common in the eribulin group (4% vs 14%; and 0% vs 8%). In patients with TN MBC, incidences of AEs were not significantly different between the 2 groups. Conclusions: In HR+/HER2− or TN MBC, nab-P was associated with increased TTD and TTNT compared with eribulin. Moreover, patients in the nab-P group required fewer doses of G-CSF.

AB2016-41: Table 1.

Efficacy and Supportive Care

AB2016-42. Pain and Health-Related Quality of Life in Relapsed/Refractory Multiple Myeloma: Patient-Reported Outcomes from the ELOQUENT-2 Study

David Cella, PhD^a; Jan McKendrick, MSc^b; Amber Kudlac, BA&Sc^b; Abderrahim Oukessou, MD^c; Teresa Zyczynski, PharmD, MBA, MPH^c; and Catherine Davis, PharmD^c

From ^aNorthwestern University, ^bPRMA Consulting Ltd, and ^cBristol-Myers Squibb

Background: Pain can be a significant burden for patients with relapsed/refractory multiple myeloma (RRMM). Elotuzumab is an immunostimulatory monoclonal antibody currently in development for RRMM. In a phase III study (ELOQUENT-2), elotuzumab + lenalidomide and dexamethasone (ELd) demonstrated a 30% risk reduction in disease progression/death versus lenalidomide + dexamethasone (Ld) alone in patients with RRMM (Lonial S, et al. N Engl J Med 2015). Here, we report selected secondary/exploratory study end points assessing pain and health-related quality of life (HRQoL) in ELOQUENT-2. Methods: Patients with RRMM were randomized to ELd or Ld and treated in 28-day cycles until disease progression/unacceptable toxicity. Secondary/exploratory analyses included patient-reported outcome (PRO) assessments; change from baseline in scores for the Brief Pain Inventory-Short Form (BPI-SF; pain interference, pain severity, worst pain), the EORTC quality of life auestionnaire-core 30 (QLQ-C30; fatigue, physical function, global health status, pain), and EORTC QLQ-myeloma module (QLQ-MY20; disease symptoms, treatment side effects). PRO data were collected at screening, on day 21 of each treatment cycle, and after treatment discontinuation. Patients with ≥1 post-baseline PRO assessment were included in analyses. Results: Overall, 646 patients were randomized to ELd (n=321) or Ld (n=325); 319 and 311 patients, respectively, had ≥1 post-baseline PRO assessment and were included in these analyses. Baseline BPI-SF mean pain severity scores were low for both groups (2.6 for ELd; 2.9 for Ld) and remained stable over time. A higher proportion of patients with a clinical response (complete or partial objective response) experienced a reduction in pain score (29% responders, 13% nonresponders). Subgroup analyses revealed a statistically significant and clinically meaningful difference in pain severity in the ELd versus Ld group for patients with initial moderate/severe pain in cycles 1–5. Similar patterns were seen with pain interference and worst pain. EORTC QLQ-C30 scores did not change significantly from baseline in either treatment group, suggesting that the addition of elotuzumab to Ld did not compromise HRQoL with regards to fatigue, physical function, global health status, and pain domains. Baseline EORTC QLQ-MY20 scores were similar between treatment groups and remained stable over time; no difference between treatment groups in disease symptoms or side effects was observed, suggesting that elotuzumab did not compromise HRQoL in these domains. Conclusions: Patients in ELOQUENT-2 generally experienced low levels of pain at baseline, which were maintained during the study duration. The addition of elotuzumab to the Ld regimen did not lead to a decrease in HRQoL in patients with RRMM.

AB2016-43. End-of-Life Costs in Advanced Melanoma

Anna O. D'Souza, BPharm, PhD^a; Tony Okoro, PharmD^b; Manan Shah, PharmD, PhD^a; Sasikiran Nunna, MS^a; and Michael Eaddy, PharmD, PhD^b

From ^aXcenda, LLC, and ^bBristol-Myers Squibb

Background: The survival advantage of new biologic therapies for advanced melanoma has the potential to defer end-of-life (EOL) costs in this patient population. Existing research on EOL costs in advanced melanoma have only evaluated costs in patients aged ≥65 years and included those diagnosed and treated before 2005. The objective of this study was to build on existing literature and estimate the incremental health care costs at EOL for adult patients diagnosed with advanced melanoma (unresectable stage III/IV) using more recent data. Methods: A retrospective matched cohort study was conducted using administrative claims data. Patients with evidence of death from January 2010 to September 2014 were identified and classified into 2 cohorts: patients with advanced melanoma and those without melanoma. Patients with advanced melanoma were matched to those without (1:4) on age at death, gender, year of death, and Charlson comorbidity index (CCI). The year preceding the index (date of death) was divided into a 6-month EOL period (immediately preceding index), and a 6-month baseline before the EOL period. All-cause health care costs and component medical and prescription costs during the EOL period were computed. Additionally, melanoma-related treatment cost components were assessed: drug treatment including biologic therapy, chemotherapy and related administration costs, radiation therapy, and surgery. All costs were statistically compared between study cohorts using paired t-tests at α=0.05. Results: A total of 1,671 patients with advanced melanoma met study criteria, of which 90% (1,504) were matched to 4 controls for a total sample of 7,520. Average age of the matched cohort sample was 62 ± 13 years with 66% male, and an average CCI of 1.9 ± 2.0. During the 6-month EOL period, the advanced melanoma cohort incurred costs 3.4 times higher than the nonmelanoma cohort ($82,010 vs $18,768; P<.001). Almost half the costs in both cohorts were incurred in the last 2 months of life (47% vs 53% in advanced melanoma vs nonmelanoma, respectively). Medical and prescription costs were significantly higher in the advanced melanoma cohort, having a 3.3 and 4.4 relative difference, respectively (P<.001, both). Medical costs comprised the majority of the total costs in both cohorts (80%–94%). For the advanced melanoma cohort, melanoma treatment–related costs comprised 27% of total costs: average treatment costs during the EOL period was $21,875, of which drug treatment was 69%, radiation 30%, and surgery 1%. Conclusions: Advanced melanoma imposes an incremental and significant burden on EOL costs relative to a comparable cohort of patients without cancer.

AB2016-44. Dosing Patterns of Patients With Non–Small Cell Lung Cancer on Epidermal Growth Factor Receptor Tyrosine Kinase Inhibitors in a Commercially Insured Population

Ancilla Fernandes, PhD^a; Elisabeth Croft, MD^a; Yen-Wen Chen, PhD^b; Ozgur Tunceli, PhD^b; and Ralph M. Turner, PhD^b

From ^aAstraZeneca Pharmaceuticals, LP, and ^bHealthCore, Inc.

Background: The benefit of epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) therapies over standard chemotherapy among patients diagnosed with EGFR mutation (EGFRm)–positive non–small cell lung cancer (NSCLC) is well established when the recommended doses are administered. However, there are limited real-world data regarding the sustainability of patients' courses of EGFR-TKI therapy. Methods: This retrospective cohort study used administrative claims data from the HealthCore Integrated Research Environment (HIRE) and HIRE-Oncology databases to identify patients aged ≥18 years with ≥1 ICD-9 code for lung cancer and EGFR-TKI treatment between May 1, 2012 and December 31, 2014. Patients who received NCCN Guideline–concordant chemotherapy for small cell lung cancer (SCLC) were assumed to have SCLC and were excluded from the analysis. The index date was the first observed TKI treatment initiation date (serving as a proxy for first-line TKI treatment), and patients were required to have 12 months of pre-index continuous health plan enrollment, but follow-up time was variable. The analysis described (1) dose at treatment initiation for patients prescribed erlotinib and afatinib, (2) dose adjustments, (3) dose interruptions (≥30-day gap between any 2 fills of TKI, but with fills after the gap during follow-up), and (4) discontinuation (no fill after the days' supply of the index medication + 30 additional days of no index drug prescription fills). Results: 72 patients (average age, 64.4 years [SD, 12.9]; 54% male) treated with erlotinib and <10 patients (aged 55 years [SD, 7]; 75% male) treated with afatinib met study criteria. Thus, erlotinib was the most commonly prescribed first-line EGFR-TKI therapy for NSCLC. Approximately 14% of patients initiated on erlotinib were started on a lower than indicated dose (indicated dose, 150 mg). Of those initiated on erlotinib, 43% experienced dose reductions, and the median time to dose reduction was 48 days (range, 5–310 days); 15% experienced dose interruptions, and the median time to dose interruption was 150 days (range, 31–353 days). Finally, 36% discontinued erlotinib at a median of 118 days (range, 31–518 days). Of the afatinib-treated patients, most (75%) started on the recommended strength (40 mg). Low patient numbers prevented further investigation into dosing patterns for afatinib, but 50% of patients discontinued afatinib at a median of 199 days (range, 150–248 days). Conclusions: Patients initiating EGFR-TKI therapy are not always started on the recommended dose, which may affect treatment outcomes. Future studies should investigate the outcomes of underdosing in this patient population.

AB2016-45. Budget Impact Model to Evaluate Post-Docetaxel Treatment Costs With Varying Cabazitaxel Use in Patients With Metastatic Castration-Resistant Prostate Cancer

Kyle Flannery, PharmD^a; Ed Drea, PharmD^a; Louis Hudspeth, PhD^a; Shelby Corman, PharmD, MS, BCPS^b; Xin Gao, PhD^b; Mei Xue, MBA^b; and Raymond Miao, MS^a

From ^aSanofi, US-HEOR, and ^bPharmerit International

This study was funded by Sanofi.

