Abstracts From the NCCN 2019 Annual Conference: Improving the Quality, Effectiveness, and Efficiency of Cancer CareTM

The following abstracts were accepted for presentation at the NCCN 2019 Annual Conference: Improving the Quality, Effectiveness, and Efficiency of Cancer CareTM General Poster Session at the Rosen Shingle Creek resort in Orlando, Florida, on March 21 and 22, 2019. Additional abstracts are available at JNCCN.org.


YIA19-001: Mechanisms of Diagnostic Delay Among Black women With Endometrial Cancer (EC): Results from Qualitative Interviews and a National Analysis of Healthcare Data

Kemi M. Doll, MD, MSa; Sara Khor, MSa; Bridgette Hempsteadb; Julianna G. Alson, MPHa; Liz Kellogg, MPHa; Erika Wolff, PhDa; David Flum, MD, MPHa; Scott Ramsey, MD, PhDc; and Barbara Goff, MDa

aUniversity of Washington, Seattle, WA; bCierra Sisters Inc., Renton, WA; and cFred Hutchinson Cancer Research Center, Seattle, WA

Background: A major contributor to the black-white mortality gap in endometrial cancer (EC) is late stage at diagnosis for black women, which may be driven by delays in diagnosis both prior to and after symptom disclosure. Methods: For phase 1, black women with EC were recruited through oncology clinics and a local cancer support group. In-depth interviews were conducted focused on experiences of menopause, postmenopausal bleeding (PMB), and symptom disclosure, and transcripts coded using directed content analysis. For phase 2, EC cases from 2001–2015 were identified in SEER-Medicare. Location, provider type, and dates of symptom report and diagnosis were defined by claims data. The diagnostic interval was then calculated and step-wise multivariate modeling used to determine factors associated with time to diagnosis. Results: Phase 1 included 11 black women from 4 states (WA, LA, GA, CA), ages 47–70, stages 1––3 at diagnosis, for a total of 147 pages of transcribed interviews. Most were insured, with access to routine medical care. Common themes were a lack of knowledge of normal vs abnormal menopausal symptoms and silencing about bleeding among friends and family. The predominant interpretation of PMB was a resumption of normal menstruation, leading to significant delay in symptom disclosure. Reporting to an MD was largely driven by increased severity of bleeding or the onset of pain. Phase 2 included 3,363 EC cases, with 293 (8%) black women. The median diagnostic interval was 28 days (IQR: 8–110 days). After adjusting for age, region, gynecologic history and other presenting symptoms, provider differences were noted with shorter time for ER MDs (84%; P<.01) and PCPs (16%; P=.05) vs OBGYN. Characterization of bleeding as ‘abnormal’ rather than ‘postmenopausal’ prolonged time to diagnosis by 60% (P<.001). Black race was associated with a 2.4-fold increased diagnostic time interval (P=.017). Step-wise modeling showed that this association was explained by differences in diagnostic work-up: Compared to those with a biopsy within 7 days of presentation, women who had an ultrasound or no procedures had longer intervals (42% and 99%, respectively) to diagnosis (P<.001 for all). Conclusion: Among black women with access to medical care, there are modifiable factors that contribute to delays in diagnosis of EC both prior to and after symptom disclosure to a physician. This is the first study to identify targets for intervention to reduce the mortality rate in this high-risk group.

YIA19-002: AXL-RTK Inhibition Modulates T-Cell Functions and Synergizes With Chimeric Antigen Receptor T-Cell Therapy in B-Cell Malignancies

Saad S. Kenderian, MDa; Reona Sakemura, MD, PhDa; Nan Yang, MDa; Michelle Cox, BSa; Sutapa Sinha, PhDa; Michael Mehrdad Hefazi, MDa; Hansen, BSa; Kendall Schick, BSa,b; Cynthia Forsman, PhDa; Justin Boysen, BSa; Wei Ding, MD, PhDa; Sameer Parikh, MDa; Neil Kay, MDa

aMayo Clinic, Rochester, MN; and bMayo Clinic Graduate School of Biomedical Sciences, Rochester, MN

Despite the remarkable outcomes of CD19-directed chimeric antigen receptor T (CART19) cell therapy in B-cell malignancies, the durable responses in diffuse large B-cell lymphoma are less than 40%, and strategies to enhance this response are desperately needed. Inhibition of AXL RTK with TP0903, a high-affinity AXL inhibitor has been found to induce robust apoptosis of malignant B cells. Here, we aimed to examine the role of AXL RTK inhibition with TP0903 on T-cell function in B-cell malignancies. First, we investigated the influence of TP0903 on CART19 cell phenotype and functions. Here, we used 41BB costimulated, lentiviral-transduced CART cells. AXL inhibition led to polarization of CART cells into a Th1 phenotype when T cells were stimulated with the CD19+ mantle cell lymphoma (MCL) cell line Jeko or with leukemic B cells isolated from patients with chronic lymphocytic leukemia (CLL), in the presence of TP0903 (Fig 1A). Exposure of activated CART cells to TP0903 also resulted in significant downregulation of inhibitory receptors on activated CART cells (Fig 1B), a reduction of conical cytokines known to be associated with the development of cytokine release syndrome (CRS) (Fig 1C). To investigate the effect of AXL RTK inhibition of CART cells with TP0903 in vivo, we established MCL xenografts through the injection of 1.0x106 of Jeko into NSG mice. A week later, mice were treated with either vehicle alone, TP0903 (20 mg/kg/day) alone, 0.5x106 of CART19 alone, or TP0903 (20mg/kg/day)+0.5x106 of CART19. Three weeks after the treatment, mice were rechallenged with 1.0x106 of Jeko. Mice treated with CART19 and TP0903 rejected the tumor challenge while mice previously treated with CART19 alone redeveloped MCL, suggesting that AXL inhibition enhanced CART cell persistence (Fig 1D). Finally, we validated our preclinical findings in correlative analyses of phase 1 clinical trial of TP0903 in patients with solid tumors (NCT02729298). Similar to our findings, there was a significant reduction in Tregs, reduction of inhibitory receptors and polarization to a Th1 phenotype. These findings will be further investigated in a planned phase 1 clinical trial of TP0903 in relapsed/refractory CLL (NCT03572634) In summary, we demonstrated for the first time that AXL inhibition polarizes T cells into a Th1 phenotype, downregulates inhibitory receptors, and synergizes with CART cells in B-cell malignancies. These findings encourage further study of TP0903 as an enhancer of T-cell immunotherapies.

