Distribution Of Risk Scores Research Articles

Abstract 024: Effects Of Diet On 10-year Atherosclerotic Cardiovascular Disease Risk Using The Pooled Cohort Equations Risk Calculator: Results From The Dash Trial

Background: The Dietary Approaches to Stop Hypertension (DASH) diet is known to reduce cardiovascular disease (CVD) risk factors, but its effects on 10-year CVD risk based on the pooled cohort estimating equation has not been reported. Objective: To determine the effects of adopting the DASH diet on 10-year atherosclerotic cardiovascular disease (ASCVD) risk compared to a typical American (control) diet or a diet rich fruit and vegetables (F/V), but otherwise similar to control. Methods: The DASH trial was a 3-arm, parallel-group, randomized controlled feeding trial of 459 adults aged 22 to 75 years without CVD and not taking anti-hypertensive or diabetes medications. These participants were randomized to a control diet, a F/V diet, or the DASH diet for 8 weeks. Weight was kept constant. Blood pressure (BP) and lipids were measured at baseline and at 8-weeks to compare 10-year ASCVD risk scores across dietary assignments. Comparisons were performed via linear regression adjusted for baseline ASCVD risk score. Results: The mean age of participants was 45 years; 49% were women, 60% were black, and 10% were current smokers. Mean systolic BP was 131.3±10.8 mm Hg, mean LDL cholesterol was 121±32 mg/dL, and mean HDL cholesterol was 48±14 mg/dL. Both DASH and F/V diets shifted the distribution of ASCVD risk scores downward compared to the control diet ( Figure, Panel A ). Compared to the control diet, the DASH and F/V diets reduced 10-year ASCVD risk by 10.0% (95% CI: -17.7, -1.5; P = 0.02) and 11.7% (95% CI: -19.3, -3.3; P = 0.007) respectively ( Figure, Panel B ). There was no difference between the DASH and F/V diets (-1.9%; 95% CI: -10.3, 7.4; P = 0.68). Conclusions: Compared to the control diet, the DASH and F/V diets reduced 10-year ASCVD risk, while the DASH and F/V had similar effects.

Frequency and Temporal Distribution of Postoperative Respiratory Depressive Events

The frequency and temporal distribution of postoperative respiratory depression (RD) events are not completely understood. This study determined the temporal distribution and frequency of RD episodes in postsurgical patients continuously monitored by bedside capnography and pulse oximetry. This was a post hoc study of a subset of postsurgical patients enrolled in The PRediction of Opioid-induced respiratory Depression In patients monitored by capnoGraphY (PRODIGY) trial from 2 sites in the United States. These patients had undergone continuous bedside monitoring on general care wards. These data were adjudicated for potential RD episodes. The number of RD episodes per patient and the time of each RD episode were determined. The first RD episode experienced by a patient was classified as an "initial" episode, and the initial and all subsequent RD episodes experienced by a patient were classified as "all" episodes. A PRODIGY risk score was calculated. Data analyzed from 250 patients contained 2539 RD episodes in 155 (62.0%, 95% confidence interval, 55.7-68.0) patients with median 2 [0-8], range of 0-545 RD episodes per patient, with a PRODIGY risk score distribution of 100 (40.0%) low, 79 (31.6%) intermediate, 70 (28.0%) high (missing data from 1 patient). Median time to the initial RD episode was 8.8 [5.1-18.0] hours postoperatively. There was a peak occurrence of initial RD events between 14:00 and 20:00 on the day of surgery, and these were associated with a large number of subsequent events in the same timeframe. The peak time of all RD episodes occurred from 02:00 to 06:00. Patients with high PRODIGY risk scores had higher incidence and greater number of RD episodes per patient (P < .001, overall comparisons between groups for both incidence [χ2] and number of episodes [Kruskal-Wallis test]). Continuous monitoring of surgical patients demonstrates that RD episodes are common, and risk increases with higher PRODIGY scores. In this patient cohort, the rate of initial RD episodes peaked in the afternoon to early evening, while peak rate of all RD episodes occurred in early morning. Further, among patients with RD episodes, the number of episodes increased with higher PRODIGY scores.

