Characterizing multiple sclerosis disease progression using a combined structural and functional connectivity metric

PURPOSE

A combined resting state functional connectivity MRI (fcMRI) and diffusion tensor imaging (DTI) metric called structural and functional connectivity index (SFCI) was recently proposed for tracking disease status and progression in multiple sclerosis (MS). The metric combines fcMRI and transverse diffusivity (TD) along different functional pathways involved in principle symptomatic domains of MS. In a longitudinal study of patients with MS receiving the same MS therapy, initial worsening of transcallosal (TC) motor and frontoparietal (FP) cognitive networks, as measured by fcMRI and DTI over the first year was followed by stabilization in the second year of follow-up. In this study we have (i) probed relationships between individual and composite neurological measures of MS with SFCI and its individual components along TC motor and FP cognitive pathways and (ii) compared sensitivity of SFCI to treatment-induced longitudinal changes with each individual imaging measure.

METHODS

Twenty five patients with MS (15 female, age 42 ± 8 y) participated in this study and were scanned at 3 T whole body MRI scanner with diffusion tensor imaging (DTI) and resting-state functional connectivity MRI (fcMRI) scan protocol at baseline and 6, 12, 18 and 24 months after starting fingolimod. fcMRI and TD along TC and FP pathways were combined to form structural and functional connectivity index (SFCI) at each time point. Correlations between individual/combined neurological measures and individual imaging components/SFCI at baseline and were evaluated and compared. In addition, efficacies of individual and combined imaging metrics in tracking network integrity were compared.

RESULTS

Individual TD along the TC pathway was significantly inversely correlated with all individual/composite neurological scores. There were moderate correlations of TC and FP components of SFCI with most of the neurological scores, and the pathway-combined SFCI was significantly correlated with all neurological scores. Trend-level increases of both TC and FP fcMRI were observed during the second year of follow-up, both TC and FP TD increased significantly in the first year and then stabilized during the second year. A trend toward a decrease in combined imaging metrics along TC and FP were observed during the first year, followed by a trend toward an increase in these metrics during the second year, while a significant decrease in SFCI during the first year followed by a significant increase during the second year was observed.

CONCLUSIONS

SFCI was more effective in tracking network integrity/disease progression than individual pathway-specific components, which supports its use as an imaging marker for MS disease status and progression.

Evaluating the impact of performance status criteria on minority eligibility for oncology clinical trials

e18701

Background: Restrictive criteria contribute to low enrollment in clinical trials. These criteria can also amplify health disparities by reducing the racial and ethnic diversity of the cohort. Many studies show that minority patients exhibit worse Eastern Cooperative Oncology Group (ECOG) performance status than white patients. Based on clinical trial data, some researchers have hypothesized that relaxing the ECOG criterion in clinical trial criteria could improve the diversity of the cohort. However, little research exists to measure ECOG’s impact on minority eligibility. Using Real-World Data (RWD), we evaluate whether relaxing the ECOG criterion monotonically increases the racial diversity of the cohort in oncology clinical trials. Methods: We used ConcertAI’s database of US oncology Electronic Medical Record (EMR) data, which includes clinical data from CancerLinQ Discovery™. We conducted sensitivity analyses for the inclusion criteria of 16 different clinical trials across multiple cancer indications. For each trial, we created five cohorts based on five different ECOG score upper limits. For example, the first cohort only included patients who had an ECOG of 0, the second cohort included patients with an ECOG of 0 or 1, etc. We then recorded the percentage of non-white patients for the resulting cohorts. We ran simple linear regressions to measure whether the change in the percentage of non-white patients was statistically significant. Results: Relaxation in ECOG status led to no uniform change in the racial diversity of patients across the 16 trials. The limited changes we did observe were not statistically significant. Conclusions: Our findings suggest that improving diversity will require more than just relaxing ECOG restrictions, and a multi-faceted approach may be needed. Future research exploring the relationship between ECOG and diversity should control for potential confounders like age, gender, and comorbidities using multivariable models. Such research needs to also account for patients with unknown race and ECOG, which represented large parts of our study population. Such research could elucidate whether the phenomenon observed in the general population—that minority patients tend to have worse performance status than whites—holds true in the sub-population of oncology patients.[Table: see text]

Impact of trial site selection on minority patient recruitment in prostate cancer trials

1558

Background: Historically, minority patients have been underrepresented in clinical cancer trials. Despite recognition of this problem, trials in the early 2000’s showed a decrease from 10.5% to 6.2% in African American trial participation when compared to trials from the early 1990’s. The drop in trial participation is also reflected in prostate cancer trials, although Black men have a 1.76 higher prostate cancer incidence rate than White men. Using prostate cancer as an example, we investigated the impact of trial site selection on potential minority patient recruitment; thus, overcoming a major system-level barrier to trial access. Methods: We created a prostate cancer cohort by filtering our real-world data sources (CancerLinQ, Electronic Medical Office Logistics) for adult male patients with ICD10 CM code C61* or ICD9 CM code 185 on 1/1/2015 or later (cohort #1). As a pre-requisite for computing site level prostate cancer patient counts, we used claims data to attribute missing site information. Finally, to identify the most promising sites for minority trial recruitment, we ranked sites by the proportion of Black patients and the overall cohort patient count. We repeated the above steps for a subset of cohort #1, which was based on the criteria for trial NCT00887198 investigating the prostate cancer drug abiraterone (cohort #2). Results: The prostate cancer cohort (#1) had 151,261 patients, of which 99,152 (65.6%) were attributed to sites. The percentage of Black patients being treated at the top ten sites ranged from 33.0% to 66.4%, with a median of 45.2% (see table). All ten sites had participated in an interventional cancer trial, and eight had participated in prostate cancer trials. Half of them were community, and half were academic sites. The abiraterone cohort (#2) had 1,267 patients, of which 1,174 (92.7%) were attributed to sites. Among the top ten sites the Black patient percentages ranged from 23.8% to 85.7%, with a median of 39.3%. Conclusions: In an analysis of 17 recent FDA drug registration trials for prostate cancer, Black trial participation ranged from only 1.4% to 6.2%, with a median of 3.0%. In contrast, Black patients being treated at the top sites in our data ranged from 33.0% to 66.4%, with a median of 45.2% (cohort #1). The percentages for the abiraterone cohort (#2) were similar, suggesting that even after applying trial criteria the Black patient percentages remain in the double-digits at top sites. Our results demonstrate that informed trial site selection could have a substantial positive impact on minority patient recruitment. [Table: see text]