YIA19-003: Size Matters: Larger Size CD19-Positive Extracellular Vesicles in Chronic Lymphocytic Leukemia (CLL) Inhibit Chimeric Antigen Receptor T-Cell Function

Saad S. Kenderian, MDa; Michelle Cox, BSa; Reona Sakemura, MD, PhDa; Fabrice Lucien-Matteoni, PhDa; Nan Yang, MDa; Sutapa Sinha, PhDa; Cynthia Forsman, PhDa; Mehrdad Hefazi, MDa; Kendall Schick, BSa,b; Justin Boysen, BSa; Sameer A. Parikh, MDa; and Neil E. Kay, MDa

aMayo Clinic, Rochester, MN; and bMayo Clinic Graduate School of Biomedical Sciences, Rochester, MN

Background: Unprecedented clinical outcomes were reported after CD19 chimeric antigen receptor T-cell (CART19) therapy. However, the complete response rate in chronic lymphocytic leukemia (CLL) is much lower, at approximately 20%–30%. Several immune defects have been identified in CLL that result from the complex interaction between CLL cells and the microenvironment. This leukemic microenvironment is rich with extracellular vesicles (EVs) secreted by B-CLL cells. Here, we aimed to investigate the role of EVs play in the diminished CART response seen in some CLL patients. Methods: EVs were isolated from peripheral blood of 16 patients with untreated CLL at different Rai stages and risk profile by FISH. Cytometry was used to determine size, number of particles per μl, and CD19 expression. To investigate the impact of EVs on CAR T-cell functions, CART19 cells were stimulated with either CLL B cells or the CD19-positive cell line JEKO and different effector functions were analyzed. Results: Two patterns of EVs in CLL patients were identified; a single versus 2 distinct EV populations, characterized by size (small [EVssmall] and large [EVslarge], Fig 1A). In 25% of patients, EVs were CD19 positive (EVCD19+). CD19 positivity was detected only in patients with the EVslarge (Fig 1B). The EVs concentration, CD19 expression (EVsCD19+ vs EVsCD19-), or the size (EVssmall vs EVslarge) did not correlate with disease stage (early vs advanced Rai stage) or risk profile of CLL (low vs high risk). To investigate our hypothesis that EVs could modulate CART19 function, CART19 cell effector functions were examined in the presence of EVsCD19+, EVsCD19-, EVssmall, or EVslarge. EVs alone were insufficient to stimulate CART19 cells. However when CART19 cells were stimulated with the CD19-positive cell line JEKO, their effector functions were reduced only in the presence of EVsCD19+ but not EVsCD19-. This included a significant reduction in CART-specific killing (Fig 1C) and a reduction in cytokine production. The impairment of CART cell functions was independent of the size of EVs, ie, there was no impairment of CART functions with large or small size EVCD19- in co-culture. Conclusion: We identify CD19-positive large size EVs from patients with CLL and demonstrate that these EVs play a role in the leukemic microenvironment by reducing CAR T-cell activity. Studies are ongoing to define the mechanism(s).

YIA19-004: Collateral Deletion of Glycolysis Genes Generates Selective Vulnerabilities to Inhibitors of Oxidative Phosphorylation

Florian Muller, PhDa; Yu-Hsi Lin, MSa; Nikunj Satani, MDa; and Naima Hammoudi, PhDa

aThe University of Texas MD Anderson Cancer Center, Houston, TX

Targeted therapies attacking specific genetic alterations in tumor cells have been an effective strategy in the treatment of many cancers and underlie the concept of “precision medicine.” Successes in various cancers have resulted almost exclusively from targeting activated oncogenes, but this approach has not been proven effective in patients with glioblastoma. We have taken a radically different approach to molecular therapy by targeting genomic deletions, which heretofore have not been considered to be therapeutically actionable. In a chemical biology drug screen of gliomas with passenger deletions in the glycolytic gene Enolase 1 (ENO1), we found that the most cytotoxic agents were those that inhibited mitochondrial oxidative phosphorylation (OxPhos), including a new drug developed at MD Anderson Cancer Center by the Institute of Applied Cancer Science (IACS), which inhibits OxPhos by binding mitochondrial complex I with nM affinity. Exquisite sensitivity to OxPhos inhibition derives from impaired glycolysis, as ENO1-deleted cells are unable to compensatory upregulate glycolysis in the face of inhibition of OxPhos, and succumb to bioenergetic failure. In addition to ENO1-deletions, we find that passenger deletion of other metabolic genes, including 6-phosphogluconate dehydrogenase (PGD), also exhibit selective sensitivity to OxPhos inhibition by the same underlying biochemical mechanism. The IACS compound was found to penetrate the blood brain-barrier and to eradicate PGD and ENO1-deleted intracranial xenografts. The lead OxPhos inhibitor, IACS-10759, is currently in phase 1 clinical trial for leukemia, and our data provide a strong rationale for advancing this agent to clinical trials in glioblastoma patients.

YIA19-005: Immunotherapy for Prostate Cancer Combining CAR-Engineered T Cells with Targeted Immune Checkpoint Inhibition