Background: As new treatment options become available for patients with metastatic castration-resistant prostate cancer (mCRPC), the economic impact becomes more important for health plan decision-makers. Budget impact models (BIMs) are becoming a requirement in many countries as part of formulary approval or reimbursem*nt decisions. Cabazitaxel is approved for patients with mCRPC post-docetaxel. This BIM, using a hypothetical 1 million member US private managed care plan, projected a 1 year budgetary impact of varying cabazitaxel use for mCRPC post-docetaxel. Methods: A BIM was developed to evaluate costs of post-docetaxel treatment in patients with mCRPC. Treatments included cabazitaxel, abiraterone acetate, enzalutamide, and radium-223, with utilization rates derived from market research data. Medication costs were calculated based on published pricing benchmarks factored by dosing and duration of therapy from prescribing information. Grade 3–4 adverse events (AEs) were incorporated using rates from phase III trials multiplied by published cost estimates, per AE, from peer-reviewed sources. 2 scenarios were assessed: (1) base-case–coinsurance and patient out-of-pocket (OOP) costs were subtracted from total treatment costs; and (2) all treatment costs were assumed to be paid by the plan. Results: In a hypothetical 1 million member health plan population, 100 patients were estimated to receive post-docetaxel treatment for mCRPC. Using current utilization rates for the 4 agents of interest, the base-case scenario estimated the cost of post-docetaxel treatment as $6,331,704 or $0.528 per member per month (PMPM). In a hypothetical scenario where cabazitaxel utilization increases from 24% to a hypothetical 33%, the PMPM cost would decrease to $0.524, reflecting a savings of $0.004 PMPM, equating to incremental savings of $49,546 or $497 per patient per year (PPPY). In the second scenario, where OOP costs were not considered, the cost of post-docetaxel treatment was estimated as $6,733,594 or $0.561 PMPM. With increased cabazitaxel utilization (24% to 33%), the PMPM cost would decrease to $0.554, reflecting savings of $0.007 PMPM, equating to incremental savings of $86,136 or $864 PPPY. The primary driver of cost savings with increased cabazitaxel utilization was lower acquisition cost. Sensitivity analyses revealed robust model results over a range of input values (utilization, prevalence, population parameters). Conclusions: In the presented BIM, increased cabazitaxel utilization projects modest savings for both health plans and patients. This BIM presents an objective, comprehensive, and robust, user-adaptable tool that health plans and medical-decision makers may use to evaluate potential economic impact of formulary and reimbursem*nt decisions.

AB2016-46. Real-World Comparative Effectiveness of Albumin-Bound Pacl*taxel Plus Carboplatin Versus Solvent-Based Pacl*taxel Plus Carboplatin in First-Line Treatment of Advanced Squamous Cell Non–Small Cell Lung Cancer in a Community Oncology Setting

Stuart L. Goldberg, MD^a; Harry D. Harper, MD^b; Manish B. Patel, PharmD^c; Claudio Faria, PharmD, MPH^c; Tommy Wu, BS^a; Samira Daswani, BS^a; Eric Schultz, BS^a; and Kelly Choi, MD^a

From ^aCOTA Inc, ^bRegional Cancer Care Associates, John Theurer Cancer Center, and ^cCelgene

Background: In a subset analysis of a multinational phase III randomized clinical trial (RCT), weekly albumin-bound pacl*taxel + carboplatin (nab-P/C) demonstrated a higher response rate (41% vs 24%; P<.01) than solvent-based pacl*taxel + carboplatin (sb-P/C) every 3 weeks, with less grade ≥3 neuropathy and arthralgia, but more thrombocytopenia and anemia in patients with advanced squamous cell carcinoma (SCC) non–small cell lung cancer (NSCLC) (Socinski MA, et al. J Clin Oncol 2012). NCCN Guidelines support the use of nab-P/C, but there is limited published experience with this combination in real-world settings. Methods: A retrospective cohort study was performed using the COTA deidentified database (representing a detailed electronic health record chart abstraction) from 13 cancer centers in New Jersey (64 oncologists). Patients with stage IIIb/IV SCC NSCLC (similar to the RCT) who received either nab-P or sb-P in combination with carboplatin as first-line treatment between October 2012 (availability of commercial nab-P) and April 2014 were analyzed. Results: 184 patients with SCC NSCLC received chemotherapy in the database; 90 had stage IIIb/IV disease, of which 24 received weekly nab-P/C and 37 received sb-P/C (25 weekly sb-P, 12 every-3-weeks sb-P). Median patient ages for nab-P/C and sb-P/C (70 vs 72 years, respectively) and smoking history (100% both groups) were comparable. The nab-P/C cohort had a lower proportion of patients that were males versus the sb-P/C cohort (66% vs 78%). More patients with stage IV disease received nab-P/C compared with sb-P/C (88% vs 60%; P=.019). Consistent with the RCT, the overall response rate (ORR) favored nab-P/C (42% vs 21%; P=.094). Progression-free survival (PFS; median, 12.1 vs 7.3 months; log-rank P=.55) and 12-month overall survival (OS) rates (61% vs 55%; chi-square P=.63) trended in favor of nab-P/C versus sb-P/C. At the time of analysis, median OS was not reached. There was numerically more grade 3/4 anemia (13% vs 5%) and thrombocytopenia (4% vs 3%), but less neutropenia in the nab-P/C cohort (0% vs 5%) consistent with the RCT. Conclusions: Confirming the RCT, patients receiving nab-P/C demonstrated a higher response rate compared with sb-P/C in this small real-world analysis from community cancer centers. PFS and 12-month OS rates also favored nab-P/C; however the difference was not statistically significant. Although RCTs provide evidence to develop NCCN guidelines, because the majority of treatments occur outside research settings, it is imperative to confirm RCT findings using real-world data sets.

AB2016-47. Comparison of Medical Costs and Health Care Resource Utilization of Postmenopausal Women With Hormone Receptor–Positive/HER2–Metastatic Breast Cancer Receiving Everolimus-Based Therapy or Chemotherapy: A US Claims Study

Annie Guerin, MSc^a; Yanni Hao, PhD^b; Nanxin Li, PhD^c; Valerie Koo, BA^c; Anna Fang, BA^c; Miranda Peeples, BA^c; and Andrew Kageleiry, BS^c

From ^aAnalysis Group, Inc., Montreal, ^bNovartis Pharmaceuticals Corporation, and ^cAnalysis Group, Inc., Boston

Background: Treatment guidelines recommend the use of endocrine therapy as first-line therapy for post-menopausal women with hormone receptor–positive (HR+)/HER2-negative (HER2−) metastatic breast cancer (mBC). For patients who fail endocrine therapy, everolimus-based therapy and chemotherapy are commonly used. Given its recent approval, there are limited studies reporting real-world health economic outcomes on everolimus. This retrospective claims analysis compared all-cause, breast cancer (BC)–related, and adverse event (AE)–related medical costs and all-cause health care resource utilization (HRU) among patients with HR+/HER2− mBC who received everolimus-based therapy or chemotherapy. Methods: This study used the MarketScan and PharMetrics claims databases (January 1, 2002–June 30, 2014) to identify postmenopausal women with HR+/HER2− mBC who failed a nonsteroidal aromatase inhibitor and later initiated a new line of therapy for mBC (index therapy/index date). Patients' drug regimens were classified into mutually exclusive treatment groups (ie, everolimus and chemotherapy) and were followed until treatment discontinuation, end of insurance eligibility, or data cut-off (June 30, 2014). All-cause, BC-related, and AE-related medical costs and all-cause HRU including inpatient, ER, outpatient, and other medical services per patient per month were assessed. Adjusted differences in costs and HRU between the everolimus and chemotherapy treatment groups were estimated pooling all treatment lines and using multivariable generalized linear models, accounting for differences in patient characteristics. Results: A total of 3,298 patients who received everolimus (n=902) or chemotherapy (n=2,636) in the first 4 lines of treatment for mBC were included in the study. Compared with chemotherapy, everolimus was associated with significantly lower all-cause (adjusted mean difference, $3,455; P<.01) and BC-related ($2,510; P<.01) total medical costs. Cost differences were driven by lower inpatient ($1,897; P<.01) and outpatient ($1,395; P<.01) service costs. Everolimus was also associated with significantly lower AE-related ($1,730; P<.01) medical costs and significantly lower HRU (inpatient incidence rate ratio [IRR], 0.74; inpatient days IRR, 0.65; outpatient IRR, 0.71; visits to oncologist IRR, 0.66; all P<.01) compared with chemotherapy. Conclusions: This retrospective claims database analysis of patients with HR+/HER2− mBC in the United States showed that everolimus was associated with substantial all-cause, BC-related, and AE-related medical cost savings and less health care resource utilization relative to chemotherapy.

AB2016-48. Treatment Patterns and Factors Associated With Everolimus Use Among Postmenopausal Women With Hormone Receptor–Positive/HER2-Negative Metastatic Breast Cancer

Yanni Hao, PhD^a; Nanxin Li, PhD^b; Miranda Peeples, BA^b; Anna Fang, BA^b; Andrew Kageleiry, BS^b; and Valerie Koo, BA^b

From ^aNovartis Pharmaceuticals Corporation, and ^bAnalysis Group, Inc.