P032 - MammaPrint risk score distribution in South African breast cancer patients with the pathogenic BRCA2 c.7934delG founder variant: towards application of genomic medicine at the point-of-care

P032 - MammaPrint risk score distribution in South African breast cancer patients with the pathogenic BRCA2 c.7934delG founder variant: towards application of genomic medicine at the point-of-care

Background characteristics and anticoagulant usage patterns of elderly non-valvular atrial fibrillation patients in the ANAFIE registry: a prospective, multicentre, observational cohort study in Japan

ObjectiveTo explore anticoagulant usage patterns stratified by stroke and bleeding risk in elderly patients with non-valvular atrial fibrillation (NVAF).DesignProspective, multicentre, observational cohort study.SettingThe real-world All Nippon AF In the Elderly...

Abstract PS5-18: Association of factors of the obesity phenotype with breast cancer recurrence risk measured by oncotype recurrence score and short-term response to neoadjuvant endocrine therapy

Abstract Background:Increased body mass index (BMI) and metabolic syndrome are associated with an increased risk of breast cancer recurrence. The mechanism(s) responsible for this increased risk are not known, though multiple mechanisms including resistance to endocrine therapy have been proposed. We sought to determine whether obesity and/or metabolic syndrome were associated with intrinsic tumor risk as determined by the Oncotype recurrence score (RS), or with resistance to neoadjuvant endocrine therapy, as determined by Ki67 response after short-term presurgical treatment. Methods:To investigate whether BMI was associated with increased intrinsic recurrence risk, we analyzed a cohort of 779 patients treated at Vanderbilt for early stage breast cancer and who had Oncotype test results. Medical records were reviewed to record BMI at the time of breast cancer diagnosis. We compared the Oncotype RS between patients with BMI greater or less than 27, and in normal, overweight, and obese ranges. To investigate whether the increase in breast cancer recurrence risk associated with obesity could be identified by a surrogate marker of short-term response to endocrine therapy, we analyzed data from a presurgical trial of endocrine therapy in early stage breast cancer patients previously completed at Vanderbilt. In this study patients were treated with 1-2 weeks of letrozole prior to surgery, then the surgical specimen was analyzed for Ki67. The original goal of the study was to compare endocrine sensitive tumors with endocrine resistant tumors as a platform for discovery of resistance mechanisms to endocrine therapy. We retrospectively analyzed medical records for 143 patients in that study who had post-treatment Ki67 values available, and for whom BMI at the time of surgery could be obtained. We also reviewed medical records for components of the metabolic syndrome (hypertension, hyperglycemia or anti-hyperglycemic medications, low HDL, high triglycerides or statin use, and BMI &amp;gt; 30 as a surrogate for waist circumference). Metabolic syndrome was defined as 3 or more of the 5 components. For endocrine therapy response, a cutoff of 1 for the natural log Ki67 expression was defined as sensitive (approximately 2.5%) and tumors with natural log of Ki67 expression between 1 and 2 (approximately 7.5%) were categorized as intermediate sensitivity. Results:We did not find any significant association between BMI and recurrence score and the distribution of risk scores was similar across all BMI groups. We did not observe any correlation between RS and BMI. However we did observe a difference in the proportion of endocrine sensitive and resistant tumors in overweight and obese groups (only a minority of patients in this cohort, 18%, had normal weight). In patients with overweight, 70% of the tumors had a post-treatment Ki67 level classified as sensitive, whereas only 40% of the tumors in patients with obesity were endocrine sensitive (chi square = 0.0518). Similarly, 72% of patients without metabolic syndrome had sensitive tumors, but only 51% of patients with metabolic syndrome had sensitive tumors (relative risk 1.408, chi square = 0.0154). Conclusions:Our findings suggest that the association between obesity and/or metabolic syndrome and increased recurrence risk is not reflected in tumor-intrinsic properties present at the time of diagnosis. Rather, the effect of obesity and metabolic syndrome appears to be in response to treatment, at least as measured by a short-term surrogate marker of long-term endocrine sensitivity. Importantly, this suggests that interventions to reverse obesity and/or metabolic syndrome may be effective in improving outcomes for breast cancer recurrence, at least in part by improving sensitivity to endocrine therapy. Citation Format: Riley Bergman, Yvonne Berko, Brent Rexer. Association of factors of the obesity phenotype with breast cancer recurrence risk measured by oncotype recurrence score and short-term response to neoadjuvant endocrine therapy [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS5-18.