Saul J. Priceman, PhDa; Yukiko Yamaguchi, PhDa; Elizabeth Epps, BSa; John Burnett, PhDa; and Stephen J. Forman, MDa

aCity of Hope National Medical Center, Duarte, CA

Repairing defects in antitumor immunity has been a longstanding challenge in prostate cancer, and in recent years cellular immunotherapy has emerged as a promising approach for controlling advanced disease. To date, therapies including tumor vaccine and adoptive T-cell immunotherapy have made remarkable headway in solid cancers. Several validated prostate-specific tumor antigens are available as targets for T-cell therapy, including prostate stem cell antigen (PSCA), which is overexpressed in metastatic disease. We are in late-stage preclinical development of PSCA-specific chimeric antigen receptor (CAR)-engineered T cells with plans to initiate a clinical trial early 2019 for the treatment of metastatic castration-resistant prostate cancer. Immune checkpoint pathways, including the programmed cell death protein-1 (PD-1) and the cytotoxic T lymphocyte-associated protein-4 (CTLA4), have emerged as critical drivers of immunosuppression in solid cancers, by limiting both adaptive antitumor immunity as well as adoptive T-cell therapies. Unfortunately in prostate cancer, checkpoint inhibition has led to underwhelming responses, likely due to the low mutational load and immunologically “cold” tumor microenvironment. We hypothesize that antitumor activity of PSCA-CAR T cells will elicit checkpoint pathways that dampen antitumor immune responses and reduce overall clinical outcomes. Herein, we utilize an shRNA approach to knockdown checkpoint receptors as a rational combinatorial strategy that targets checkpoint pathways to improve overall therapy with PSCA-CAR T cells for metastatic prostate cancer. We have successfully developed a multiple shRNA knockdown approach to simultaneously disrupt 3 pathways that may hamper CAR T-cell activity in the tumor. These CAR T cells with shRNA knockdown of checkpoint receptors will be directly compared with checkpoint pathway inhibitor antibody therapies in xenograft models of prostate cancer, with the hope that next generation CARs will resist this break on the immune system in solid tumors.

YIA19-007: Patient-Derived Tumor Organoids for Drug Discovery of High-Risk Hepatoblastoma

Liqin Zhu, PhDa; Liyuan Li, PhDa; Laura Hamel, PhDa; and Dolores Lopez-Terrada, MD, PhDb

aSt. Jude Children’s Research Hospital, Memphis, TN; and bTexas Children's Hospital, Houston, TX

Hepatoblastoma (HB), the most common liver cancer in children, has become the fastest growing pediatric cancer in the United States. The extent of tumor resection remains the most significant prognostic factor for HB, leaving children with unresectable HB limited treatment options and a largely stagnant survival rate below 40%. Due to the rareness and heterogeneity of this pediatric solid tumor, clinically-relevant, patient-derived models are in urgent need in order to develop more effective therapy for individual patients. Recent studies on many adult cancers have proven patient-derived organoids (PDOs) are faithful in vitro models mimicking disease biology and patient drug response. Via the support from a 2017 NCCN Young Investigator Award, our lab tested HB organoid culture using a large cohort of patient-derived xenografts (PDXs). To date, we have grown HB PDOs from PDXs derived from primary, metastatic, and recurrent tumors at 82% (11/14), 100% (7/7), and 100% (2/2) successful rate, respectively. These HB PDOs displayed histopathology consistent with that of their parental tumors. When we tested the first group of 6 PDO lines to the 10 drugs included in recently opened COG AHEP1531 clinical trial for pediatric liver cancer, we found 2 lines whose parental tumors did not respond to HB standard-of-care cisplatin. They showed the same resistance to cisplatin as well as 2 other platinum-based drugs, carboplatin and oxaliplatin. PDOs derived from 2 matched primary and metastatic tumors showed most similar overall response. These results provided encouraging evidence supporting the notion that HB PDOs mimic patient drug response. Intriguingly, in a subset of cisplatin-resistant HB PDOs, vincristine and gemcitabine showed strong antitumor activities, which we will further validate in vivo using a novel neonatal liver xenograft model developed in our lab. To accurately model the unique early developmental tumor microenvironment (TME) of HB, we have successfully engrafted HB cells into the liver of postnatal day 5–7 mice. We found TME of the neonatal liver, indeed, has a strong and unique metastasis-promoting effect compared to that of the adult liver. With this first PDO effort on pediatric HB and our exciting preliminary data, we are confident that we are approaching our long-term goal of building an integrated “HB PDO Clinic” to enable accurate evaluation of disease risk and therapeutic response to facilitate personalized HB patient care.


BPI19-008: Racial Disparities in Breast Cancer Treatment Based on NCCN Quality Guidelines

Surbhi Agarwal, MD, MPHa; Ruta Rao, MDb; and David Ansell, MD, MPHb

aRush Medical College, Chicago, IL; and bRush University Medical Center, Chicago, IL

Background: NCCN quality measures for breast cancer include (1) radiation therapy administered within 1 year of diagnosis for women under age 70 receiving breast-conserving surgery; (2) chemotherapy considered in 4 months of diagnosis for women under 70 with T1c or stage II/III ER/PR- tumors; (3) endocrine therapy administered within 1 year of diagnosis for women with AJCC T1 or stage II/III ER/PR+ breast cancer. These evidence-based measures promote accountability for providers and allow transparency in quality of care. Black women are less likely than white women to receive these therapies that are associated with a survival benefit. Improving adherence to guidelines can decrease the gap in mortality rates for minority women with breast cancer. Methods: We performed a retrospective chart review on patients with breast cancer between April 2010 and October 2015 at Rush University. Information collected included time of diagnosis, clinical stage, ER/PR status, surgical procedures, radiation, chemotherapy, endocrine therapy, and demographics. Chi-squared analysis was done to compare percent of black versus white women who met each quality guideline. Results: In total, 2,436 women were analyzed, of whom 30.3% were black, 66% were white, and 3.7% were other. Of this cohort, 779 women met inclusion criteria for quality guideline 1, and there was no significant difference between black and white women who did not receive radiation therapy (P=.21; 24.7% vs 20.4%). For quality guideline 2 (n=382), there was also no significant difference between black and white women who did not get chemotherapy within 4 months of diagnosis (P=.32; 36.6% vs 31.4%). However, for quality guideline 3 (n=1,222), there was a statistically significant difference between black and white women who did not get hormone therapy within a year of diagnosis (P=.0008; 36.9% vs 26.1%). Conclusions: Endocrine therapy reduces risk of recurrence and mortality in women with ER/PR positive breast cancer; however, there is a disparity between black versus white women who meet this NCCN quality measure. Further studies are needed to understand the reason for this gap in quality of care so that specific interventions can be implemented to eliminate this disparity.