Influence of equipment changes on MRI measures of brain atrophy and brain microstructure in a placebo-controlled trial of ibudilast in progressive multiple sclerosis

Background

Hardware changes can be an unavoidable confound in imaging trials. Understanding the impact of such changes may play an important role in the analysis of imaging data.

Objective

To characterize the effect of equipment changes in a longitudinal, multi-site multiple sclerosis trial.

Methods

Using data from a clinical trial in progressive multiple sclerosis, we explored how major changes in imaging hardware affected data. We analyzed the extent to which these changes affected imaging biomarkers and the estimated treatment effects by including such changes as a time-dependent covariate.

Results

Significant differences whole brain atrophy (brain parenchymal fraction, BPF) and microstructure (transverse diffusivity, TD) between scans with and without changes were found and depended on the type of hardware change. A switch from GE HDxt to Siemens Skyra led to significant shifts in BPF (p < 0.04) and TD (p < 0.0001). However, we could not detect the influence of hardware changes on overall trial outcomes- differences between placebo and treatment arms in change over time of BPF and TD (p > 0.5).

Conclusions

The results suggest that differences among hardware types should be considered when planning and analyzing brain atrophy and diffusivity in a longitudinal clinical trial.

Neurofilament light chain in a phase 2 clinical trial of ibudilast in progressive multiple sclerosis

BACKGROUND

Sensitive and specific biomarkers for use in progressive multiple sclerosis (MS) have not been established. We investigate neurofilament light (NfL) as a treatment response biomarker in progressive MS.

OBJECTIVE

To evaluate whether ibudilast 100 mg/day alters serum and cerebrospinal fluid (CSF) levels of NfL in progressive MS.

METHODS

In a protocol-defined exploratory analysis from a 2-year, phase 2 clinical trial of ibudilast in progressive MS (NCT01982942), serum samples were collected from 239 subjects and a subset contributed CSF and assayed using single-molecule assay (SIMOA) immunoassay. A mixed model for repeated measurements yielded log(NfL) as the response variable.

RESULTS

The geometric mean baseline serum NfL was 31.9 and 28.8 pg/mL in placebo and ibudilast groups, respectively. The geometric mean baseline CSF NfL was 1150.8 and 1290.3 pg/mL in placebo and ibudilast groups, respectively. Serum and CSF NfL correlations were r = 0.52 and r = 0.78 at weeks 48 and 96, respectively. Over 96 weeks, there was no between-group difference in NfL in either serum (p = 0.76) or CSF (p = 0.46). After controlling for factors that may affect NfL, no effect of ibudilast on NfL in either serum or CSF was observed.

CONCLUSION

Ibudilast treatment was not associated with a change in either serum or CSF NfL.

Minority patient reported attitudes regarding tissue donation and participation in cancer research

6558

Background: Minority populations are underrepresented in biospecimen banks created for cancer research, which has implications for future therapeutic approaches. Patients’ reluctance to donate biospecimens is perceived to be a potential barrier but is poorly studied. We present interim analyses of a survey to assess attitudes in a patient cohort comprised of racial/ethnic minorities. Methods: Patients filled out a validated 23-item survey [ J Cancer Educ, 2014. 29: p. 580-7] for this prospective cohort study approved by the Cleveland Clinic IRB. Surveys were provided in the outpatient oncology clinic of a Cleveland Clinic community hospital. Eligibility requirements included tissue diagnosis of any solid tumor malignancy in non-curative setting; age ≥ 18 years, ECOG PS 0-2, self-reported race/ethnicity as any other than non-Hispanic White, undergoing cancer therapy in the next 30 days. Data are presented for the first 90 patients surveyed in 2015-2016. Results: Median age was 69 years (range, 35-92). Only 24 (27%, 1 missing) had been asked to donate samples in the past; of those, 20 (83%) had donated. The majority (n = 60, 67%) were willing to donate samples. A higher proportion (75%) responded as being likely to donate samples if they learned more about the research and reasons for sample donation. A smaller proportion was likely to donate samples if they received money (30%) or health services (40%) in return. Many (55-73%) disagreed with negative statements such as, “I will be treated as a guinea pig,” and only (3-4%) disagreed with trust statements such as, “I trust sample banks/medical researchers/procedures.” However, 42% endorsed being “concerned that something like the Tuskegee study could happen again.” Conclusions: The majority of racial and ethnic minority patients in this study were willing to donate biospecimens for research, with an even greater likelihood of participation if appropriate rationales were provided. While systemic mistrust persists, the vast majority trusted medical researchers and procedures. Our findings suggest that underrepresentation of minorities in cancer biospecimen repositories, not likely attributable to patient reluctance, must be addressed to achieve health equity.