Background: The effectiveness and safety of everolimus for hormone receptor–positive (HR+)/HER2-negative (HER2−) metastatic breast cancer (mBC) has been demonstrated in both clinical trials and real-world practice. This study aimed to assess how everolimus is used in the real world. Methods: Postmenopausal women with HR+/HER2− mBC who were initiated on a new line of therapy for mBC between July 20, 2012 and March 31, 2014 after a nonsteroidal aromatase inhibitor were identified from the MarketScan and PharMetrics databases (January 1, 2002–June 30, 2014) and classified into treatment groups by treatment type and line of therapy. Multivariate logistic regression models were used to identify factors associated with the use of everolimus-based therapy versus endocrine monotherapy and versus chemotherapy, pooling data across lines of therapy (LOT). Dosing patterns for everolimus were summarized by line of therapy. Adherence to everolimus was assessed using the medication possession ratio (MPR), stratified by line of therapy. A multivariate logistic regression model was used to assess the association between adherence to everolimus (MPR >0.8) and line of therapy, adjusting for differences in patient characteristics at treatment initiation. Results: A total of 940 everolimus, 6,134 endocrine monotherapy, and 3,410 chemotherapy regimens were included across the first 4 LOT for mBC. Patients with bone metastasis (odds ratio [OR], 6.27; P<.01) and visceral metastasis (OR, 2.71; P<.01) were more likely to use everolimus than endocrine monotherapy. Patients with a higher comorbidity burden (Charlson comorbidity index; OR, 0.58; P<.05), visceral metastasis (OR, 0.75; P<.01), central nervous system metastasis (OR, 0.55; P<.01), and prior use of chemotherapy for mBC (OR, 0.42; P<.01) were less likely to use everolimus versus chemotherapy. Across LOT, approximately 80% of patients initiated everolimus at the label-recommended dose of 10 mg/d; <5% of patients experienced a dose increase relative to the starting dose, although approximately 20% experienced a dose decrease relative to the starting dose. Across the first 4 LOT, 60%–70% of patients demonstrated high adherence to everolimus (defined as MPR >0.8); adherence to everolimus was not significantly different across LOT. Conclusions: This retrospective claims database analysis found that, in a real-world setting, everolimus is more commonly used than endocrine monotherapy in more severe patients and more commonly used than chemotherapy in less severe patients. The majority of patients used everolimus according to the label-recommended dose and showed high adherence to everolimus across LOT for HR+/HER2− mBC.

AB2016-49. Real-World Treatment Patterns and Characteristics Associated With Systemic Therapy in Patients With Advanced Renal Cell Carcinoma: Results from a US Cohort Chart Review

Robert Hawkins, MD^a; Caitlyn Solem, PhD^b; Justin Doan, MPH, MSc^c; Cynthia Macahilig, MA^d; Jvawnna Bell, MPH, MBA^c; and Shan Jiang, PhD^b

From ^aChristie NHS Foundation Trust, ^bPharmerit International, ^cBristol-Myers Squibb, and ^dMedical Data Analytics

Background: Although guidelines recommend systemic therapy for patients with advanced renal cell carcinoma (aRCC), not all patients receive this treatment. We describe real-world treatment patterns and patient characteristics associated with receiving and discontinuing systemic therapy for aRCC across multiple lines of therapy (LOT) from the US cohort of a global study. Methods: In this global, retrospective, observational medical chart review we identified a random sample of US patients (aged ≥18 years) diagnosed with stage IIIB or IV RCC between January 1, 2010, and December 31, 2012. Patient demographic and clinical characteristics and treatment patterns were extracted using a standardized abstraction form. Student's t-test and Pearson's chi-square test were used to examine differences on continuous and categorical variables between those who did and did not receive ≥2 LOT. Results: A total of 41 US sites contributed data on 307 patients with aRCC; mean age was 64 years, 56% were male, and 4% had a family history of RCC. The majority (82%) were diagnosed with stage IV aRCC, and 9% were staged low-, 53% were intermediate-, and 31% were high-risk (7% unknown). Surgery was more frequent in patients with low- and high-risk disease (42% and 20%, respectively) than in patients with intermediate-risk disease (15%), with 68% of surgeries being a radical nephrectomy. Of the 292 patients (95%) who received 1 LOT, 280 (91%) discontinued and 12 (4%) remained on therapy. Of the 181 patients (59%) who received ≥2 LOT, 171 (56%) discontinued and 10 (3%) remained on therapy. 62% of patients received a tyrosine kinase inhibitor as first-line monotherapy, and 40% received it as second-line monotherapy. 12% of patients received an mTOR inhibitor as first-line monotherapy, and 40% received it as second-line monotherapy. Younger patients (63 vs 66 years) and those with PPO coverage (20% vs 10%) were more likely to receive ≥2 LOT, whereas those with HMO coverage (17% vs 6%) and brain metastases (9% vs 1%) were less likely to receive ≥2 LOT. Of the 171 patients who discontinued treatment in the ≥2 LOT group, reasons included progression (53%), completed planned cycles (15%), “other” (15%), death (9%), patient choice (4%), maximum clinical benefit achieved (4%), and toxicity (3%). Conclusions: >50% of patients who received 2 LOT discontinued therapy because of progression or death, highlighting that better treatment options with more durable clinical benefits are needed for patients with aRCC in the United States.

AB2016-50. Treatment Patterns in Metastatic Melanoma After Introduction of New Drug Therapies

Kenneth Kennedy, PharmD, MS, BCOP^a; Sharanya Murty, PhD^b; Joseph T. Dye, RPh, PhD^c; Tyler Whisman, PharmD, BCOP^d; and Karen Worley, PhD^c

From ^aHumana Pharmacy Solutions, ^bComprehensive Health Insights (at time of study), ^cComprehensive Health Insights; and ^dHumana Pharmacy Solutions (at time of study)

Background: New drug therapies introduced between 2011 and 2013 for metastatic melanoma likely improved outcomes, but limited information exists regarding their incorporation into current treatment regimens. The objective of this study was to characterize initial and concurrent or subsequent cancer therapy for patients receiving ipilimumab (IPI), vemurafenib (VEM), dabrafenib (DAB), or trametinib (TRAM) for unresectable metastatic malignant melanoma (MMM). Methods: The study population was identified as patients with commercial insurance or Medicare coverage from a large national health plan, aged 18–89 years who received their first diagnosis of MMM between January 1, 2011 and December 31, 2013, and had ≥1 claim for IPI, VEM, DAB, or TRAM after index diagnosis for MMM. Initial therapy was defined as the first cancer therapy (surgery, radiation [RT], or chemotherapy) received within 90 days from index diagnosis. Patients were followed until death, disenrollment, or June 30, 2014, whichever came first, to track concurrent and subsequent cancer therapies, and specifically the new agents. Initiation and duration for all cancer treatments were recorded for each patient and results were reported using descriptive statistics. Results: 196 patients with MMM were included, and the most common initial treatments were IPI (n=40; 20.4%), VEM (n=17; 8.7%), and surgery, RT, or chemotherapy (n=83, 42.3%). After 2 years on the market, initial therapy with IPI or VEM was more consistently prescribed closer to time of diagnosis. In 15 patients (37.5%), initial IPI was administered for the full recommended duration of 63 days (4 doses); mean duration of VEM was 227 days. Concurrent melanoma therapies (started after, but continued with, initial therapy) were noted in 46 patients (34.1%); RT (14.1%) and IPI (11.1%) were most commonly prescribed. Subsequent MMM treatment was administered to 74 patients(54.8%); IPI (30.1%), other chemotherapy (16.3%), VEM (14.1%), and RT (14.1%) were prescribed most frequently. Subsequent therapy was initiated <90 days after initial therapy (51%), within 91–180 days (32%), or >6 months (69%). DAB and TRAM had limited use as initial, concurrent, or subsequent therapy in MMM. Conclusions: Use of new cancer drugs in first-line treatment for MMM was low, especially for VEM, DAB, and TRAM. New therapies were more likely to be prescribed in conjunction with traditional cancer treatments or as second-line options. Understanding how new therapies can be incorporated into current treatment regimens and implications for health outcomes are important areas for future research.

AB2016-51. Expectations and Decision-Making in Receiving Chemotherapy for Elderly Patients With Advanced Cancer

Yousef Khelfa, MD^a; Todd Gress, MD^a; Ajeet Gajra, MD^b; and Toni Pacioles, MD^a

From ^aMarshall University School of Medicine, and ^bSUNY Upstate Medical University

Background: Chemotherapy for patients with advanced cancer can prolong life by weeks or months but is not curative. Several factors may guide patients in their decision-making about receiving chemotherapy in the palliative setting, but these are under-studied, There is a gap of knowledge regarding perception and expectations about why elderly patients (≥65 years) with advanced cancer decide to pursue chemotherapy. To study this decision-making process, we used a simple questionnaire. Methods: This study was reviewed and approved by our institution IRB. Patients ≥65 years with advanced (incurable) cancer (solid tumor) were administered a questionnaire after obtaining informed consent. Baseline data including demographic, type of cancer, living situation, performance status (PS), and baseline laboratory work (WBC, hemoglobin, platelet count, albumin, and creatinine levels) were collected. The questionnaire consisted of 16 questions that address patient expectations before initiating a course of chemotherapy. The role of the spouse/family/physician in the patients' decision-making was also explored. The questions address the patient's willingness to accept tradeoffs (eg, chemotherapy-associated toxicity for a 2–3 month prolongation of life). Patients were asked if a palliative care consult/advanced illness consult was offered in addition to or in place of chemotherapy. Results: Between October 15, 2014 and July 25, 2015, a total of 26 patients were interviewed, median age was 75 years, with 96% of patients being white, 46% male, and 54% female (Table 1). Among the 26 patients, 23 said they expect that by receiving chemotherapy, they will live longer and feel more active (88%); 20 said that they expect to feel better physically (77%); and 22 would feel better mentally (84%). 86% were satisfied with their decision to undergo chemotherapy. 62% (16 patients) considered chemotherapy worthwhile if it made them feel better but did not help control the cancer or help them live longer; 69% considered it worthwhile if it stopped cancer growth for 2–3 months but did not help them live longer. When asked about if chemotherapy was worthwhile if it made them really tired but helped them live longer, 77% thought it is worth it; 65% of those who answered yes thought living >1 year more would make it worth it, 10% if living 1–3 months, 15% if living for 4–6 months, and 10% if living 7–12 months. 42% would prefer chemotherapy even if it caused them to require assistance eating, dressing, and toileting. 58% said chemotherapy is worthwhile if it prolonged their life for >1 year, 23% if it made them live 7–12 months longer, 4% if it prolonged their life for 4–6 months, and 15% if they lived for extra 1–3 months. 31% stated that they would withstand the side effects of chemotherapy (eg, nausea, vomiting, diarrhea, or constipation) if it helped them live 2–3 months longer. Only 54% were aware of chemotherapy alternatives, and 31% of patients were offered supportive or hospice care. When asked about their reason for accepting chemotherapy treatment, 17 patients stated it was because their doctor recommended it, 8 said because they were convinced of its benefits, and 5 said that they took it because family or friends recommended it (note: 3 patients had >1 reason). Half of the patients stated that overall functioning and quality of life were more important to them than prolongation of life with chemotherapy. On a multivariate analysis, among the 23 patients (80%) satisfied with their decision to take chemotherapy, 87% finished at least a high school education. Of those who considered chemotherapy worthwhile if it made them feel better but did not help control the cancer or help them live longer, 3, 5, 7, and 1 patients had an ECOG PS of 0, 1, 2, and 3, respectively. All patients who would prefer chemotherapy even if it would cause them to need help eating, dressing, and toileting had an ECOG PS ≤2 (4, 3, and 4 patients with ECOG PS of 0, 1, and 2 respectively). Of the 13 patients who thought that overall functioning and quality of life are more important than prolongation of life with chemotherapy, only 1 patient had received chemotherapy previously, whereas 3 patients of those who disagree had taken chemotherapy in the past. Only 1 of the 8 patients said they would take the side effects of nausea, vomiting, diarrhea, or constipation for chemotherapy if it helped him live 2–3 months longer had previous chemotherapy, 2 of the 8 patients said they would not, and 1 of those said maybe would take it. Conclusion: The majority of patients (17/26; 65%) decided to receive palliative chemotherapy because their oncologist recommended it and only 8/26 (31%) said they were offered hospice option. Physicians have the greatest influence on patient decision-making with regards to cancer treatment. We recommend that a comprehensive geriatric assessment and multidisciplinary discussion be part of the evaluation for elderly patients to help with the decision-making, which includes making patients aware of the hospice option upfront. The majority of patients who enrolled had a functional status ≤2. Previous chemotherapy treatment does not seem to be a factor in the decision-making for palliative chemotherapy, although our total number of patients surveyed is low to make such a conclusion. Of note, all patients who were interviewed had good social support and a caregiver was present at the survey.