BPI19-013: Use of Clinical Decision Support and Peer Review to Increase NCCN Guidelines Adherence

Eric Gratias, MD, FAAPa; David Spangler, BAa; and Margaret Rausa, PharmDa

aeviCore Healthcare, Bluffton, SC

Background: eviCore healthcare uses the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) to support its proprietary program for medical oncology drug management. All treatment regimens assigned NCCN Category of Evidence 1, 2A, or 2B are considered NCCN-adherent treatment selections in the eviCore program. The purpose of this study was to evaluate the pattern of NCCN adherence during the first year following program launch in regional payer markets. Methods: All cancer drug treatment authorization requests submitted in month 1 and month 12 following program launch for 4 regional third-party payers representing 13 different states were included, each of whom had management of high cost oncology drugs in place prior to eviCore program launch. Month 1 data were used as a surrogate for pre-program NCCN adherence, which is an overestimate as there is significant eviCore program impact on patients initiating therapy during that time. Requests with incomplete clinical data were excluded from analysis. Included requests were stratified by month 1 or month 12 from initial program launch date for each health plan. NCCN adherence was assigned based on the results of the clinical decision support and peer consultation processes utilized by eviCore to adjudicate the treatment request. NCCN adherence rate was calculated for each subgroup and a cumulative NCCN adherence rate for all included cases was calculated using weighted average accounting for volume differences by market. Results: There were 2,028 treatment regimen requests that were fully evaluable, with 1,285 occurring in month 1 and 743 occurring in month 12 following program launch. The rate of NCCN adherence increased for each health plan during the first program year, ranging from 69%–84% in month 1 and rising to 79%–91% in month 12. The weighted cumulative NCCN adherence during month 1 for all included plans was 75% and rose to 88% at month 12 following program launch. Conclusions: Use of clinical decision support supplemented by peer consultation is an effective means of increasing oncologists’ adherence to NCCN-recommended therapies across a broad range of regional provider markets. Additional study is warranted to determine whether this methodology can be applied to NCCN Categories of Preference to direct more patients toward preferred regimens with superior efficacy, safety, and affordability to further improve quality of care and lower total medical costs.

BPI19-018: Cancer Screening Rates and Adherence of Uninsured Cancer Patients in Free Clinics

Katherine Robinson, MDa; Amber Todd, MDa; Abu-Sayeef Mirza, MD, MPHa; Madeline Macdonald, BSb; Noura Ayoubi, BSb; Rahul Mhaskar, MPH, PhDb; Richard Roetzheim, MDa; Laurie Woodard, MDa; and Smitha Pabbathi, MDc

aUniversity of South Florida, Tampa, FL; bMorsani College of Medicine, Tampa, FL; and cMoffitt Cancer Center, Tampa, FL

Background: There are limited studies documenting the prevalence of malignancies and the cancer screening practices of the uninsured population. Cancer survivors require continued cancer surveillance and screening for recurrence and second primaries. However, screening may be suboptimal among the uninsured. Our objective was to identify and document the screening rates and adherence to ACS guidelines in our local uninsured community. Methods: Demographic data, cancer history and associated cancer screening measures were extracted from electronic medical records of patients managed in 8 free clinics between January 2016 and December 2017 in the Tampa Bay Area. Frequencies, proportions, and Pearson correlation coefficients were used to describe the population and statistically significant relationships. Using the ACS cancer screening recommendations, the charts were reviewed for appropriate cancer screening. Results: From manual chart review, 6,958 charts were reviewed and 201 (2.89%) patients had a diagnosis of cancer. The average age was 55.58 years and 134 (66.67%) were women. Most common malignancies included breast cancer (49, 24.38%), prostate (18, 8.96%), colorectal (13, 6.47%), leukemia/lymphoma (11, 5.47%), cervical (10, 4.98%), melanoma (10, 4.98%), ovarian (9, 4.48%), thyroid (9, 4.48%), renal (6, 2.99%), bladder (5, 2.49%), and uterine (5, 2.49%). Of the 201 patients diagnosed with cancer, 104 (51.74%) met the guidelines for a screening mammogram; however, only 49 (47.12%) had this completed. 115 (57.21%) met the guidelines for a screening Papanicolaou smear; 28 (24.35%) had it completed. 145 (72.14%) met the guidelines for a screening colonoscopy; 23 (15.86%) had it completed. 39 (19.4%) met the guidelines for prostate screening; 3 (7.69%) had it completed. Of the 201 patients, 14 (6.97%) reported a greater than 30 pack smoking history but no patients were screened with a low-dose CT of the thorax. Of the 10 patients with melanoma, 3 (30%) mentioned having routine skin exams. Conclusions: The uninsured population have many barriers to obtaining health care and appropriate screening for malignancies. This retrospective chart review highlights the need for easier access to screening. Increasing screening rates in the uninsured population will decrease cancer mortality as well as being cost effective to the community. It is important for free clinic providers to emphasize guideline-directed cancer screening at every visit.


CLO19-037: Reducing the Duration, Incidence and Severity of Mucosal Injury Due to Cancer Radiation Therapy (RT); Positive Randomized Phase 2b Trial Results With GC4419 (Avasopasem Manganese), a Small Molecule Superoxide (SO) Dismutase (SOD) Mimetic

Jon T. Holmlund, MDa; Carryn M. Anderson, MDb; Stephen T. Sonis, DMD, PhDc; Robert Beardsley, PhDa; Dennis Riley, PhDa; Jeffrey Mark Brill, BAa; Melissa Brookes, BAa; Kara Terry, BAa; and J. Mel Sorensen, MDa

aGalera Therapeutics, Inc., Malvern, PA; bUniversity of Iowa Hospitals and Clinics, Iowa City, IA; and cBioModels, Boston, MA