Informed decision making (IDM) for prostate cancer (PCa) screening in a high-risk population

112

Background: PCa incidence and mortality in African Americans (AA) is higher than in Caucasians. Health-education programs and culturally appropriate outreach to high-risk groups in accordance with American Cancer Society IDM guidelines can reduce disparities. Data show that it is hard to provide comprehensive unbiased education about screening to patients (pts). This study aims to examine whether IDM guidelines in a large high risk group setting can improve knowledge on PCa and screening decision, and whether such education program is overall beneficial to pts. Methods: Pts were included in one-day outreach event and were given a 15-question pre and post- test focused on standard informative educational PowerPoint and then were offered screening (PSA + DRE). Components of IDM were reviewed during this educational intervention. Demographics and family history was collected and UCSF 10-year mortality index was assessed to help IDM. Pre- and post- test number of correct answers were compared (Wilcoxon signed rank); pts were surveyed on their opinion on the program. The decision regarding screening after the intervention was tracked as well as the % of PCa diagnosed. Pts were tracked via an established navigation system to ensure follow up care. Results: 106 pts were included in the current analysis. Median number of correct answers at pre and post test was 8 and 11 (p < 0.001). Overall, 86% responded that they wanted screening. Of those, 92% were AA and 21% had family history of PCa; 21 pts had PSA only, 60 had PSA + DRE. 13 pts (16%) had abnormal PSA per NCCN guidelines, 5 (8%) had abnormal DRE. 5 PCa were biopsy-diagnosed, 4 had abnormal DRE + PSA; 1 had only abnormal DRE. Overall, 82% pts favored IDM before screening, 18% would prefer screening without IDM. 75% of all pts found the information “very helpful” in decision-making (within a 5-point Likert scale). Conclusions: Our education-based IDM led to significant improvement in knowledge about PCa screening. Most pts preferred education prior to screening. Our approach paired with the use of navigation program is feasible and was positively received by a large high risk group. Project is ongoing with more pts and follow up, and further validation is pending. Clinical trial information: NCT02419846.

Abstract P6-09-22: Detecting high mutational load ER+ breast cancer patients through Foundation One cancer gene panel mutations

Background:

Cancer gene panels such as Foundation One are widely used clinically for aiding with cancer treatment decision-making since single point mutations in important genes and gene pathways in the tumor can point to tumor susceptibility that can be targeted with specific drugs. Tumor mutational load has more recently been proposed as a useful prognostic factor and indicator of clinical benefit in treatment with PD-1 and CTLA-4 blockade therapy. Breast cancer estrogen receptor positive patients (ER+) in particular appear to have a subset of high mutational load patients that could benefit if identified. However whole exome sequencing needed to directly measure mutational load is not yet widely available as a clinical tool. In this study we evaluate the predictive value of Foundation One cancer gene panel mutations for estimating tumor genome-wide mutational load and its use in identifying a clinically meaningful subset of breast cancer patients.

Methods:

The Cancer Genome Atlas breast cancer sequencing data on 569 ER+ patients was used to establish mutational load distribution. ER+ patients were divided into low mutational load and high mutational load groups according to 3 criteria: mean mutational load, the point of inflection in the mutational load distribution and the mutational load that optimally separates groups in terms of survival. Foundation One (FO) mutational load was then calculated as the number of mutations present within the 314 genes queried by the panel. FO mutational load was used to predict whether patients fell into the low or high mutational load groups found through analysis of the full exome data. Receiver Operating Characteristic (ROC) curves were constructed and optimal values for specificity and sensitivity of the FO mutational load classification were found.

Results:

Mean mutational load for ER+ patients was found to be 57 mutations, the point of inflection of the mutational load distribution was established at 100 mutations, and the number of mutations that best separated groups in terms of survival was 160, (HR = 6.6, p-value=0.004). FO mutational load was found to be a good predictor for the low and high classifications established by all three criteria, with areas under the curve of 0.74, 0.91 and 0.945 respectively. The optimal predictive value of the FO mutational load classification was found at 5 mutations as the cut-off, with 94.2% specificity and 88% sensitivity for predicting groups defined by survival and 95% specificity and 71% sensitivity for those defined by the mean.

Conclusion:

The Foundation One cancer gene panel can be used to effectively identify a clinically meaningful subgroup of ER+ patients with high mutational load. These patients may benefit from targeted treatments such as PD-1 inhibitors being offered through clinical trials. The compromise in sensitivity that results from the reduction in number of genes queried by a panel means an important proportion of patients with high mutational load will be missed but this still translates to a large improvement in the identification of these patients given the wide availability of gene panels in the clinic. Basic and clinical follow-up studies need to take place to clinically validate the high mutational load ER+ patient subgroup.

Citation Format: Raska P, Abraham J, Budd T. Detecting high mutational load ER+ breast cancer patients through Foundation One cancer gene panel mutations [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-09-22.

Abstract P2-12-04: Impact of institution of young women's breast cancer clinic on time to treatment and utilization of fertility, genetics and social work consultations in women under age 50 with new diagnosis of early stage breast cancer

Background: Genetic counseling and fertility resources are often underutilized in young women with early stage breast cancer (ESBC) due, in part, to concerns about treatment delays. At our institution, women newly diagnosed with ESBC typically see a breast surgeon, medical oncologist and radiation oncologist in a multidisciplinary clinic with additional cancer related subspecialist referrals occurring at those providers' discretion. We hypothesized that time to treatment (TTT) and utilization of fertility, genetics and social work consultations would improve after implementing a Young Women's Breast Cancer Clinic. As of January 1, 2015, all patients under age 50 seen at Cleveland Clinic for new diagnosis of ESBC were automatically offered scheduling of appointments with medical genetics, reproductive endocrinology and social work in addition to the usual multidisciplinary team.