Credit Access and Mobility during the Flint Water Crisis

How do credit-constrained communities cope with the financial consequences of environmental crises? Beginning in April 2014, the residents of Flint, Michigan, were exposed to lead-contaminated water resulting from a series of governmental missteps. In this paper, we use the spatial distribution of lead and galvanized pipes in Flint to study the effect of the crisis on households’ financial health, including loan balances, repayment of outstanding debt, and Equifax Risk Scores, as well as on household mobility. We find that relatively more affected households, as measured by exposure to lead pipes, experienced a modest increase in the balance and frequency of past due loans. Equifax Risk Scores declined slightly on average, but more so at the bottom of the Risk Score distribution. In addition, we find that there was no effect on mobility out of the state or county, but that more affected households were more likely to move within the city when the crisis began, away from lead-pipe-dense areas.

Inclusion of medication-related fall risk in fall risk assessment tool in geriatric care units

BackgroundFalls are common undesirable events for older adults in institutions. Even though the patient’s fall risk may be scored on admission, the medication-induced fall risk may be ignored. This study developed a preliminary categorization of fall-risk-increasing drugs (FRIDs) to be added as a risk factor to the existing fall risk assessment tool routinely used in geriatric care units.MethodsMedication use data of older adults who had experienced at least one fall during a hospital ward or a nursing home stay within a 2-year study period were retrospectively collected from patient records. Medicines used were classified into three risk categories (high, moderate and none) according to the fall risk information in statutory summaries of product characteristics (SmPCs). The fall risk categorization incorporated the relative frequency of such adverse drug effects (ADEs) in SmPCs that were known to be connected to fall risk (sedation, orthostatic hypotension, syncope, dizziness, drowsiness, changes in blood pressure or impaired balance). Also, distribution of fall risk scores assessed on admission without considering medications was counted.ResultsThe fall-experienced patients (n = 188, 128 from the hospital and 60 from nursing home records) used altogether 1748 medicaments, including 216 different active substances. Of the active substances, 102 (47%) were categorized as high risk (category A) for increasing fall risk. Fall-experienced patients (n = 188) received a mean of 3.8 category A medicines (n = 710), 53% (n = 375) of which affected the nervous and 40% (n = 281) the cardiovascular system. Without considering medication-related fall risk, 53% (n = 100) of the patients were scored having a high fall risk (3 or 4 risk scores).ConclusionIt was possible to develop a preliminary categorization of FRIDs basing on their adverse drug effect profile in SmPCs and frequency of use in older patients who had experienced at least one documented fall in a geriatric care unit. Even though more than half of the fall-experienced study participants had high fall risk scores on admission, their fall risk might have been underestimated as use of high fall risk medicines was common, even concomitant use. Further studies are needed to develop the FRID categorization and assess its impact on fall risk.

Comparison of two algorithms to support medication surveillance for drug-drug interactions between QTc-prolonging drugs

BackgroundQTc-prolongation is an independent risk factor for developing life-threatening arrhythmias. Risk management of drug-induced QTc-prolongation is complex and digital support tools could be of assistance. Bindraban et al. and Berger et al. developed two algorithms to identify patients at risk for QTc-prolongation. ObjectiveThe main aim of this study was to compare the performances of these algorithms for managing QTc-prolonging drug-drug interactions (QT-DDIs). Materials and MethodsA retrospective data analysis was performed. A dataset was created from QT-DDI alerts generated for in- and outpatients at a general teaching hospital between November 2016 and March 2018. ECGs recorded within 7 days of the QT-DDI alert were collected. Main outcomes were the performance characteristics of both algorithms. QTc-intervals of > 500 ms on the first ECG after the alert were taken as outcome parameter, to which the performances were compared. Secondary outcome was the distribution of risk scores in the study cohort. ResultsIn total, 10,870 QT-DDI alerts of 4987 patients were included. ECGs were recorded in 26.2 % of the QT-DDI alerts. Application of the algorithms resulted in area under the ROC-curves of 0.81 (95 % CI 0.79–0.84) for Bindraban et al. and 0.73 (0.70–0.75) for Berger et al. Cut-off values of ≥ 3 and ≥ 6 led to sensitivities of 85.7 % and 89.1 %, and specificities of 60.8 % and 44.3 % respectively. ConclusionsBoth algorithms showed good discriminative abilities to identify patients at risk for QTc-prolongation when using ≥ 2 QTc-prolonging drugs. Implementation of digital algorithms in clinical decision support systems could support the risk management of QT-DDIs.