Introduction: RT-induced SO contributes to initiation of mucosal injury; eg, oral mucositis (OM) and esophagitis. GC4419 specifically mimics SOD’s dismutation of SO to hydrogen peroxide (H2O2), interdicting OM initiation. GC4419 reduced RT-severe OM (SOM) in a hamster cheek pouch model, and protected mucosa and other normal tissues from radiation-induced injury in other animal models. In a published phase 1b/2a open-label trial (Anderson et al, IJROBP, 1 Feb 2018), GC4419 attenuated SOM in patients (Pts) receiving intensity-modulated RT (IMRT) plus concurrent cisplatin (CDDP) for locally advanced head & neck cancer (HNC). Objectives: Determine whether GC4419 reduces duration, incidence, & severity of SOM. Methods: Pts with locally advanced oral cavity or oropharyngeal cancer; definitive or postoperative intensity-modulated (IM)RT (approximately 70 Gy [>50 Gy to > 2 oral sites]) plus CDDP (weekly or q3wk) were randomized (stratification: tumor HPV status, CDDP schedule) to 30 or 90 mg of GC4419, or placebo (PBO), 60-minute IV infusion, M–F, ending <60 minutes before IMRT delivered in 35 fractions over 7 weeks. WHO grade OM was assessed by trained evaluators biw during IMRT & qwk for up to 8 wks after IMRT. Primary endpoint: duration of SOM. Efficacy was tested for each active dose vs PBO (ITT population) by a sequential, conditional approach (2-sided alpha, 0.05). Results: 223 pts (44 sites): 90 mg (n=76), 30 mg (n=73), or PBO (n=74). Baseline patient and tumor characteristics and treatment delivery were balanced. Efficacy: At 90 mg GC4419 vs PBO, duration of SOM was significantly reduced (median, 1.5 vs 19 d; P=.024). SOM incidence (43% vs 65%; P=.009), and grade 4 incidence (16% vs 30%; P=.045) also improved. There were intermediate improvements with 30 mg. Safety was comparable across arms; no significant GC4419-specific toxicity; other known toxicities of IMRT/CDDP were not increased. Conclusions: GC4419 demonstrated a significant, clinically meaningful reduction of SOM duration, and dose-dependent improvements in other SOM parameters, with acceptable safety. A confirmatory phase 3 trial (NCT03689712) is in progress. Clinical trials to reduce RT-related esophagitis are also planned.

CLO19-040: The Role of Adjuvant Therapy in Patients With Pathological T2N0 Resected Gastric Adenocarcinoma

John Khoury, MDa; David Macari, MDa; Daniel Ezekwudo, MD, PHDa; Ayoda Weredeb; and Ishmael A. Jaiyesimi, DO, MSa

aOakland University William Beaumont School of Medicine, Royal Oak, MI; and bOakland University William Beaumont School of Medicine, Rochester, MI

Background: There is controversy surrounding the benefit of adjuvant therapy for patients with pT2N0, stage IB gastric adenocarcinoma following surgical resection. Methods: Patients with T2N0 gastric adenocarcinoma (tumor invasion into the muscularis propria) who underwent surgical resection with pathological evaluation of at least 15 lymph nodes were identified from the Surveillance Epidemiology and End Results Registry (SEER) database. Demographics, adjuvant therapy, and survival data were collected and analyzed using SPSS statistical software. Results: A total of 452 cases were identified between 2004 and 2014. Median age at diagnosis was 69. 60.2% of the patients were white, 27.7% Asian, 10.8% black, and 1.3% from other races. Adjuvant therapy was administered to 30.5% of the patients, of which 44.2% received chemoradiation, 48% chemotherapy only, and 7.2% radiation therapy only. After a median follow up of 39 months, the median overall survival (OS) was not reached in the group of patients who received adjuvant therapy versus 100 months in the group that did not receive adjuvant therapy (P=.005). The 5-year OS rate (5-YOS) was 77% for the adjuvant therapy group versus 62% for those who did not receive adjuvant therapy. Univariate analysis revealed that the hazard ratio for death [adjuvant therapy vs observation] was 0.54; 95% CI, 0.35 to 0.83. Adjuvant therapy showed statistically significant survival benefit in patients younger than 60 years of age (5-YOS, 95% vs 79%) and failed to show survival benefit in patients older than 60 (5-YOS, 63% vs 58%). Multivariate analysis revealed that age was associated with mortality, whereas sex, race, grade, tumor size, and number of lymph nodes examined were not associated with increased mortality. Conclusions: Adjuvant therapy provided survival benefit for pT2N0, stage IB resected gastric adenocarcinoma. Our results suggest that patients younger than 60 year of age may benefit the most from this therapy.

CLO19-060: Blood Exosomal Long RNA Profiling Identifies Diagnostic and Prognostic Markers in Pancreatic Ductal Adenocarcinoma

Shulin Yu, PhDa; Peng Wang, PhDa; and Zhen Chen, PhDa

aFudan University Shanghai Cancer Center, Shanghai, China

Background: Long RNAs have been recently identified in human blood exosomes, posing clinical implications. Whether exosomal long RNAs (exoLRs) could constitute key future biomarkers for noninvasive diagnosis, therapeutic evaluation, and prognosis in cancer remains unknown. The study aimed to explore the exoLR landscape of human blood exosomes and evaluate the feasibility of developing a diagnostic or prognostic signature for early detection and prognostic prediction of pancreatic ductal adenocarcinoma (PDAC) based on exoLR profiling. Methods: A case-control study of 267 cases including 137 patients with PDAC and 39 with chronic pancreatitis (CP) plus 91 blood donors as healthy participants was conducted. The exoLR profile of pretreated blood samples was analyzed by exoLR-sequencing (exoLR-seq). Results: An average of 15,000 exoLRs were reliably detected for each sample through exoLR-seq, and 1,053 exoLRs were differentially expressed in PDAC. Based on these data, we constructed a diagnostic signature (d-signature) that showed high accuracy with an area under the curve (AUC) of 0.977 (95% CI: 0.958–0.996), a sensitivity of 92.42% (95% CI: 83.2%–97.49%), and a specificity of 95% (95% CI: 87.69%–98.62%) in a training cohort (n=146), which was further confirmed in a validation cohort (n=93). Notably, the combination of d-signature and CA19-9 yielded an AUC of 0.963 (95% CI: 0.909–1.017), with a sensitivity of 98.13% (95% CI: 93.41%–99.77%) and specificity of 94.59% (95% CI: 81.81%–99.34%). Additionally, we constructed a prognostic prediction model (exoLR p-signature) that effectively predicted prognosis and survival in patients with PDAC (P=9.838e-08). Conclusions: This study clearly demonstrated the value of exoLR profiling in cancer marker discovery and the feasibility of developing a diagnostic or prognostic signature for early detection and prognostic prediction of PDAC.