Methods: Women under age 50 diagnosed with ESBC seen at Cleveland Clinic from 1/2014-12/2015 were identified using our tumor registry. Demographics, tumor pathology, clinical and treatment histories were obtained through medical chart review as per IRB approved protocol. Time from initial visit in our system to date of treatment initiation was calculated for all patients and compared between the 2014 (pre-intervention) and 2015 (post-intervention) cohorts as was time from diagnosis (biopsy date) to treatment initiation. Completed reproductive endocrinology, genetic counseling and social work consultations were documented. Welch two sample t-test was used to compare time to treatment between groups. Chi squared test was used to compare frequency of subspecialty consultations between groups.

Results: 207 young women with ESBC were identified over the 2 year period, 99 in 2014 and 108 in 2015. Median age was 45 in 2014 and 44 in 2015. Most were diagnosed outside of our hospital system, 58% in 2014 and 76% in 2015. The most common initial treatment was surgery with reconstruction (S+R) (54% and 50% for 2014 and 2015 respectively) followed by chemotherapy (23% and 27%) then surgery without reconstruction (S) (20% and 24%). Median TTT from first encounter was 30 days in 2014 and 28 days in 2015 (p=0.33) and was 36 days versus 33.5 days (p=0.23) when calculated from biopsy date. TTT in the S and S+R groups was 37 vs 28 days (p=0.84) and 36.5 vs 32 days, (p=0.21), respectively. Genetics, reproductive endocrinology and social work consults in 2014 vs 2015 were documented as 89% vs 94%, 4% vs 9% and 58 vs% 55% (p=0.22, 0.32, 0.77). For patients under age 40, 27% in 2014 and 30% in 2015 completed reproductive endocrinology consultations.

Conclusions: Offering upfront scheduling of breast cancer related subspecialty appointments for young women with newly diagnosed ESBC did not significantly improve overall TTT. There was a trend towards improved TTT in those receiving surgery with or without reconstruction as first treatment and no suggestion of delay in TTT. A modest numeric increase in completed genetic counseling and reproductive endocrinology consultations was not statistically significant, but may have been clinically meaningful for affected individuals.

Citation Format: Kruse ML, Raska P, Abraham J, Budd GT, Montero A, Grobmyer S, Moore H. Impact of institution of young women's breast cancer clinic on time to treatment and utilization of fertility, genetics and social work consultations in women under age 50 with new diagnosis of early stage breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-12-04.

Abstract C02: Electronic personal health records, minorities and prevention

Introduction:

Cancer outcome disparities are in large part attributable to impediments to access to care for minority populations and individuals of lower socio-economic status. The use of electronic personal health records (ePHR) such as myChart, which allow patients to directly interact with their health record and their healthcare providers, has been drastically increasing with cancer patients overall but disproportionately less so in the underserved. We need to better understand the factors that underlie disparities in usage of ePHR, particularly for prevention, in order to be better able to leverage on this technology to improve access to care.

Methods:

In 2012 Cleveland Clinic implemented an opt out program where every patient is given a myChart activation code at check in. This has resulted in close to a million “newly active” patients and almost half a million activated MyChart users. All patients with an email address on file were invited to participate in a survey inquiring about patient knowledge, use and preferences for myChart. We used logistic regression to test for differences across ethnicity after controlling for factors that confound factors such as age and illness.

Results:

A total of 18,599 369 myChart users responded to the survey. 633 (3%) responders self-identified as Hispanic (His) and 1265 (7%) as African American (AA). Both His and AA tended to be younger and healthier (less chronic conditions) than non-Hispanic White (WHI) responders (AA OR=1.4, p = 0.001, His OR 2.0, p = 0.001). Despite this His and AA have higher frequencies of visits to their PCP and to the ER after controlling for chronic conditions and age. Although younger WHI responders were more likely to say that they would use myChart more if they visited their PCP more, this was not the case for His and AA. Instead we find that His and AA are more likely to say that they do not discuss preventative measures with their PCP (AA OR =1.3, p=0.01, His OR = 1.5, p = 0.002)

Conclusion: Our results suggest that adding functionality to ePHR that will more actively promote discussion and planning of prevention between primary care providers and His and AA may be a way to improve the impact of this technology on how minorities engage with prevention.

Citation Format: Paola Raska, Katherine Schneider, Jordan Zivoder, Scott Steele, Jame Abraham. Electronic personal health records, minorities and prevention. [abstract]. In: Proceedings of the Ninth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2016 Sep 25-28; Fort Lauderdale, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2017;26(2 Suppl):Abstract nr C02.

Retrospective study of the efficacy and safety of neoadjuvant docetaxel, carboplatin, trastuzumab/pertuzumab (TCH-P) in nonmetastatic HER2-positive breast cancer

BACKGROUND

Pertuzumab is FDA approved in the preoperative setting in combination with trastuzumab and chemotherapy, in women with nonmetastatic HER2 + breast cancer. The TRYPHAENA trial (n = 77) reported a pathologic complete response rate (pCR), i.e., ypT0ypN0, of 52 % in patients treated with neoadjuvant (docetaxel, carboplatin, trastuzumab, & pertuzumab) TCH-P. Aside from this study, there is limited information regarding the safety and efficacy of TCH-P in the neoadjuvant setting. Our goal was to evaluate the safety and efficacy of neoadjuvant TCH-P in a non-clinical trial setting.