AB2016-51: Table 1.

Patient Characteristics (N=26)

AB2016-52. Genetic Testing of a 7-Gene Panel Could Cost-Effectively Aid in Risk Reduction for Women at High Risk of Hereditary Breast Cancer

Yonghong Li, PhD^a; Lance A. Bare, PhD^a; Richard A. Bender, MD^b; Charles M. Strom, MD, PhD^a; and James J. Devlin, PhD^a

From ^aQuest Diagnostics, and ^bSignal Genetics

Background: Women with mutations in the BRCA1, BRCA2, TP53, PTEN, CDH1, STK11, or PALB2 genes have an increased risk of breast cancer. NCCN recommends that women with these gene mutations consider prophylactic surgery and MRI surveillance, which improve life expectancy for women at high risk of breast cancer. The objective of this study was to determine whether genotype results from a panel of these 7 breast cancer–associated genes would cost-effectively aid in risk reduction for women at high risk of hereditary breast cancer compared with genotype results from only BRCA1 and BRCA2. Methods: A decision model was developed to estimate the cost and effectiveness of 2 testing strategies in a hypothetical cohort of 50-year-old women with a family history of breast cancer. In the usual practice strategy, patients were tested for BRCA1 and BRCA2 mutations. In the 7-gene test strategy, patients were tested for mutations in all 7 genes. In both strategies, those who tested positive were assumed to undergo either appropriate prophylactic surgery or enhanced surveillance (annual MRI and mammography), and those who test negative were assumed to undergo routine surveillance (annual mammography). Costs were based on the Medicare clinical laboratory fee schedule or diagnostic laboratory pricing; other model parameters were taken from peer-reviewed literature. The effect of uncertainties of model parameters on the cost-effectiveness outcome was assessed. Results: For each at-risk woman in the base case, the 7-gene test strategy resulted in a gain of 0.0067 life-years and increased costs by $278 compared with the BRCA1/2 test strategy. Therefore, the incremental cost-effectiveness ratio (ICER) for the 7-gene test strategy versus the BRCA1/2 test strategy was $41,394 per life year gained. The ICER was most sensitive to the incremental cost of the 7-gene test over the BRCA1/2 test, cost of breast cancer treatment, risk of breast cancer in those with mutations in the 5 additional genes in the 7-gene test, and proportion of BRCA1/2 noncarriers who tested positive for mutations in these 5 genes. Probabilistic sensitivity analysis suggested that the 7-gene test strategy had a 96.3% chance of costing less than $100,000 per life-year gained. Conclusions: Compared with genetic testing of BRCA1/2 alone, testing 7 breast cancer–associated genes, followed by appropriate risk-reduction management, could cost-effectively improve life expectancy for women at high risk of breast cancer.

AB2016-53. Do Adverse Events Related to Tyrosine Kinase Inhibitor Treatment Among Patients With Chronic Myelogenous Leukemia Represent a Health Care Cost Burden in the Real-World Setting?

Jay Lin, PhD, MBA^a; Dinara Makenbaeva, MD, MBA^b; Robyn Bilmes, PharmD, BCOP^b; Melissa Lingohr-Smith, PhD^a; and Nicola T. Wallis, MRCPath, FFPM^b

From ^aNovosys Health, and ^bBristol-Myers Squibb

Background: Although introduction of BCR-ABL1 tyrosine kinase inhibitors (TKIs) for the treatment of chronic myelogenous leukemia (CML) has markedly improved overall survival, TKI treatment has been found to be associated with a number of clinically important adverse events (AEs). A recent study by Cortes et al (Am J Hematol 2015) used the FDA Adverse Event Reporting System (FAERS) to identify potential AEs associated with imatinib, dasatinib, and nilotinib treatment. The objective of our study was to evaluate the real-world health care costs associated with AEs identified in the FAERS study among TKI-treated patients with CML. Methods: Adult patients with ≥1 inpatient or ≥2 outpatient ICD-9-CM diagnosis codes for CML and ≥1 claim for a TKI treatment from January 1, 2006 to September 30, 2012 were identified from the commercial and Medicare MarketScan research databases. The first claim for a TKI was defined as the index event. Patients were required to have continuous enrollment during 12-month periods before (baseline period) and after (follow-up period) the index event. Health care costs associated with AEs, selected from those identified in the FAERS study and based on the highest probability of association with a TKI, were evaluated during the follow-up period based on ICD-9-CM codes. The evaluated AEs included femoral arterial stenosis, intermittent claudication, peripheral arterial occlusive disease (PAOD), coronary artery stenosis, conjunctival hemorrhage, pleural effusion, malignant pleural effusion, and pericardial effusion. Results: The study population identified from the MarketScan databases included 2,005 patients with CML who received TKI treatment. Mean age of the study population was 56 years, 56% were male, and the majority (59.1%) had a Charlson comorbidity index score of 1–2. Mean total health care costs were highest for femoral arterial stenosis ($17,015 per patient), followed by PAOD ($15,154 per patient), intermittent claudication ($6,112 per patient), and coronary arterial stenosis ($4,944 per patient). Mean total health care costs for fluid retention–related AEs, including pericardial effusion and pleural effusion, were $2,797 and $1,908 per patient, respectively. In this analysis, patients with conjunctival hemorrhage were found to have received treatment in outpatient settings only, with a mean total health care cost of $134 per patient. No patients with malignant pleural effusion were observed in the study population. Conclusions: AEs identified in the FAERS as having the strongest association with TKI treatment are associated with a significant cost burden, with vascular stenosis-related AEs, including femoral arterial stenosis and PAOD, having the highest costs.

AB2016-54. Characteristics and Outcomes of Patients With Advanced Cancer Discharged From a Phase I Trial Inpatient Unit to Hospice Care

Abhishek Maiti, MBBS^a; Holly Kinahan, ANP-C^b; Kenneth Hess, PhD^b; Jennifer Dempsey, PA-C^b; Laura Beatty, PA-C^b; Sarah Baldwin, ANP-C^b; David S. Hong, MD^b; Aung Naing, MD^b; Siqing Fu, MD^b; Apostolia M. Tsimberidou, MD^b; Sarina Piha-Paul, MD^b; Filip Janku, MD^b; Daniel Karp, MD^b; Suresh Reddy, MD^b; Sriram Yennu, MD^b; Eduardo Bruera, MD^b; Funda Meric-Bernstam, MD^b; Gerald Falchook, MD^c; and Vivek Subbiah, MD^b

From ^aThe University of Texas Health Science Center at Houston, ^bThe University of Texas MD Anderson Cancer Center, and ^cSarah Cannon Research Institute at HealthONE

Background: Patients with advanced cancer whose disease progresses on standard therapy are potential candidates for phase I clinical trials. Because of the advanced disease and comorbidities, these patients often need inpatient admission. This study assessed the outcomes of these patients after discharge to hospice care. Methods: We performed a retrospective analysis of patients with solid tumor malignancies discharged to hospice care from the inpatient service of the Clinical Center for Targeted Therapy at The University of Texas MD Anderson Cancer Center. Data were obtained from electronic records and follow-up phone calls. Descriptive statistics, Kaplan-Meier survival analysis, log-rank tests, and univariate Cox proportional hazards regression analysis were used to evaluate overall survival (OS) after discharge to hospice and potential clinical predictors for OS. Results: 133 patients were included in the study cohort (Table 1). All patients had metastatic disease and an ECOG performance status (PS) ≥3. Median survival after discharge to hospice from an inpatient setting was 16 days (interquartile range, 5–29 days). Survival rate was 23% at 1 month, and 5% at 3 months post-discharge. Median survival after the last cancer treatment was 46 days, with survival of 17% at 3 months and 5% at 6 months. Patients with serum LDH >618 IU/L had a shorter median post-discharge survival of 11 versus 20 days for patients with LDH ≤618 IU/L (hazard ratio, 1.48; 95% CI, 1.03–2.13; P=.032). Other studied parameters such as serum albumin, bilirubin, and sites of metastasis did not identify any other at-risk subgroups. A relatively high proportion of our patients, 13% and 34% were enrolled in hospice within the last 3 and 7 days of life, respectively. Conclusions: Patients with metastatic cancer participating in phase I trials who have a poor PS and require inpatient admission have a very short survival after discharge to hospice. A high LDH level predicts an even shorter survival.