CLO19-60 Figure 1
CLO19-60 Figure 1

ExoLR profiling in the diagnosis of PDAC. (A–B) ROC for performance of exoLR d-signature in the training (A) and validation cohorts (B). (C–D) Unsupervised hierarchical clustering of 8 exoLRs selected for use in the d-signature in the training (C) and validation cohorts (D). (E) ExoLR d-signature in healthy, chronic pancreatitis (CP), and pancreatic ductal adenocarcinoma (PDAC) subjects. (F) ExoLR d-signature in PDAC patients in stage I–IV. (G) ROC for performance of exoLR d-signature compared with that of CA19-9 in the differential diagnosis of PDAC and CP. (H) DCA to compare the net benefit of combined exoLR d-signature and CA19-9 (red line) with that of CA19-9 alone (blue line) for PDAC vs CP. AUC, area under the curve. CI, confidence interval.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 17, 5.5; 10.6004/jnccn.2019.5002

CLO19-60 Figure 2
CLO19-60 Figure 2

ExoLRs in the prognostic prediction of PDAC. (A–B) Kaplan–Meier curves of overall survival for PDAC patients with low or high risk, according to the prognostic signature (P-signature) in the training (A) and validation cohorts (B). (C) ROC for the p-signature, tumor stage, and p-signature combined with tumor stage in the whole PDAC cohort. (D) Kaplan–Meier survival curves of PDAC patients with combinations of exoLR p-signature and tumor stage in the whole PDAC cohort. AUC, area under the curve. CI, confidence interval.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 17, 5.5; 10.6004/jnccn.2019.5002


HSR19-077: Primary Tumor Location (PTL) and Survival Outcomes in a Real World Cohort of KRAS Wild-Type (WT) Metastatic Colorectal Cancer (mCRC) Patients in the United States

Himani Aggarwal, MPhil, PhDa; Kristin M. Sheffield, PhDa; Li Li, PhDa; David Lenis, PhD, MSb; Rachael Sorg, MPHb; and Rebecca Miksad, MD, MPHb

aEli Lilly and Company, Indianapolis, IN; and bFlatiron Health, New York, NY

Background: PTL is a prognostic factor for mCRC. Recent data suggest PTL is also predictive of survival benefit with cetuximab (CET) and bevacizumab (BEV). This study evaluated the prognostic and predictive effect of PTL in patients with KRAS WT mCRC who initiated first-line (1L) therapy with CET vs BEV in the real world. Methods: This retrospective study selected patients with KRAS WT mCRC who initiated 1L therapy with CET or BEV + FOLFIRI or FOLFOX between January 2013 and April 2017 from Flatiron Health’s electronic health record-derived database. PTL was abstracted from patients’ charts. Left-sided PTL (LPTL): splenic flexure to rectum; right-sided PTL (RPTL): cecum to splenic flexure. Propensity score matching was used to balance treatment cohorts on baseline characteristics. Kaplan Meier and Cox regression methods were used for survival analyses. Results: 1,312 patients met the selection criteria. Of 248 CET + FOLFIRI or FOLFOX patients, 164 had LPTL and 84 had RPTL; of 1,064 BEV + FOLFIRI or FOLFOX patients, 679 had LPTL and 385 had RPTL. CET LPTL and RPTL patients were more likely to receive FOLFIRI vs BEV patients (LPTL: 64.0% vs 24.3%; P<.001; RPTL: 76.2% vs 24.9%; P<.001). Stage at initial diagnosis was different between CET RPTL vs BEV RPTL patients (P<.001). CET RPTL patients were more likely to be stage III (44.0% vs 22.6%) while BEV RPTL patients were more likely to be stage IV (48.8% vs 65.7%) at initial diagnosis. CET RPTL patients were more likely to have a history of adjuvant chemotherapy vs BEV RPTL patients (47.6% vs 22.3%; P<.001). In the matched sample, median overall survival (OS) was 29.7 months (95% CI, 26.9–35.2) for LPTL patients vs 18.3 months (95% CI, 15.8–21.3) for RPTL patients (P<.001). Median OS was 29.7 months (95% CI, 27.4–NA) for CET LPTL vs 29.1 months (95% CI, 26.6–35.6) for BEV LPTL patients (HR, 0.87; 95% CI, 0.63–1.19; P=.378), and 17.0 months (95% CI, 12.0–32.6) for CET RPTL vs 18.8 months (95% CI, 15.8–22.3) for BEV RPTL patients (HR, 1.00; 95% CI, 0.68–1.46; P=.996). The interaction of treatment and PTL was not significant in the Cox regression. Conclusions: This study found a prognostic effect of PTL but not a predictive effect. LPTL patients had significantly longer OS vs RPTL patients. However, the treatment effect for CET vs BEV by PTL was not significantly different. Future research is needed to examine differences between real-world and clinical trial populations that may have contributed to divergent results.

HSR19-096: The Impact of New Oncology Drugs on Disability and Health Care Spending: An Assessment of Real-World Evidence

Razelle Kurzrock, MDa; Robert Goldberg, PhDb; Alice C. Ceacareanu, PharmD, PhDc; and Zachary A. P. Wintrob, MScc

aUniversity of California San Diego, San Diego, CA; bCenter for Medicine in the Public Interest, New York, NY; and cROAKETIN Inc., Clinical Services, Buffalo, NY

Background: It is sometimes suggested that newly-approved cancer treatments have only marginal effectiveness, which raises questions concerning their cost-vs-benefit ratio. Such concerns appear at odds with the lower cancer-related hospitalization rate and improved survival. However, such cost-effectiveness analyses rely on population-based averages obtained from the analysis of clinical trial data. By failing to analyze data from longitudinal datasets, such assessments are unable to account for real-world patient conditions and treatment patterns in evaluating clinical and cost-effectiveness. Longitudinally surveyed clinical data has the potential to objectively reveal any association between patient outcomes and new cancer treatment utilization. Methods: We investigated the effect of being prescribed a higher proportion of new oncology drugs on quality of life, medical services use, and productivity measures as reported by the Medical Expenditure Panel Survey (MEPS, 1996–2015). General linear models with Taylor series variance estimation were applied. New oncology drugs were defined as cancer treatments marketed after the year 2000. Included subjects (N=16,677) had a solid or hematologic malignancy diagnosis (CCCodex 11–47) and available prescription data. Individual age and employment status were accounted for as covariates. All analyses were performed using SAS version 9.4 (Cary, NC). Results: Unadjusted regression data show that individuals using newer oncology treatments missed on average 2.5 (±0.3 SE) fewer days of work or school per year as compared to patients using older drugs (43% improved productivity, P<.0001). The effect persisted even after adjusting for the magnitude of the effect (P<.0001). Accounting for age, the use of newer drugs was, on average, associated with ∼35% fewer missed work or school days. Cancer patients using newer treatments had 0.06 (±0.01 SE) fewer hospital admissions/year compared to patients using older treatments (P<.0001) and spent less time in the emergency room (P<.0411) with ∼45% fewer hospitalizations. Patients using newer medicines also had fewer health-related visits (P<.0001). Conclusion: Analysis of longitudinal real-world evidence gives a more comprehensive and reliable view of the clinical and economic impact of new oncology treatments. Our data suggests significant reductions in lost work/school days, hospitalizations, and use of medical services in general.