MATERIALS AND METHODS

Cancer data registry was utilized to identify patients with HER2 + nonmetastatic breast cancer that received neoadjuvant TCH-P. pCR was defined as the absence of invasive or noninvasive cancer in breast and lymph nodes, i.e., ypT0ypN0.

RESULTS

70 patients with a median age of 52 years met our inclusion criteria. Clinical staging was I-8.5 %; II-68.5 %; and III-22.8 %. 60 % of patients had hormone receptor (HR)-positive tumors. 23 % (16/71) of patients required dose reduction for rash, diarrhea, neuropathy, or thrombocytopenia. Overall, no patients developed grade 3-4 left ventricular systolic dysfunction(LVSD); an asymptomatic reduction in LVEF of >10 % was observed in three patients. The overall observed pCR rate was 53 %. As expected, the pCR rate was higher in patients with HR-negative breast cancer than for patients with HR+ disease: 69 % (20/29) vs. 42 % (17/41), respectively. The axillary downstaging rate was approximately 53 % (19/36).

CONCLUSION

Neoadjuvant TCH-P, in a nonclinical trial setting, was associated with a pCR rate of 53 % similar the reported rate in TRYPHAENA. Toxicity was manageable, with no patients experiencing symptomatic heart failure.

Abstract P6-12-08: Healthcare barrier profiles in patients navigated for cancer screening and treatment and the impact of the affordable care act

Background: The underserved community experiences barriers to cancer screening that result in overall greater mortality rates across all cancers. Insurance coverage brought forth by the Affordable Care Act has the potential to significantly impact these barriers through reducing the burden of health care cost on the patient. In this study the authors observe the impact of the Affordable Care Act on the barrier profiles presented by the patients navigated for cancer screening and treatment.

Methods: Patient navigation encounters were recorded for a total of 1146 patients navigated for cancer screening and treatment at the Cleveland Clinic Foundation from the years 2012 through 2015. A total of 3259 encounters were classified into barrier types. Health care billed encounters were retrieved from EPIC for this group of patients from the time they entered patient navigation and classified in terms of insurance coverage. Patients were categorized according to their barrier profile. Appropriate generalized linear regression models were used to test for association of these profiles to number and types of navigation and health care encounters and cost, and to test for change in types of encounters and patient barrier profiles through time.

Results: The insurance barrier is present in 23% of all navigation encounters. Patients presenting with an insurance barrier had a greater mean number of navigation (p&lt;0.001) and health care encounters (p&lt;0.006), had a greater proportion of self-paid health care encounters (p&lt;0.001) and a lower total cost billed for health care encounters after controlling for number of encounters (p&lt;0.001) . The access barrier is present in 53% of navigation encounters while patients that present with only the access barrier account for 42% of the entire sample. Patients that present with only the access barrier have doubled every year (OR 2.2 per year, 95% CI [1.8 2.6]) from 2012 to 2015, while the proportion of self-paid health care encounters (OR 0.26 per year, 95% CI [0.25 0.28]) and the presence of the insurance barrier (OR 0.55 per year, 95% CI [ 0.49 0.62]) have more than halved during this time period.

Conclusion: Although the Affordable Care Act has clearly had an impact by lowering the number of insurance barrier navigation encounters through time, it has uncovered access as the predominant remaining barrier. Understanding and targeting the access barrier will be the most effective way to potentiate the effects of the ACA on patients being navigated for cancer screening and treatment.

Citation Format: Raska P, Ott MA, Steele A, Bailey A, Hickman D, Allen D, Urbanek D, Glass K, Bailey J, Bell K, Montero A, Abraham J. Healthcare barrier profiles in patients navigated for cancer screening and treatment and the impact of the affordable care act. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P6-12-08.

Abstract P4-13-24: Impact of genomic medicine on clinical decision making in patients with advanced breast cancer at two academic medical centers

Background: A deeper molecular understanding of cancer biology has led to the development of therapies targeting driver mutations. Genomic profiling of tumors is commercially available and has become integrated into many clinical practices as part of a paradigm shift towards personalized care of cancer patients. The current impact of genomic profiling on clinical decision making for patients with advanced breast cancer is not well described.

Methods: Patients with metastatic breast cancer (mBC) who had tumors submitted for commercial genomic analysis from 2013-2015 were identified consecutively at two large academic cancer centers with genomic basket trials open for the majority of the collection period. Demographics, tumor pathology, clinical, and treatment histories were extracted through medical chart review as per an IRB approved protocol. Data from genomic analysis reports was extracted including number and type of mutations, FDA approved therapies and clinical trials available. Genomic analysis was determined to have impacted clinical decision making if the next line of therapy was influenced either by accrual to clinical trial, or a decision to prescribe an FDA-approved therapy. The most frequent somatic mutations and their relative frequencies were determined.