AB2016-55: Table 1.

Demographics and Clinical Characteristics (N=133)

AB2016-56. Comparative Effectiveness and Safety of Nab-Pacl*taxel Plus Carboplatin Versus Gemcitabine Plus Carboplatin in First-Line Treatment of Advanced Squamous Cell Non–Small Cell Lung Cancer in a US Community Oncology Setting

From Raja Mudad, MD^a; Manish B. Patel, PharmD^b; Sandra Margunato-Debay, PharmD, MBA^b; Lincy S. Lal, PharmD, PhD^c; and David Garofalo, MS, MBA^c

From ^aUniversity of Miami Miller School of Medicine, Sylvester Comprehensive Cancer Center, ^bCelgene Corporation, and ^cCardinal Health

Background: In a subset analysis of the pivotal phase III multinational randomized clinical trial (Socinski MA, et al. J Clin Oncol 2012), nab pacl*taxel + carboplatin (nab-P/C) significantly improved the primary end point of time to treatment discontinuation (TTD) and overall survival (OS) in patients with squamous cell carcinoma (SCC) non–small cell lung cancer (NSCLC) compared with sb-pacl*taxel + carboplatin (overall response rate, 41% vs 24%; P<.001). In a SCC NSCLC subset analysis of a noninferiority clinical trial (Scagliotti GV, et al. J Clin Oncol 2008), patients receiving gemcitabine (gem)/cisplatin experienced significantly improved survival versus pemetrexed/cisplatin (median OS, 10.8 vs 9.4 months; P=.05); however there are no clinical trials or real-world studies comparing nab-P/C and gem/platinum in SCC NSCLC. This study analyzes the median TTD, median OS, and incidence of adverse events (AEs) among patients treated in the first line setting with 2 NCCN category 1 regimens, nab-P/C versus gem/platinum, for advanced (stage IIIB and IV) SCC NSCLC in community oncology practices. Methods: A retrospective cohort study was performed using fully deidentified data from a nationally representative electronic medical record platform of >975 community oncologists. Patients diagnosed with advanced SCC NSCLC from October 1, 2012 to December 31, 2014; aged ≥18 years; and completed ≥1 cycle of first-line line therapy with either nab-P/C or gem/platinum were included in the study. Median TTD and median OS were compared using the Kaplan-Meier method and log-rank tests. Results: 193 patient records had complete demographic information and met eligibility criteria (nab P/C, n=61; gem + carboplatin [gem/C], n=132). No patients were treated with cisplatin combination therapy. Mean age (68.7 vs 67.9 years; P=.57), male gender (66% vs 61%; P=.51), and Charlson comorbidity index at therapy initiation (2.7 vs 4.1; P=.09) were not statistically different between nab-P/C and gem/C, respectively. Patients treated with nab-P/C had a statistically significant longer median TTD (4.3 vs 3.5 months; P=.03) and median OS (12.8 vs 9.0 months; P=.03) compared with gem/C. Patients receiving nab-P/C experienced fewer hematologic AEs versus those receiving gem/C (Table 1). Conclusions: Patients receiving first-line nab-P/C experienced significantly longer TTD and OS in this real world analysis compared with those receiving gem/C. Additionally, patients receiving nab-P/C had a lower incidence of hematologic AEs versus those receiving gem/C.

AB2016-56: Table 1.

Incidence of Hematologic AEs

AB2016-57. Treatment Patterns and Economic Outcomes in Patients With Relapsed/Refractory Multiple Myeloma Using Pomalidomide or Carfilzomib

Kejal Parikh, BPharm, MS^a; Ali McBride, PharmD, MS, BCP^b; Quanhong Ni, MS^a; Yasir Nagarwala, MD^a; and Safiya Abouzaid, PharmD, MPH^a

From ^aCelgene Corporation, and ^bThe University of Arizona Cancer Center

Background: Pomalyst (pomalidomide [POM]; Celgene Corporation, Summit, NJ) has shown an overall survival advantage and improved health-related quality of life in heavily pretreated patients with relapsed/refractory multiple myeloma (rrMM). The objective of this study was to compare percent persistence, time to treatment discontinuation (TTTD), and healthcare costs associated with POM versus carfilzomib (CAR) during a 1-year period from treatment initiation. Methods: Patients with MM initiating treatment with POM or CAR between October 2009 and September 2014 were identified in a large US claims database. Patients had to be continuously enrolled during baseline (6 months) and study (12 months) periods after treatment initiation, and patients were excluded if they initiated index treatments with an overlap of <90 days. Percent persistence (continuous use of the index drug with an allowable gap <90 days) and TTTD (number of days from index drug initiation to discontinuation date with an allowable gap <90 days) were compared using logistic regression and multivariate Cox proportional hazard models, respectively. Cost differences were estimated using a generalized linear model adjusting for variables that were statistically different at baseline. Results: A total of 1,631 patients initiated either of the index drugs; the final sample included 145 POM-treated and 124 CAR-treated patients. Compared with CAR-treated patients, POM-treated patients were older (mean age, 66 vs 61 years; P<.001), more likely to have Medicare coverage (57% vs 32%; P<.001), and use dexamethasone (57.2% vs 39.5%; P=.003). During the study period, POM-treated patients experienced a significantly higher percent persistence and longer median TTTD compared with CAR-treated patients (Table 1). Compared with CAR-treated patients, POM-treated patients had comparable total adjusted health care costs, significantly higher adjusted pharmacy costs, and significantly lower adjusted chemotherapy-related costs. Conclusions: POM-treated patients had significantly better real-world outcomes compared with CAR-treated patients. Regardless of the treatment initiated, total health care costs were comparable and differences in specific components of health care costs were associated with differences in the place of treatment administration.

AB2016-57: Table 1.

Summary of Results

AB2016-58. Real-World Treatment Patterns in Patients With Chronic Myelogenous Leukemia Newly Initiated on Tyrosine Kinase Inhibitors in an Integrated Health Care System

Nazia Rashid, PharmD, MS^a; Han A. Koh, MD^b; Christian Dimaano, PhD, MPH^c; Kathy J. Lin, MS^a; Brian Stwalley, PharmD^c; and Eugene Felber, PhD^c

From ^aKaiser Permanente Southern California, DIS Research Group, ^bSouthern California Permanente Medical Group, and ^cBristol-Myers Squibb

This study was funded by Bristol-Myers Squibb.

Background: Tyrosine kinase inhibitors (TKIs) have shown efficacy in patients with chronic myeloid leukemia (CML) in the clinical trial setting. However, additional real-world data evaluating TKI treatment patterns within a managed care organization are needed. This study evaluated treatment patterns among newly diagnosed patients with CML initiated on TKI therapy in an integrated health care system. Methods: A retrospective cohort study based on administrative claims and clinical data from Kaiser Permanente Southern California (KPSC) was conducted. Patients were TKI-naïve; had an incident CML diagnosis between January 1, 2007, and December 31, 2013; initiated TKI therapy (index date) between January 1, 2007, and December 31, 2014; were aged ≥18 years on the index date; had no history of cancer or stem cell transplant 12 months before diagnosis; and had continuous enrollment for 12 months before the index date. Patients were followed from their index date until the end of the study period, death, or disenrollment from health plan. TKI therapy was evaluated in terms of persistence and switching. Multivariate regression analyses were conducted to evaluate factors associated with discontinuation and switching. Chart notes were evaluated for reasons that led to discontinuation. Results: 216 patients with incident CML treated with a TKI were identified: 19 (8.8%) on dasatinib, 189 (87.5%) on imatinib, and 8 (3.7%) on nilotinib; the mean age was 53, and 63% were male. 103 patients (48%) continued their index therapy, 62 (29%) switched it, and 51 (24%) discontinued it. Discontinuation of therapy was associated with age, Medicare insurance, Charlson comorbidity index (CCI) score ≥2, and elevated BMI. Higher CCI score, prior hydroxyurea use, and higher use of the ER or hospitalizations were significantly associated with switching therapy. The most common reason leading to discontinuation of therapy was tolerability. 16 patients (31%) who discontinued their index therapy were initiated on another TKI after >60 days. Conclusions: This is the first study to evaluate CML TKI treatment patterns and factors associated with discontinuation and switching within a US integrated managed care organization. Imatinib was the initial TKI therapy for 87.5% of patients. Of those with incident CML initiated on TKIs, 53% either switched or discontinued their index therapy on average <1 year from treatment initiation. The primary reason for discontinuation was related to tolerability. Factors such as age, higher comorbidity burden, increase in BMI, and prior health care use were associated with switching or discontinuation.

AB2016-59. Effects of Vitamin D₃ on Patient-Reported Symptoms in Women on Aromatase Inhibitors

Alice C. Shapiro, PhD, RDN; Rachel Lerner, MD, MS; Sara Richter, MS; and Susan A. Adli, MS

From Frauenshuh Cancer Center, Park Nicollet Institute

Background: Treatment with aromatase inhibitors (AIs) can produce a cluster of symptoms that affects the quality of life in many cancer survivors. The patient perspective in reporting these symptoms is not systematically collected and may provide valuable insight into strategies for symptom management. As part of a larger study, we report secondary data that characterizes the effect of 2 doses of vitamin D₃ supplements on self-reported symptoms in this population. Methods: The larger study was a double-blind clinical trial designed to study the effects of vitamin D₃ on arthralgias in women taking AIs. Patients (n=113) were postmenopausal women, aged ≥18 years, and with stage I–IIIA breast cancer taking AIs. Eligible participants were enrolled and randomized at the screening visit, went through a 4-week wash in/out period on 600 IU vitamin D₃ daily, followed by group assignment with usual care dose of 600 IU or a high dose of 4,000 IU/d per day for 6 months. Here, we report on 10 additional symptom scales from The Breast Cancer Prevention Trial (BCPT) symptom checklist. Data were collected at enrollment, baseline, and 6-month visits. Changes in symptom scales within and between groups were compared using paired or 2-sample t-tests or their nonparametric equivalents, as appropriate. Results: The groups did not differ on demographic or clinical characteristics. A statistically significant change was demonstrated in patient-reported fatigue in the usual dose group from baseline to 6 months (mean, 1.8 vs 1.4; P=.0061). Statistically significant changes in hot flashes were seen in the high-dose group from baseline to 6 months (mean, 1.5 vs 1.2; P=.0034), and between the vitamin D₃ groups for the change from baseline to 6 months (mean change, 0.1 vs −0.3; P=.0232). There were no significant changes in other subscales. Conclusions: Vitamin D₃ supplements did not improve patient-reported symptoms in most of the BCPT subscales. The statistically significant improvements observed in fatigue and hot flashes reflect small improvements that most likely have little clinical impact. Vitamin D₃ does not appear effective in reducing the side effects of AIs, however, the BCPT is a useful tool to incorporate the patient experience and to evaluate potential treatments for AI adverse events in breast cancer survivors.