HSR19-097: Understanding U.S. Mutational Testing Patterns and Treatment Implications in Gastrointestinal Stromal Tumors (GIST) Patients

Khalid Mamlouk, PharmDa; Zach Crouch, PharmDa; Philina Lee, PhDa; Clive Mendonca, PhDb; Maria Gumina, PhDb; and Brian Rubin, MD, PhDc

aBlueprint Medicines, Inc., Cambridge, MA; bTrinity Partners, Waltham, MA; and cCleveland Clinic, Cleveland, OH

Background: In the United States, 3,000–4,000 cases of GIST occur each year. Approximately 80% of newly diagnosed GIST harbor mutations in KIT and up to 10% harbor mutations in PDGFRA, including an activation loop mutation at amino acid 842 that is associated with treatment resistance. Most patients (pts) with metastatic GIST (mGIST) respond initially to imatinib but progress and require alternative lines of therapy. However, pts with a PDGFRα D842V mutation have absolute resistance to imatinib and little to no benefit from other therapies. NCCN and ESMO GIST guidelines recommend routine mutational testing in all pts initiating therapy, though recent publications report low testing rates in this setting. This retrospective chart audit of GIST treaters further examines mutational testing and treatment decisions to better understand current practices and educational needs. Methods: Hematologist/oncologists who practiced in academic or community settings were eligible if they treated ≥5 pts with mGIST in the past 12 months. Respondents were asked to retrieve the last 4 charts of pts with metastatic/unresectable GIST of which ≥1 received ≥3 lines of therapy and answer questions geared toward assessing patient journey from diagnosis and mutation testing through treatment choices and outcomes. Results: 105 U.S. physicians (academic: 32; community: 73) reviewed and answered questions based on 403 patient charts. Overall KIT/PDGFRA mutation testing rates prior to initiating therapy was 59%, with lower rates in community settings. Further, only 34% of surveyed community physicians indicated they always perform mutational testing before initiating therapy despite it being strongly recommended in NCCN Guidelines. Top reasons cited for not performing mutation testing included questionable therapeutic impact, not part of practice’s standard protocol, cost, and lack of tissue. In our data set, 15 pts with a PDGFRα D842V mutation were identified, 12 of whom received imatinib despite known inactivity in this population. In the overall population, pts were primarily sequenced through imatinib, sunitinib, and regorafenib, with community clinicians being less likely to use regorafenib in the third-line setting. Conclusion: Continued education is needed about the role of mutational testing in GIST to ensure optimal treatment, especially among U.S. community oncology practitioners. The data also highlight the need for novel therapies that are effective in PDGFRα D842V-driven GIST.

HSR19-100: CancerSupportSource®-15: Development and Evaluation of a Short Form of a Distress Screening Program for Cancer Survivors

Shauna McManus, BSa; Alexandra K. Zaleta, PhDa; Melissa F. Miller, PhD, MPHa; Joanne S. Buzaglo, PhDb; Julie S. Olson, PhDa; Sara Goldberger, LCSW-Rc; and Kevin Stein, PhD, FAPOSa

aCancer Support Community, Research and Training Institute, Philadelphia, PA; bPRO Solutions, Vector Oncology, Memphis, TN; and cCancer Support Community, New York, NY

Background: CancerSupportSource (CSS) is a 25-item distress screening tool implemented at community-based cancer support organizations and hospitals nationwide. CSS assesses distress over 5 domains: (1) emotional concerns (including depression and anxiety risk screening subscales), (2) symptom burden, (3) body and healthy lifestyle, (4) healthcare team communication, and (5) relationships. This study developed a short form of CSS and examined its psychometric properties. Methods: 2,379 cancer survivors enrolled in the Cancer Support Community’s Cancer Experience Registry. Participants provided demographic and clinical background and completed CSS-25 and PROMIS-29, a measure of health-related quality of life. Item reduction was conducted with a subsample of 1,435 survivors and included external item quality (correlations between items and PROMIS-29 scales), internal item quality (inter-item and inter-factor correlations, factor loadings and structure, and item communalities from an exploratory factor analysis of CSS-25), and professional judgement (ranking/prioritization of items by CSS-25 developers, accounting for theoretical and practical implications). Pearson correlations and confirmatory factor analysis were conducted on a separate subsample of 944 survivors to corroborate psychometric properties and dimensionality of the shortened scale. Results: Scale refinement resulted in a 15-item short form of CSS (CSS-15). At least 1 item from each of the 5 CSS-25 domains was retained to preserve multidimensionality, including anxiety and depression risk screening subscale items. Additionally, 1 item about tobacco/substance use was kept due to clinical significance for risk assessment. In confirmatory factor analysis, the model explained 59% of the variance and demonstrated good fit (RMSEA=0.068, 90% CI=0.061–0.075; SRMR=0.033; CFI=0.959; χ2(68)=334.75, P<.001). Correlation between CSS-15 and CSS-25 was 0.986, P<.001. Total distress was associated with PROMIS subscales (rs=−.65–.75, ps<.001); internal consistency reliability was excellent (α=.92). Conclusions: CSS-15 is a brief, reliable, and valid multidimensional measure of distress. The reduced measure retained excellent internal consistency and a stable factor structure, while correlating well with CSS-25 and PROMIS-29. CSS-15 can serve as a practical tool to efficiently screen for distress among cancer patients and survivors.