Results: A total of 82 patients with mBC who had undergone commercially available genomic testing were identified. The median age was 49 (range: 29-70). 42 patients (51%) had ER-positive HER2-negative disease, 33 (40%) had ER-negative HER2-negative disease, 4 (5%) had ER-negative HER2-positive disease and 3 (4%) had ER-positive HER2-positive disease. The median number of lines of therapy received prior to genomic profiling was 2 (range 0-15). Genomic analysis reports suggested that 61 (74%) of these patients had at least one FDA approved medication available for at least one somatic mutation, and 79 (96%) had at least one clinical trial available (39 (46%) in the same state, 11 (13%) in the same institution). Genomic testing influenced management in 8 patients (10%), with 6 patients (7%) experiencing a change in next line of therapy attributable to genomic information. In 74 patients (90%), genomic testing results did not affect clinical decision-making. The most frequently observed somatic mutations included: TP53, PI3KCA, MYC, CCDN1, FGF, ZNF703, GATA3, ARID1A, MCL1, PTEN, MYST3, and BRCA1.

Conclusions: Genomic testing did not affect management in the vast majority of mBC patients treated at two major academic cancer centers. Furthermore, the most identified mutated genes found were not targetable. The real world clinical utility of genomic analysis remains limited in breast cancer but may influence clinical decision making in a minority of patients.

Citation Format: Kruse ML, Santa-Maria CA, Raska P, Swoboda A, Jain S, Sohal D, Moore H, Budd GT, Abraham J, Montero AJ. Impact of genomic medicine on clinical decision making in patients with advanced breast cancer at two academic medical centers. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P4-13-24.

Abstract P1-09-03: Socioeconomic characteristics of African American women with breast cancer

Background: Breast cancer is the most common cancer among African American (AA) women. Despite having a lower incidence of breast cancer compared to white women (124.4 compared to 127.9 per 100,000), AAs have a higher death rate (30.2 compared to 21.3 per 100,000). One explanation for this discrepancy is that breast cancer in AAs is often detected at a later stage compared to white women. We conducted this retrospective study to examine socioeconomic characteristics among AA women with breast cancer to see if there were factors associated with stage of diagnosis which may contribute to the known disparities. Methods: We identified all AA women diagnosed with any stage breast cancer from 2006-2014 within the Cleveland Clinic Cancer Data Warehouse and classified them into either early or late stage disease at time of diagnosis. Stages 0-II were classified as early and stages III-IV as late. We examined several variables at diagnosis including age, marital status, tobacco use, alcohol use, Medicaid insurance status, and breast cancer subtype which included HER-2 positive (HER+), hormone receptor positive/HER2 negative (HR+/HER-), and triple negative(TN). AA median income was obtained from US census data according to the zip code at diagnosis. We conducted univariate logistic regression for individual estimates and confidence intervals and multiple logistic regression and model selection to determine significant predictors of stage of diagnosis. Results: Of the 771 AA women identified, 108 (14%) were diagnosed at a late stage of disease with a median age of 59 years. Receptor status distribution was 12.4%, 31%, and 16.6% for HER+, HR+/HER-, and TN respectively for early stage, and 15.7%, 27%, and 25% for late stage. Among early stage 50% were current or previous smokers and 2.6% had Medicaid insurance compared to late stage patients where 63% were current or previous smokers and 9.2% had Medicaid insurance. Multiplicative effect estimates and 95% confidence intervals from univariate logistic regressions identified the following significant factors: tobacco use 1.48 [1.11-1.96] and Medicaid 3.73 [1.56-8.51] (p-values&lt;0.01), and TNBC 1.67 [1.02-2.68] (p-value&lt;0.05). In a stepwise model selection, only tobacco use and Medicaid were retained in the model, as well as age at diagnosis. Conclusions: There are socioeconomic differences among AA women with breast cancer. Only tobacco use, Medicaid insurance, and age at diagnosis were predictive of late stage in this study.

Citation Format: Winter A, Raska P, Ornstein M, Moore H, Montero A, Budd GT, Tullio K, Bailey J, Abraham J. Socioeconomic characteristics of African American women with breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P1-09-03.

Population structure at different minor allele frequency levels

Inferring population genetic structure from large-scale genotyping of single-nucleotide polymorphisms or variants is an important technique for studying the history and distribution of extant human populations, but it is also a very important tool for adjusting tests of association. However, the structures inferred depend on the minor allele frequency of the variants; this is very important when considering the phenotypic association of rare variants.

Using the Genetic Analysis Workshop 18 data set for 142 unrelated individuals, which includes genotypes for many rare variants, we study the following hypothesis: the difference in detected structure is the result of a "scale" effect; that is, rare variants are likely to be shared only locally (smaller scale), while common variants can be spread over longer distances. The result is similar to that of using kernel principal component analysis, as the bandwidth of the kernel is changed. We show how different structures become evident as we consider rare or common variants.

Identification and molecular characterization of a new ovarian cancer susceptibility locus at 17q21.31

Epithelial ovarian cancer (EOC) has a heritable component that remains to be fully characterized. Most identified common susceptibility variants lie in non-protein-coding sequences. We hypothesized that variants in the 3' untranslated region at putative microRNA (miRNA)-binding sites represent functional targets that influence EOC susceptibility. Here, we evaluate the association between 767 miRNA-related single-nucleotide polymorphisms (miRSNPs) and EOC risk in 18,174 EOC cases and 26,134 controls from 43 studies genotyped through the Collaborative Oncological Gene-environment Study. We identify several miRSNPs associated with invasive serous EOC risk (odds ratio=1.12, P=10(-8)) mapping to an inversion polymorphism at 17q21.31. Additional genotyping of non-miRSNPs at 17q21.31 reveals stronger signals outside the inversion (P=10(-10)). Variation at 17q21.31 is associated with neurological diseases, and our collaboration is the first to report an association with EOC susceptibility. An integrated molecular analysis in this region provides evidence for ARHGAP27 and PLEKHM1 as candidate EOC susceptibility genes.