AB2016-60. Primary Thromboprophylaxis in Ambulatory Patients With Lung Cancer Receiving Chemotherapy: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Kyaw Zin Thein, MD; Patricia A. Brock, MD; Valda D. Page, MPH; and Thein Hlaing Oo, MD, FRCP

From The University of Texas MD Anderson Cancer Center

Background: Lung cancer is the most common cause of cancer mortality in the United States. Venous thromboembolism (VTE) is the second-leading cause of death in patients with cancer; it was hypothesized that outcomes may be improved by preventing VTE. Primary thromboprophylaxis (PTP) in ambulatory patients with solid tumors remain uncertain. We conducted a systematic review and meta-analysis of randomized controlled trials (RCTs) to determine the benefit and risk of PTP with low-molecular-weight heparins (LMWH) in ambulatory patients with locally advanced or metastatic lung cancer receiving chemotherapy. Methods: We performed a comprehensive literature search using MEDLINE, Embase, and the Cochrane Database of Systematic Reviews through September 30, 2015. References of all potential studies were also reviewed for any additional relevant studies. RCTs with a reduction in symptomatic VTE as the primary or secondary end point and of major bleeding as a safety outcome were incorporated in the analysis. The Mantel-Haenszel method was used to calculate the estimated pooled risk ratio (RR), and risk difference (RD) with 95% CIs. Results: 2 RCTs and a subgroup of another 3 RCTs with a total of 2,113 patients with lung cancer were included. Prophylactic doses of dalteparin, nadroparin, certoparin, semuloparin, and bemiparin were used in the studies. The duration of LMWH ranged from 3–6 months. VTE incidence was 28 (2.5%) in LMWH group and 59 (5.9%) in control group (RR, 0.408; 95% CI, 0.263–0.634; P<.001). The absolute RD in VTE was −0.036 (95% CI, −0.054 to −0.019; P<.001) with an estimate of the number needed to treat (NNT) of 29 to prevent 1 symptomatic VTE event. Major bleeding events were reported in 12 patients (2.24%) in the PTP group compared with 7 (1.69%) in control group, according to an analysis of the 3 RCTs. The pooled relative risk for major bleeding was statistically nonsignificant at 1.438 (95% CI, 0.596–3.468; P=.419). Conclusions: Approximately 60 patients were treated with LMWH to prevent 1 symptomatic VTE among all ambulatory unselected patients with cancer on chemotherapy in a previous meta-analysis. In our study, the relative risk reduction was 57.9% with an NNT of 29 to prevent 1 VTE without an increase in major bleeding events. Our study findings are similar to the preliminary results from the FRAGMATIC Study with a relative risk reduction of 40% and NNT of 35 to prevent 1 VTE event. Routine PTP for patients with lung cancer receiving chemotherapy is not recommended and further studies are necessary to define a subset of ambulatory patients with lung cancer receiving chemotherapy who will benefit from PTP.

AB2016-61. Health Care Resource Utilization by Treatment Response Level in Patients With Relapsed Multiple Myeloma: A Delphi Panel Study

Michele Thomas, MS, MBA^a; Yaozhu J. Chen, MPA^b; Ken Bridges, MD^b; Maria Lankford, BS^a; Ze Cong, PhD^b; and Won Chan Lee, PhD^a

From ^aXcenda LLc, and ^bOnyx Pharmaceuticals, Inc.

Background: With advancements in the treatment of relapsed multiple myeloma (RMM) and health care resource constraints, it is important to understand the impact of RMM treatment response on health care resource utilization (HRU). Methods: A 2-round Delphi panel was formed to generate consensus-based HRU estimates in US patients with RMM. 10 hematologists/oncologists meeting the selection criteria (ie, treated ≥6 patients with RMM in the last year; see ≥1 patients with newly diagnosed MM monthly; spend ≥75% of time in direct patient care) were recruited and remained anonymous. For round 1, each clinician was presented with 16 RMM patient types that varied in age (<65; ≥65 years), ECOG performance status (0–1; ≥2), and modified International Myeloma Working Group response levels (complete response [CR], partial response [PR], stable disease [SD], and progressive disease [PD]). For each patient type, physicians were asked to provide annual HRU estimates (in an open-ended format) for service categories as informed by pilot phone interviews with clinicians. Round 2 focused on services likely to have more significant impact on care for patients with RMM: physical examinations, ER visits with/without a resulting inpatient stay (IP), bone marrow biopsies, supportive medications, transfusions, and PET scans. Panelists reviewed blinded individual estimates by patient type and service category from round 1 in a structured format, and were asked to revisit their own estimates, if appropriate. Consensus was defined as a priori when ≥75% agreement (near consensus reached for 60%–74% agreement) on categorical items and ≤20% between 25th–75th percentiles for continuous items. Results: Overall, physicians emphasized the variability of HRU based on patient type and response level. Full consensus was rare and most often cited in CR. Near consensus was more common, typically in a specific patient type, not by response level. However, some generalities can be noted: for the most costly services, those who achieved a CR were projected to have fewer ER visits (range, 0–1) than those who did not, irrespective of IP (ER + IP: PR, 0–2; SD, 1; PD, 1–2; and ER with no IP: PR, SD, and PD, 1–2). Bone marrow biopsy use was higher in those who did not experience a CR (PR and SD, 0–1; PD, 1 vs CR, 0). Patients who did not achieve a CR were expected to require more physical examinations (range, 6–17) per year than those who achieved a CR (mode, 6). Erythropoietin-stimulating agents were unlikely for patients who did not have PD, although transfusions were only for patients with PD. Consistent with NCCN Guidelines, bisphosphonates should be ordered for all patients across response levels. However, antibiotics and granulocyte-colony stimulating factors were likely unnecessary in patients at any response level. The need for PET scans and bone scans/surveys/radiographs could not be adequately determined. Conclusions: This study demonstrates that CR may be associated with less resource use in patients with RMM. Novel RMM treatment options that control the disease more effectively with deeper response may yield cost offsets.

AB2016-62. Home Administration of Omacetaxine: Initial Experience from the SYNCare Program

Jill Van Wart Hood, PhD^a; Stephen Lott, PharmD, MS, CSP^b; Gordon Strachan, PhD^a; Boxiong Tang, MD, PhD^a; and Udo Mueller, MD, PhD^a

From ^aTeva Pharmaceutical Industries, Ltd, and ^bDiplomat Pharmacy

Background: Omacetaxine mepesuccinate (Synribo; Teva Pharmaceutical Industries Ltd., North Wales, PA) was FDA-approved in 2012 for the treatment of adults with chronic- or accelerated-phase chronic myeloid leukemia with resistance and/or intolerance to ≥2 tyrosine kinase inhibitors. In May 2014, home administration was approved for appropriate candidates. The SYNCare program was developed by Teva to support all aspects of home administration, including preparation, distribution, administration with education, training materials, nurse follow-up, and a 24/7 clinician-supported hotline. Before SYNCare, it is estimated that omacetaxine-treated patients received a mean of 29 doses (based on shipment records to physician offices). This analysis describes early SYNCare experience with the first myelosuppressive agent approved for home administration. Methods: Eligible patients are enrolled in SYNCare and supported at no cost to the physician or patient. At enrollment, prescribers provide information about patient demographics, comorbidities, and prior therapies. During treatment, nurses contact patients at every cycle to provide education, assess adherence, and monitor for adverse events (AEs). Data are collected regarding the number of doses administered, duration of treatment, and patient-reported AEs. Data collected during the first 74 weeks (September 2014–January 2016) of the SYNCare program are presented using descriptive statistics. Results: Enrollees are classified by number of omacetaxine doses received: ≥56 doses (group 1; n=21) and <56 doses (group 2; n=44); 2 additional patients were new starts. 3,283 doses of omacetaxine were prepared and delivered to 67 enrollees during the defined period. Overall, patients received a mean of 90.2 doses; in groups 1 and 2, corresponding means were 124.9 and 32.4 doses. Median (range) duration of omacetaxine treatment was 83.1 days (3–511 days) overall and 178.9 (43–511) and 37.4 days (3–193) for groups 1 and 2, respectively. During monthly follow-up calls, patient/caregiver-reported AEs were documented in 49 patients (73.1%). 19 patients reported AEs related to low blood counts (neutropenia, 20.1%; thrombocytopenia, 14.9%; anemia, 10.4%). Other AEs included injection site reactions (13.4%), nausea/vomiting (19.4%), and fatigue (16.4%). Patients and caregivers were counseled regarding AE management, as appropriate. Conclusions: Preliminary data suggest that SYNCare may expand access to omacetaxine, with many patients receiving home administration for an extended duration. No new safety signals were observed. By providing education and live support, SYNCare offers benefits aimed at improving patient care and health outcomes. As enrollment increases, further analyses will be used to inform clinical practice and enhance patient safety.