HSR19-102: Direct and Indirect Economic Burden of Cervical Cancer (CxCa) in the United States in 2015: A Mixed-Methods Analysis

Chizoba Nwankwo, PhDa; Shelby L. Corman, PharmD, MS, BCPSb; Ruchit Shah, PhDb; and Youngmin Kwon, BAb

aMerck & Co., Inc., Kenilworth, NJ; and bPharmerit International, Bethesda, MD

Background: An estimated 12,820 women in the United States will be diagnosed with CxCa in 2018, with 4,210 deaths from the disease. The economic burden of CxCa, both in terms of healthcare costs and lost productivity, has not been adequately studied. Methods: This was a mixed-methods study that evaluated the direct and indirect costs of CxCa using data from the Medical Expenditure Panel Survey (MEPS) for prevalent CxCa cases and the National Center for Health Statistics (NCHS) for deaths due to CxCa. Total healthcare costs and number of work days missed were compared between CxCa cases and controls in MEPS, using propensity scores calculated from baseline demographics and comorbidities. Missed work was converted to costs using the average hourly wage for women in 2015. Per-patient incremental healthcare and lost work productivity costs were then multiplied by the number of prevalent cases of CxCa in 2015 obtained from the Surveillance, Epidemiology, and End Results Program (SEER). NCHS data on the age-stratified number of CxCa deaths per year (1935–2015) and life expectancy data from the Social Security Administration were then used to calcluate the number of women who would be alive in 2015 if they had not died from CxCa and the lost earnings resulting from early mortality. The primary study outcome was the total direct and indirect cost of CxCa in 2015, calculated as the sum of the incremental direct healthcare costs, incremental lost productivity costs due to missed work, and lost productivity costs resulting from early death due to CxCa. Results: An estimated 257,524 women were alive with CxCa in 2015. Total healthcare costs were $4,221 higher, and an additional 0.37 work days were missed in women with CxCa compared to propensity-matched controls. Of the 488,475 women who died of CxCa prior to 2015, 108,832 would be alive in 2015 and 38,540 would be part of the workforce. Lost earnings in 2015 attributable to these deaths totaled $2.19 billion. The total economic burden of CxCa in the United States in 2015 was thus estimated at $3.3 billion (Table 1). Conclusions: CxCa was responsible for nearly $3.3 billion in direct and indirect costs in 2015. Early death among women with CxCa was the biggest driver of total economic burden.

HSR19-103: Burden of Self-Reported Mental Health Among Women With Cervical Cancer (CC) in the United States

Chizoba Nwankwo, PhDa; and Michael J. Doane, PhDa

aMerck & Co., Inc., Kenilworth, NJ

Background: Globally, CC is the fourth most common cancer in women, with 569,847 new cases and 311,365 deaths from CC reported in 2018. Little is known about the burden of living with CC, especially related to mental health. This study examined patient-reported outcomes, including symptoms of depression and anxiety, amongst women with and without a diagnosis of CC. Methods: Data were aggregated from the 2016–2018 U.S. National Health and Wellness Surveys, a nationally representative, self-administered, internet-based survey of adults (N=247,484). Women who reported a physician-diagnosis of CC were matched 1:1 by propensity scores to a sample of women who did not report any cancer diagnoses. Propensity score matching was conducted using the following sociodemographic characteristics: age, race, possession of health insurance, smoking status, comorbidity status, body mass index, income, and year of survey completion. Bivariate analyses (ie, chi-square and t-tests) assessed differences in mental health outcomes between these 2 matched groups of female respondents. Outcomes included the following: (1) depressive severity via the Patient Health Questionnaire (PHQ-9), (2) suicidal ideation via the PHQ-9 (ie, thoughts of being better off dead on several days or more during the past 2 weeks), (3) anxiety severity via the Generalized Anxiety Disorder 7-Item Scale (GAD-7), and (4) healthcare resource use for mental health services (ie, visits to psychologists and psychiatrists during past 6 months). Results: Analyses of the propensity score matched sample of 1,044 women with a CC diagnosis versus 1,044 without a CC diagnosis showed that CC respondents reported significantly more severe scores of both depression (7.3 vs 6.0; P<.001) and anxiety (5.7 vs 4.7; P<.001). Although not statistically significant, a numerically greater proportion of CC respondents reported suicidal ideation during the past 2 weeks (19.0% vs 16.0%; P=.158). Respondents with CC were marginally more likely to visit a psychologist (8.6% vs 6.4%; P=.056) and were significantly more likely to visit a psychiatrist (8.6% vs 6.2%; P=.037) at least once during the prior 6 months than respondents without CC. Conclusions: CC is associated with mental health burden, including more severe symptoms of depression and anxiety as well as greater use of mental health services. This study highlights the likely impact of CC for both patients as well as the healthcare system.

  • View in gallery View in gallery
  • View in gallery

    ExoLR profiling in the diagnosis of PDAC. (A–B) ROC for performance of exoLR d-signature in the training (A) and validation cohorts (B). (C–D) Unsupervised hierarchical clustering of 8 exoLRs selected for use in the d-signature in the training (C) and validation cohorts (D). (E) ExoLR d-signature in healthy, chronic pancreatitis (CP), and pancreatic ductal adenocarcinoma (PDAC) subjects. (F) ExoLR d-signature in PDAC patients in stage I–IV. (G) ROC for performance of exoLR d-signature compared with that of CA19-9 in the differential diagnosis of PDAC and CP. (H) DCA to compare the net benefit of combined exoLR d-signature and CA19-9 (red line) with that of CA19-9 alone (blue line) for PDAC vs CP. AUC, area under the curve. CI, confidence interval.

  • View in gallery

    ExoLRs in the prognostic prediction of PDAC. (A–B) Kaplan–Meier curves of overall survival for PDAC patients with low or high risk, according to the prognostic signature (P-signature) in the training (A) and validation cohorts (B). (C) ROC for the p-signature, tumor stage, and p-signature combined with tumor stage in the whole PDAC cohort. (D) Kaplan–Meier survival curves of PDAC patients with combinations of exoLR p-signature and tumor stage in the whole PDAC cohort. AUC, area under the curve. CI, confidence interval.

  • View in gallery
All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 522 64 1
PDF Downloads 224 56 1
EPUB Downloads 0 0 0