Epstein-Barr Virus-Based Nasopharyngeal Carcinoma (NPC) Risk Prediction Scores Are Elevated in NPC Multiplex Family Members in Taiwan.

Nasopharyngeal carcinoma (NPC) is caused by Epstein-Barr virus (EBV) and is more likely to occur in susceptible families. Whether genetic susceptibility operates through altered EBV control is incompletely understood. We used a NPC risk prediction model based on 14 EBV markers to compare risk score distribution in unaffected members from multiplex families with that in population-based controls. Despite the absence of NPC at the time of antibody measurement, we observed an upward shift in risk score among multiplex family members compared to the general population, consistent with the possibility that genetic factors affect NPC risk through alterations in EBV control.

Prognostic Value of Disease Risk Score Versus Gait Speed in Older Adults with Lymphoma

Background: Frailty is associated with poor tolerance to chemotherapy and mortality. While recent evidence suggests measures of frailty such as gait speed predict survival in older patients with hematologic malignancies such as lymphoma (Liu, Blood, 2019), its predictive ability has not been compared to disease-based prognostic risk scores. Methods: From February 2015 to April 2019, all patients aged 75 years and older who presented for initial consultation for at the Dana-Farber Cancer Institute for diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL) and Waldenstrom macroglobulinemia (WM) were approached for a screening frailty evaluation by a trained research assistant. As part of the evaluation, 4-meter gait speed was assessed (at normal pace with non-ambulatory patients recorded as having a speed of zero). Disease-specific prognostic scores were obtained retrospectively, including the International Prognostic Index (IPI) for DLBCL, the Follicular Lymphoma International Prognostic Index (FLIPI) for FL, and the International Prognostic Scoring System for Waldenstrom Macroglobulinemia (ISSWM). Univariate and multivariable Cox proportional hazard models were used to determine predictors of overall survival (OS; from date of screening geriatric assessment) including gait speed and disease-specific scores. Those with incomplete risk data excluded from disease-specific analyses. Hazard ratios and generalized R2 (higher the R2, stronger association with OS) were also calculated. Results: A total of 145 patients were included, 59 with DLBCL, 30 with FL, 56 with WM. Median age was 79 (IQR 77 to 82) and 42.1% were female. For the overall cohort, mean gait speed was 0.78 (±0.27). Risk score distribution and disease-specific gait speed are presented in the table. With a median follow-up of 20.9 months (IQR 10.2 to 34.5), median OS was 14.4 months for DLBCL, 17.2 months for FL and 32.0 months for WM. Disease-specific prognostic scores were not predictive of survival for DLBCL (p = 0.76) or FL (p = 0.22) but were for WM (p = 0.04). Overall, faster gait speed was significantly predictive of OS (HR 0.10, 95% CI 0.03 to 0.32, p &amp;lt; 0.0001) for all three lymphomas combined, which would mean an HR of 0.56 for an increase of 0.25 m/s of gait speed. In disease-specific analyses, faster gait speed was predictive of survival for DLBCL (HR 0.10, 95% CI 0.02 to 0.46, p = 0.003) and WM (HR 0.11, 95% CI 0.01 to 0.83, P = 0.03) but not for FL (P = 0.26). In DLBCL, gait speed explained significant variability in OS (R2 = 22.95%) compared with IPI score (R2 = 1.87%; p = 0.005) whereas it did not for WM compared with IPSSWM (R2 = 10.22% vs. 5.58%, p = 0.11) or FL compared with the FLIPI (R2= 12.21% vs. 8.10%, p = 0.35). Conclusion: Gait speed may help to further refine the prediction of outcomes in patients with aggressive lymphomas beyond standard prognostic scores but may have less of an effect for indolent lymphomas. These data suggest that gait speed should be incorporated into the standard assessment of patients with aggressive lymphomas. Table Disclosures Buckstein: Takeda: Research Funding; Celgene: Consultancy, Honoraria, Research Funding. Soiffer:Mana therapeutic: Consultancy; Kiadis: Other: supervisory board; Juno, kiadis: Membership on an entity's Board of Directors or advisory committees, Other: DSMB; Gilead, Mana therapeutic, Cugene, Jazz: Consultancy; Cugene: Consultancy; Jazz: Consultancy.