European American stratification in ovarian cancer case control data: the utility of genome-wide data for inferring ancestry

We investigated the ability of several principal components analysis (PCA)-based strategies to detect and control for population stratification using data from a multi-center study of epithelial ovarian cancer among women of European-American ethnicity. These include a correction based on an ancestry informative markers (AIMs) panel designed to capture European ancestral variation and corrections utilizing un-thinned genome-wide SNP data; case-control samples were drawn from four geographically distinct North-American sites. The AIMs-only and genome-wide first principal components (PC1) both corresponded to the previously described North or Northwest-Southeast axis of European variation. We found that the genome-wide PCA captured this primary dimension of variation more precisely and identified additional axes of genome-wide variation of relevance to epithelial ovarian cancer. Associations evident between the genome-wide PCs and study site corroborate North American immigration history and suggest that undiscovered dimensions of variation lie within Northern Europe. The structure captured by the genome-wide PCA was also found within control individuals and did not reflect the case-control variation present in the data. The genome-wide PCA highlighted three regions of local LD, corresponding to the lactase (LCT) gene on chromosome 2, the human leukocyte antigen system (HLA) on chromosome 6 and to a common inversion polymorphism on chromosome 8. These features did not compromise the efficacy of PCs from this analysis for ancestry control. This study concludes that although AIMs panels are a cost-effective way of capturing population structure, genome-wide data should preferably be used when available.

Analysis of exome sequences with and without incorporating prior biological knowledge

Next-generation sequencing technology provides new opportunities and challenges in the search for genetic variants that underlie complex traits. It will also presumably uncover many new rare variants, but exactly how these variants should be incorporated into the data analysis remains a question. Several papers in our group from Genetic Analysis Workshop 17 evaluated different methods of rare variant analysis, including single-variant, gene-based, and pathway-based analyses and analyses that incorporated biological information. Although the performance of some of these methods strongly depends on the underlying disease model, integration of known biological information is helpful in detecting causal genes. Two work groups demonstrated that use of a Bayesian network and a collapsing receiver operating characteristic curve approach improves risk prediction when a disease is caused by many rare variants. Another work group suggested that modeling local rather than global ancestry may be beneficial when controlling the effect of population structure in rare variant association analysis.

Abstract 1678: European ancestry and ovarian cancer: Testing for their association using case control data

Population stratification in case-control studies can result from differences in ancestral make-up between cases and controls due to sampling variation. Ancestry is therefore routinely estimated and controlled for in these studies. Although less common, it is also possible for ancestry to truly be associated to the disease of interest. Controlling for ancestry can then compromise the power to detect variants that underlie both ancestry and disease. Using data from a multi-center case-control study of epithelial ovarian cancer (OC) among women of European-American ethnicity in which case-control samples were drawn from four geographically distinct North-American sites, we developed a three-step method to investigate whether the stratification observed across the four sites is due to chance or to a non-spurious association between European ancestry and OC. First, we collapsed genome-wide data into two principal components (PCs) that represent the major axes of European ancestral variation, which correspond to north-south and to west-east geographical clines. Next, we determined the discriminant function based on these two PCs that best separates cases and controls for each individual study site and we used the largest angle between the four discriminant function vectors as a measure of similarity ranging from 0 degrees, where all four vectors are exactly the same, to 180 degrees where two of the vectors face exactly opposite directions. If stratification is due to sampling, it is unlikely that the four sites will present similar discriminant functions by chance alone. We then gauged significance of similarity by permuting the case control status across individuals within each site and recalculating the maximum angle between the discriminant function vectors, thus obtaining a null distribution for the statistic. We found that all four sites present the greatest separation between cases and controls on the west-east axis of ancestral variation, with a higher proportion of ovarian cancer cases presenting western ancestry. The largest angle between the four discriminant function vectors observed was of 36.5 degrees. Permutation testing showed this level of similarity between the functions to have a p-value &lt; 0.01 and suggests that OC risk is somehow tied to European ancestry. It remains to be seen whether shared culture or other exposures such as parity or oral contraceptive use explain this connection or whether genetic variants underlie it and are potentially being missed due to routine ancestry control in association testing.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1678. doi:1538-7445.AM2012-1678

Development of a panel of genome-wide ancestry informative markers to study admixture throughout the Americas

Most individuals throughout the Americas are admixed descendants of Native American, European, and African ancestors. Complex historical factors have resulted in varying proportions of ancestral contributions between individuals within and among ethnic groups. We developed a panel of 446 ancestry informative markers (AIMs) optimized to estimate ancestral proportions in individuals and populations throughout Latin America. We used genome-wide data from 953 individuals from diverse African, European, and Native American populations to select AIMs optimized for each of the three main continental populations that form the basis of modern Latin American populations. We selected markers on the basis of locus-specific branch length to be informative, well distributed throughout the genome, capable of being genotyped on widely available commercial platforms, and applicable throughout the Americas by minimizing within-continent heterogeneity. We then validated the panel in samples from four admixed populations by comparing ancestry estimates based on the AIMs panel to estimates based on genome-wide association study (GWAS) data. The panel provided balanced discriminatory power among the three ancestral populations and accurate estimates of individual ancestry proportions (R²>0.9 for ancestral components with significant between-subject variance). Finally, we genotyped samples from 18 populations from Latin America using the AIMs panel and estimated variability in ancestry within and between these populations. This panel and its reference genotype information will be useful resources to explore population history of admixture in Latin America and to correct for the potential effects of population stratification in admixed samples in the region.