Preclinical Oncology

AB2016-63. Combined Hepatocellular-Cholangiocarcinoma: A Case Series at a North American Center

Varun Kapur, MD; Mazen E. Iskandar, MD; Rifat Abul Ala Syed Mannan, MD; Steven T. Brower MD; and Neil D. Thiese, MD

From Mount Sinai Beth Israel Medical Center

Background: With the advances in immunohistochemistry (IHC), combined hepatocellular-cholangio-carcinoma (HCC-CC) has emerged as a unique entity among primary liver tumors. However, there is insufficient data on its behavior and treatment approaches, resulting in variability among physicians treating patients with HCC-CC. Methods: All primary liver cancers with HCC elements diagnosed at Beth Israel Medical Center, New York, between 2008 and 2013, were retrospectively reviewed. 8 combined HCC-CC tumors were identified using IHC hepatocellular markers (Hep-Par-1, CD10, pCEA, arginase-1), biliary markers (CK7, CK19), and mucin. Additional immunostaining was performed for liver fatty acid binding protein (LFABP), CA19.9, glypican-3, CD117 (c-kit), epithelial adhesion molecule (EpCAM), beta-catenin, and C-reactive protein (CRP). Charts of the patients were also reviewed to collect socio-demographic, clinical, radiologic, and outcomes data. Results: The overall incidence of HCC-CC was 11.4% (8/70), with a male predominance (7/8) and proportionate distribution among races. 5 patients had liver cirrhosis, including 4 cases of viral hepatitis (Table 1). IHC staining profile of the HCC and CC components are summarized in Table 2 showing pCEA,CD10, HepPar1, and EpCAM as HCC specific markers, whereas CK7, CK19, CA19-9 were more specific to CC. Using different treatment approaches, survival was unexpectedly high where 6/8 patients (75%) were still alive at follow up (mean, 22.4 months; median, 24 months; range, 2–50) (Table 3). Conclusions: HCC-CC is a histopathologically distinct tumor. Larger multicenter data will be needed to better understand its clinical behavior and tailor specific treatment rather than treating it like HCC or CC.

AB2016-63: Table 1.

Clinical Features (N=8)

AB2016-63: Table 2.

IHC Results

AB2016-63: Table 3.

Treatment and Outcomes

Quality Improvement

AB2016-64. Influence of US Stroke Belt Residence on Cardiovascular Disease Risk in Chronic Myelogenous Leukemia

Anna O. D'Souza, BPharm, PhD^a; Dinara Makenbaeva, MD, MBA^b; Eileen Farrelly, MPH^a; Pamela B. Landsman-Blumberg, MPH, DrPH^a; and Bjorn Bolinder, MSc, MBA^b

From ^aXcenda, LLC, and ^bBristol-Myers Squibb

Background: NCCN Guidelines for chronic myelogenous leukemia (CML; version 1.2016) recommend patient comorbidities, including cardiovascular disease (CVD) and its risk factors, be considered when evaluating tyrosine kinase inhibitor (TKI) therapy options. The objective of this study was to assess both the risk for CVD and the influence of residence in the US Stroke Belt (SB) on prevalence of CVD and its risk factors among patients with CML. The SB is an 11-state geographic region with a higher than average incidence of CVD and its risk factors. Methods: A retrospective cohort analysis was conducted using the International Oncology Network database (January 2005–January 2015). Patients aged ≥18 years with a diagnosis of CML were identified with date of first observed diagnosis defined as study index. CVD and its risk factors were identified using diagnoses and/or treatment information from predefined electronic medical record fields and electronic text search of physician progress notes. Prevalence of CVD and its risk factors were estimated at baseline and during a 1-year follow-up period. Annual prevalence rates among SB vs non-SB residents were compared using chi-square tests. Annual prevalence rates for CVD risk factors for patients with CML were age–gender standardized to the 2010 US census population for comparison with published rates for the general US population overall and stratified by SB residence status. Standardized rate ratios and z-tests were used to assess the magnitude and significance of annual rate differences. Results: 1,639 patients with CML were included with a mean age of 59 years (SD, 16), with 56% residing in the SB. At 1-year follow-up, 27% of patients with CML showed evidence of CVD and 72% of those without CVD showed evidence of its risk factors. SB residents with CML had similar or slightly higher prevalence rates of CVD and its risk factors versus non-SB residents with CML. Relative to the general US adult population, annual standardized prevalence rates in the overall CML population were 20%, 30%, and 40% higher for hypertension, diabetes, and obesity, respectively. When stratified by SB residence status, annual standardized prevalence rates among SB residents with CML were 1.5, 2.0, and 1.9 times significantly higher for hypertension, diabetes, and obesity, respectively, relative to the general SB population. Corresponding estimates in non-SB residents with CML were also significantly higher relative to the general non-SB resident population −1.4, 2.1, and 2.0 times, respectively. Conclusions: This study suggests that patients with CML may have both a higher risk and prevalence of CVD relative to the general US adult population that is not explained by SB residence.

AB2016-65. Optimizing Functionality and Efficiency in Oncology Treatment Verification Processes

Carly Duncan, PharmD Candidate; David Veasey, PharmD Candidate; Mohammed Ibrahim, RPh, BCOP, BCPS, BCACP; Steven Roshon, MD; and William Kernan, PharmD, MBA

From Cleveland Clinic, Florida

Background: The process of verifying oncology treatment regimens varies substantially between institutions, and since the transition from paper orders to electronic health records (EHR), even more variability has been introduced into the process. The goal of this study was to identify key components that allow for optimum functionality and efficiency in an oncology treatment verification process. Methods: A 10-question multiple choice/short answer survey was prepared encompassing the key components related to an oncology treatment verification process. The survey was then distributed via e-mail to oncology pharmacists, and the results were collected during a 1-month time period. Results were analyzed to determine how the majority of institutions conduct their oncology order processing and the components the respondents believed most vital to the process. Results: 73 responses were collected between September 16 and October 16, 2015. 47% reported using Beacon Oncology software to process orders, and the remainder used either paper order forms or another EHR program. 52% reported needing double verification for chemotherapy and single verification for nonchemotherapy orders. 48% reported that their order processing workflows progress from physician to nurse to pharmacy. 83% reported using software that does not alert the user if a patient requires prescreening for a new medication. Furthermore, 80% reported that their software does not alert the user when a patient fails to meet treatment parameters. 57% reported needing secondary validation by a physician to create a new protocol. 65% reported that the specific monitoring parameters required for each regimen are included when a protocol is created. 68% reported not having links to national guidelines with each regimen. 74% reported that all regimen-specific premedications are included in each protocol. 65% stated that an alert during the verification process for patients who fail to meet treatment parameters would be the most beneficial component to their system. Conclusions: Based on the responses received, several components were identified as pertinent to an oncology treatment verification process (eg, alerts during the verification process if the patient does not meet treatment parameters, links to national guidelines, requiring secondary physician validation when a new protocol is created, incorporating specific laboratory monitoring parameters for each regimen). The next phase to this study would be to implement these changes into an oncology treatment verification process at an institution and assess the changes in functionality and efficiency it may provide.

AB2016-67. Evaluating Iatrogenic Prescribing in Oncology: Consensual Establishment of a List of Preventable Adverse Drug Events

Guillaume Hébert, MSc^a,b; Florence Netzer, PharmD, PhD^a; Christophe Bardin, PharmD^c; Pierrick Bedouch, PharmD, PhD^c; Muriel Carvalho, PharmD^c; Mikael Daouphars, PharmD, PhD^c; Marie Ferrua^a; Anne-Christine Joly, PharmD^c; Isabelle Madelaine, PharmD^c; Nathalie Pons-Kerjean, PharmD^c; Rémi Varin, PharmD, PhD^c; Etienne Minvielle, MD, PhD^a,b; and Francois Lemare, PharmD, PhD^a,b

From ^aInstitut Gustave Roussy, ^bEcole des Hautes Etudes en Sante’ Publique, and ^cIATRIGGER Work Group, France

Background: IATRIGGER is a multicenter study to evaluate drug-related iatrogenesis in oncology, and is based on a trigger tool method. The first phase was a monocentric feasibility test. The second phase is a multicenter trial, with the first step consisting of building a consensus list of adverse drug events (ADEs) that will be tracked in different hospitals with oncology/hematology patients. The main objective of this study is to establish a consensus list of ADEs in patients with cancer undergoing treatment. The secondary objective is to establish flowcharts for each ADE to evaluate drug-imputability, preventability, and severity of harm. Methods: A list of 26 events was prepared and submitted to experts to establish a consensual ranking of ADEs. During the first round, experts were asked to score each ADE from 1 to 5 in terms of frequency, severity, and preventability, and an overall score was obtained. After round 1, all of the ADEs that have scored above mean were selected; those with a mean + SD greater than the average score were selected for round 2. Experts were then asked to select the 4 ADEs they believe are the most imporant. A literature review of those ADEs was then conducted in Medline, and a search algorithm was constructed for each ADE via Boolean combinations. Flowcharts were drawn to assess drug-causality of the ADE, severity according to the NCI-CTCAE, and its preventability in relation to the implementation of international guidelines such as ASCO, NCCN, and ESMO. 30 experts (oncologists and pharmacists) from 9 different hospitals participated in the study. Results: After 2 rounds of electronic votes (1st round, 30 experts; 2nd round, 26 experts), a final list of 15 ADEs established consensus. The ADEs with the most consensus were neutropenia, nausea/vomiting, and treatment-induced anemia. For each ADE, preventability was defined based on nonapplication of international management guidelines and the analytical flowcharts. For instance, the use of granulocyte-colony stimulating factors (G-CSFs) in febrile neutropenia and the day of G-CSF initiation were observed. Conclusions: The consensus method identified a list of ADEs that experts believe have varying practices and for which benchmarking is susceptible to define the best methods of care and prevention.