A genome-wide association study of prostate cancer in Latinos.

Latinos represent <1% of samples analyzed to date in genome‐wide association studies of cancer. The clinical value of genetic information in guiding personalized medicine in populations of non‐European ancestry will require additional discovery and risk locus characterization efforts across populations. In the present study, we performed a GWAS of prostate cancer (PrCa) in 2,820 Latino PrCa cases and 5,293 controls to search for novel PrCa risk loci and to examine the generalizability of known PrCa risk loci in Latino men. We also conducted a genetic admixture‐mapping scan to identify PrCa risk alleles associated with local ancestry. Genome‐wide significant associations were observed with 84 variants all located at the known PrCa risk regions at 8q24 (128.484–128.548) and 10q11.22 (MSMB gene). In admixture mapping, we observed genome‐wide significant associations with local African ancestry at 8q24. Of the 162 established PrCa risk variants that are common in Latino men, 135 (83.3%) had effects that were directionally consistent as previously reported, among which 55 (34.0%) were statistically significant with p < 0.05. A polygenic risk model of the known PrCa risk variants showed that, compared to men with average risk (25th–75th percentile of the polygenic risk score distribution), men in the top 10% had a 3.19‐fold (95% CI: 2.65, 3.84) increased PrCa risk. In conclusion, we found that the known PrCa risk variants can effectively stratify PrCa risk in Latino men. Larger studies in Latino populations will be required to discover and characterize genetic risk variants for PrCa and improve risk stratification for this population.

Compliance Effects of Risk-Based Tax Audits

Tax administrations use machine learning to predict risk scores as a basis for selecting individual taxpayers for audit. Audits detect noncompliance immediately, but may also alter future filing behavior. This analysis is the first to estimate compliance effects of audits among high-risk wage earners. We exploit a sharp audit assignment discontinuity in Norway based on individual tax payers risk score. Additional data from a random audit allow us to estimate how the audit effect vary across the risk score distribution. We show that the current risk score audit threshold is set far above the one that maximizes net public revenue.

Prognostic Biomarkers Converge on a Germinal Centre Dark Zone Phenotype As a Determinant of Adverse Outcome in Follicular Lymphoma Patients Treated with Rituximab and Chemotherapy