Individuals from Latin America are descendants of multiple ancestral populations, primarily Native American, European, and African ancestors. The relative proportions of these ancestries can be estimated using genetic markers, known as ancestry informative markers (AIMs), whose allele frequency varies between the ancestral groups. Once determined, these ancestral proportions can be correlated with normal phenotypes, can be associated with disease, can be used to control for confounding due to population stratification, or can inform on the history of admixture in a population. In this study, we identified a panel of AIMs relevant to Latin American populations, validated the panel by comparing estimates of ancestry using the panel to ancestry determined from genome-wide data, and tested the panel in a diverse set of populations from the Americas. The panel of AIMs produces ancestry estimates that are highly accurate and appropriately controlled for population stratification, and it was used to genotype 18 populations from throughout Latin America. We have made the panel of AIMs available to any researcher interested in estimating ancestral proportions for populations from the Americas.

Rare variant density across the genome and across populations

Next-generation sequencing allows for a new focus on rare variant density for conducting analyses of association to disease and for narrowing down the genomic regions that show evidence of functionality. In this study we use the 1000 Genomes Project pilot data as distributed by Genetic Analysis Workshop 17 to compare rare variant densities across seven populations. We made the comparisons using regressions of rare variants on total variant counts per gene for each population and Tajima's D values calculated for each gene in each population, using data on 3,205 genes. We found that the populations clustered by continent for both the regression slopes and Tajima's D values, with the African populations (Yoruba and Luhya) showing the highest density of rare variants, followed by the Asian populations (Han and Denver Chinese followed by the Japanese) and the European populations (CEPH [European-descent] and Tuscan) with the lowest densities. These significant differences in rare variant densities across populations seem to translate to measures of the rare variant density more commonly used in rare variant association analyses, suggesting the need to adjust for ancestry in such analyses. The selection signal was high for AHNAK, HLA-A, RANBP2, and RGPD4, among others. RANBP2 and RGPD4 showed a marked difference in rare variant density and potential selection between the Luhya and the other populations. This may suggest that differences between populations should be considered when delimiting genomic regions according to functionality and that these differences can create potential for disease heterogeneity.

Replication of GWAS "Hits" by Race for Breast and Prostate Cancers in European Americans and African Americans

In this study, we assessed association of genome-wide association studies (GWAS) “hits” by race with adjustment for potential population stratification (PS) in two large, diverse study populations; the Carolina Breast Cancer Study (CBCS; N total = 3693 individuals) and the University of Pennsylvania Study of Clinical Outcomes, Risk, and Ethnicity (SCORE; N total = 1135 individuals). In both study populations, 136 ancestry information markers and GWAS “hits” (CBCS: FGFR2, 8q24; SCORE: JAZF1, MSMB, 8q24) were genotyped. Principal component analysis was used to assess ancestral differences by race. Multivariable unconditional logistic regression was used to assess differences in cancer risk with and without adjustment for the first ancestral principal component (PC1) and for an interaction effect between PC1 and the GWAS “hit” (SNP) of interest. PC1 explained 53.7% of the variance for CBCS and 49.5% of the variance for SCORE. European Americans and African Americans were similar in their ancestral structure between CBCS and SCORE and cases and controls were well matched by ancestry. In the CBCS European Americans, 9/11 SNPs were significant after PC1 adjustment, but after adjustment for the PC1 by SNP interaction effect, only one SNP remained significant (rs1219648 in FGFR2); for CBCS African Americans, 6/11 SNPs were significant after PC1 adjustment and after adjustment for the PC1 by SNP interaction effect, all six SNPs remained significant and an additional SNP now became significant. In the SCORE European Americans, 0/9 SNPs were significant after PC1 adjustment and no changes were seen after additional adjustment for the PC1 by SNP interaction effect; for SCORE African Americans, 2/9 SNPs were significant after PC1 adjustment and after adjustment for the PC1 by SNP interaction effect, only one SNP remained significant (rs16901979 at 8q24). We show that genetic associations by race are modified by interaction between individual SNPs and PS.

Defining genetic determinants of the Metabolic Syndrome in the Framingham Heart Study using association and structural equation modeling methods

The Metabolic Syndrome (MetSyn), which is a clustering of traits including insulin resistance, obesity, hypertension and dyslipidemia, is estimated to have a substantial genetic component, yet few specific genetic targets have been identified. Factor analysis, a sub-type of structural equation modeling (SEM), has been used to model the complex relationships in MetSyn. Therefore, we aimed to define the genetic determinants of MetSyn in the Framingham Heart Study (Offspring Cohort, Exam 7) using the Affymetrix 50 k Human Gene Panel and three different approaches: 1) an association-based "one-SNP-at-a-time" analysis with MetSyn as a binary trait using the World Health Organization criteria; 2) an association-based "one-SNP-at-a-time" analysis with MetSyn as a continuous trait using second-order factor scores derived from four first-order factors; and, 3) a multivariate SEM analysis with MetSyn as a continuous, second-order factor modeled with multiple putative genes, which were represented by latent constructs defined using multiple SNPs in each gene. Results were similar between approaches in that CSMD1 SNPs were associated with MetSyn in Approaches 1 and 2; however, the effects of CSMD1 diminished in Approach 3 when modeled simultaneously with six other genes, most notably CETP and STARD13, which were strongly associated with the Lipids and MetSyn factors, respectively. We conclude that modeling multiple genes as latent constructs on first-order trait factors, most proximal to the gene's function with limited paths directly from genes to the second-order MetSyn factor, using SEM is the most viable approach toward understanding overall gene variation effects in the presence of multiple putative SNPs.