AB2016-68. Effectiveness of Colorectal Cancer Screening Messaging Among Individuals Noncompliant With Guidelines

Malhar Jhaveri, MBBS, PhD; Jesse Cambon, MEng; Tristan Cordier, MPH; Laura Scheuer-Sutton, BA; Stephanie Weidenborner, MBA^a; and Gilbert Haugh, MS

From Humana Inc.

Background: Colon cancer is the third leading cause of cancer-related death in the United States. Current guidelines recommend regular colorectal cancer (CRC) screening for adults aged 50–74 years for the early detection of polyps or early-stage cancer when treatment is most effective. Despite these guidelines, CRC screening rates are low and better strategies to improve screening rates are needed. This study was conducted to assess the impact of a mail outreach among Medicare Advantage (MA) enrollees previously noncompliant with CRC screening. Methods: MA enrollees with a large national insurer who were reminded about CRC screening in 2012 and 2013, and remained eligible for but were noncompliant with CRC screening in 2012–2014, were sent reminder letters in September 2014. For those with an attributed primary care physician (PCP), their letter included the PCP name; letters to individuals without an attributed PCP included a business reply card and an offer to send a fecal immunochemical test (FIT) kit to their home. Control groups were randomly selected from the eligible population and not sent a reminder letter. The main outcome was CRC screening rates which included a FIT, sigmoidoscopy, or colonoscopy. Messaged and control groups were compared via a log-rank test with a statistical significance of P<.05. Sample sizes were calculated to detect a 1.75% increase in outcome for each arm by December 31, 2014, with 80% power and 95% confidence using a 10% referent screening rate. Results: 38,675 individuals met the eligibility criteria. PCP's name was available for 20,085 individuals, of which 2,483 (12.4%) were withheld as controls. The remaining 18,590 were eligible for the FIT outreach letters, of which 2,651 (14.3%) controls were withheld. Compared with controls, a 1.3% higher CRC screening rate was observed for those receiving letters with their PCP's name (6.1% vs 4.8%, respectively; P=.014). Those receiving the FIT outreach had a 1.2% increased CRC screening rate (5.6% vs 4.4%; P=.011). Overall, an additional 414 CRC screenings were attributed to this outreach. Conclusions: Results indicate that a mail outreach could significantly increase future CRC screening rates in MA enrollees who were previously noncompliant with CRC screening. More research is needed to further understand barriers to screening and the types of interventions to best overcome those barriers.

AB2016-69. High Adherence to Oral Medications Among Hematology Patients in a Rural Community-Based Hospital Network

Darla Liles, MD; Suzanne Lea, PhD; Tiffanie Moore, MPH; Bryan Dangott, MD; Larrin Stewart, BA; Chelsea Passwater, RN, BSN; and Charles Knupp, MD

From East Carolina University, Brody School of Medicine

Background: In patients with multiple myeloma (MM) and chronic myeloid leukemia (CML), a lack of medication adherence is related to poor therapeutic response. Compliance may be influenced by treatment beliefs, behavior choices, or patient barriers. We sought to understand patient compliance in a predominately African American (AA) patient population in a rural community-based hospital network. Methods: The validated Ask-12 survey (GlaxoSmithKline, Philadelphia, PA) was offered to 42 patients with MM or CML seen in 3 community-based oncology clinics affiliated with an academic medical center during routine visits from March to October 2015. The survey was comprised of 3 sections: inconvenience/forgetfulness, treatment beliefs, and behavior. Patients responded using a Likert scale ranging from “in the last week” (least adherent) to “never” (most adherent). Data based on the NCCN Guidelines were collected for each participant and analyzed in SPSS. Results: Table 1 reports baseline results for 5 behavior questions from 42 patients (30 MM, 12 CML). With 31 AA participants (76%), 23 women and 19 men completed the survey. Limited compliance differences were found by cancer type. Median time since diagnosis among patients with MM and CML were 2 and 2.5 years, respectively. Patients with MM were older than those with CML (median, 63 and 48, respectively). In at least 69% of respondents, behavioral factors were not a barrier to adherence. Cost was the least reported reason for noncompliance. Adherence was lowest when medicine was not physically available, suggesting most intended to take their medicine but were unable to, which was consistent with taking medicine “more or less than prescribed.” Conclusions: High adherence to oral medications was found. Cost was not a barrier. Provider improvements may include presenting strategies to help patients plan to take medications at the prescribed time.

Table 1.

Number (%) of Behavioral Measures for Medication Compliance in 42 MM and CML Patients

AB2016-70. Patterns of Bone Marrow Cytogenetic Testing

Richard Lewis Martin III, MD, MPH; Catherine Leith, MB, BChir; Rodolfo Henrich Lobo, MD; and Ryan Mattison, MD

From University of Wisconsin Hospital and Clinics

Background: NCCN and ASH identify baseline cytogenetic testing as a core quality metric for myelodysplastic/myeloproliferative syndrome (MDS/MPN), acute myeloid leukemia (AML), acute lymphoblastic leukemia, multiple myeloma (MM), chronic myeloid leukemia (CML), and chronic lymphocytic leukemia (CLL). In other settings, however, for routine follow up of bone marrows with an initial normal karyotype, cytogenetic testing provides minimal additional information and adds significant cost to analysis ($500 via Wisconsin State Laboratory of Hygiene). For this reason, the University of Wisconsin Hospital and Clinics (UWHC) implemented a hold feature on cytogenetic orders so that marrows obtained with equivocal need for cytogenetic testing could be canceled at a later date. It remains unknown whether this hold feature has reduced unnecessary cytogenetic testing. It is also unknown how cytogenetic ordering is distributed among hematologic diseases. Evaluating these 2 variables is critical to the development of quality improvement systems within the cytogenetic testing process. Methods: We performed a retrospective analysis of 216 consecutive adult bone marrow biopsies from July 1, 2013 through September 30, 2013. Ordering sheets were abstracted for biopsy indication and hematologic disease. Routine monitoring was defined as patients previously treated with transplant or chemotherapy with a subsequent normal follow-up marrow. We also reviewed previous bone marrow biopsies and cytogenetic testing to assess longer term patterns of cytogenetic ordering. Results: Overall, 26 of 216 marrows (12%) had cytogenetic studies drawn but processing was canceled. Diseases contributing the greatest number of cytogenetic orders were MM (20%), AML (16%), and nonspecific cytopenias (17%). Hematologic diseases with the most frequently drawn and canceled cytogenetic studies were non-Hodgkin lymphoma (12/37), Hodgkin lymphoma (1/3), and posttransplant lymphoproliferative disease (PTLD; 1/3). Regardless of indication, cytogenetic studies were always obtained on MDS/MPN, CLL, and CML marrows. Among biopsy indications, cancelations were highest in PTLD and NHL staging marrows (11/22) and response marrows (7/37), and least frequent in routine monitoring marrows (47/48). Conclusions: While UWHC meets NCCN and ASH quality metric guidelines, we observed a pattern of cytogenetic ordering that was minimally influenced by the hold feature. AML, MM, and routine follow up marrows contributed the greatest number of cytogenetic tests ordered. In our next “plan, do, study, act” cycle, we will develop a fishbone and root cause analysis framework targeting the specific problem of inappropriately high cytogenetic ordering in AML and MM marrows with an initial normal karyotype to identify potential interventions acceptable to pathologists and clinicians.

AB2016-71. A Survey of Outpatient Chemotherapy Dosing Practices Among Obese Patients in a Tertiary Care Cancer Center: Identification of Barriers and Implementation of Full Weight-Based Dosing

Rutika Mehta, MD, MPH; Pallawi Torka, MD; Priyank Patel, MD; Amanda Przespolewski, DO; and Ellis Levine, MD

From Roswell Park Cancer Institute

Background: Obesity is associated with an increased risk of and mortality from a range of cancers; underdosing of chemotherapy in this population is thought to be a contributing factor. Up to 40% of obese patients receive limited chemotherapy doses that are reduced from their actual weight-based dose. However, there is no evidence that toxicity is increased among obese patients receiving actual body weight–based therapy. To improve the quality of patient care at Roswell Park Cancer Institute (RPCI), we analyzed chemotherapy-dosing practices across various clinics and assessed institutional compliance with ASCO 2012 guidelines for appropriate chemotherapy dosing in obese adult patients with cancer. Methods: The FADE model (focus, analyze, develop, execute, and evaluate) was adopted for operations. A 2012 survey at RPCI identified up to two-thirds of clinicians self-report routine adjustment of chemotherapy doses for obesity; this was deemed as the baseline for assessment. The study population was defined as all patients with BMI ≥25 who initiated outpatient chemotherapy at RPCI infusion center from January 1, 2015, through June 31, 2015. Data on patient characteristics, chemotherapy doses, dose reduction (yes/no), and reason for dose reduction were collected and analyzed. Results: 287 patients were included in final analyses, with the following characteristics: median age was 63.7 years (range, 27–89); 59% were male; median BMI was 30 (range, 25–57); BMI distribution was 47.8% with BMI 25–29.9 (overweight), 32.4% with BMI 30–34.9 (mildly obese), 15.5% with BMI 35–39.9 (moderately obese), and 4.3% with BMI ≥40 (severely obese). 60 dose reductions occurred in 44 patients (15.8%). Median reduction in dose was 25% (range, 15%–100%). Most common reasons for dose reduction were cytopenias, neuropathy, poor performance status, renal dysfunction, and advanced age. Dose reductions due to obesity were noted only in 2 patients (0.007%). Patients with gastrointestinal cancers had significantly more dose reductions than other cancer types; conversely, no dose reductions were performed in patients treated for lymphoma. Dose reductions did not differ across gender, race, BMI class, performance status, or cancer stage. Contrary to expectation, patients undergoing palliative chemotherapy underwent no more dose reductions than those receiving curative intent therapy (including adjuvant, neoadjuvant, and curative without surgery settings). Conclusions: Our study shows significant improvement among institutional chemotherapy dosing practices in obese patients and establishes compliance for appropriate weight-based dosing according to ASCO 2012 guidelines. Pharmacy educational initiatives and physician self-learning are likely factors leading to this change.