Introduction: Follicular lymphoma (FL) is a clinically and genetically heterogeneous disease with highly variable patient outcomes. Recently, Huet et al. proposed a 23-gene expression-based risk score for predicting progression-free survival (PFS) in FL patients treated with rituximab and chemotherapy (Huet et al. Lancet Oncology 2018). The m7-FLIPI risk score has also been described as a clinico-genetic model predicting patient outcomes (Pastore et al. Lancet Oncology 2015). Moreover, EZH2 wild-type status and high expression of the FOXP1 transcription factor are associated with increased risk of lymphoma progression (Mottok et al. Blood 2018). This multitude of prognostic tools in FL raises the question whether they identify common biology. The aims of this study were to assess whether the 23-gene predictor score identifies a poor risk group of patients in our own gene expression dataset, and whether commonality exists between the 23-gene score, the m7-FLIPI, EZH2 mutation status and FOXP1 expression.Methods: In our previous work, we generated Illumina DASL microarray expression profiles for 137 FL patients who were treated with rituximab and CVP chemotherapy (cyclophosphamide, vincristine and prednisone). Using genes from the 23-gene linear risk predictor, we determined each patient's risk score by setting coefficients at -1 and +1 for genes associated with favorable and unfavorable PFS, respectively. We dichotomized the distribution of scores using the maximally selected log-rank statistic. We also performed unsupervised, hierarchical clustering to identify underlying subgroups in an unbiased fashion. Survival analyses were performed using the log-rank test and Cox regression analyses. We used gene set enrichment analysis to identify concordant differences of relevant gene signatures between specimens with either low or high expression of FOXP1.Results: Twenty genes from the 23-gene predictor (87%) were identified in the DASL gene expression dataset. The coefficients from univariate Cox regression analysis from our data were correlated with coefficients from Huet et al. (Pearson r = .7, P < .001; Spearman r = .44, P = .051). All poor-risk genes from the 23-gene predictor were associated with poor PFS in our data, and vice versa. Concordantly, calculated risk scores were significantly associated with PFS in the univariate Cox regression analysis (P = .007). Dichotomizing the distribution of risk scores identified 68% of cases with high risk score who had inferior PFS and OS compared to 32% of cases with low risk score (5-year PFS 54% vs. 77%, P = .004; 5-year OS 73% vs. 86%, P = .04). Hence, the risk score stratified patients into groups with diverging outcomes. This association was found to be independent of the Follicular Lymphoma Prognostic Index (FLIPI). In addition, the mean risk score was significantly higher in cases with high expression of FOXP1 (P < .001) and in cases with high m7-FLIPI risk score (P = .023). Unsupervised hierarchical clustering identified two main clusters (“cluster 1” and “cluster 2”) that were characterized by low and high expression of genes associated with poor outcome, respectively. Patients from “cluster 2” experienced worse PFS compared to patients in “cluster 1” (P = .046; 5-year PFS 54% vs. 68%). The 5-year OS was 72% for patients in “cluster 2”, vs. 81% in “cluster 1” (P = .13).We have previously reported that a germinal centre dark zone signature is enriched in cases with high FOXP1 expression, and the ICA13 signature reported by Huet et al. has been described as being highly expressed in centroblasts. Using gene set enrichment analysis, we found that genes with positive weight and coefficients in the ICA13 and the 23-gene predictor score, respectively, were enriched in the FOXP1-high phenotype (adjusted P = .009 and .005, respectively). GeneMANIA illustrated co-expression interconnectivity among ORAI2, TCF4, AFF3, FOXO1, CXCR4 and FOXP1, suggesting that genes with prognostic significance operate in tightly regulated networks.Conclusions: Our results exemplify the robustness of the predictor model by Huet et al. Further, we demonstrate biomarker convergence on a common phenotype: FOXP1 expression, EZH2 wild-type status and expression of dark zone-related genes, which characterize a subset of FL cases with adverse outcome following rituximab and chemotherapy. DisclosuresSarkozy:Roche/Genentech: Consultancy. Sehn:Roche/Genentech: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Karyopharm: Consultancy, Honoraria; Lundbeck: Consultancy, Honoraria; Seattle Genetics: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Abbvie: Consultancy, Honoraria; Celgene: Consultancy, Honoraria; TG Therapeutics: Consultancy, Honoraria; Merck: Consultancy, Honoraria; Morphosys: Consultancy, Honoraria. Weigert:Novartis: Research Funding; Roche: Research Funding. Steidl:Juno Therapeutics: Consultancy; Tioma: Research Funding; Bristol-Myers Squibb: Research Funding; Roche: Consultancy; Seattle Genetics: Consultancy; Nanostring: Patents & Royalties: patent holding.

Association between the Risk of Low/High-Count Monoclonal B-Cell Lymphocytosis (MBL) and the Chronic Lymphocytic Leukemia (CLL) Polygenic Risk Score (PRS)

BackgroundMBL is a precursor state to CLL, with a prevalence of ~5-8% in the general population and 15-18% in CLL families. To date, 41 single nucleotide polymorphisms (SNPs) have been found to be associated with CLL. We previously showed that a PRS of the weighted average of the number of risk alleles of these SNPs is associated with CLL risk using cases and controls from the International Lymphoma Epidemiology (InterLymph) Consortium. We validated this score in an independent sample of CLL cases and controls from the Genetic Epidemiology of CLL (GEC) Consortium, a cohort of families with ≥2 members with CLL. In the CLL families we also reported an association between the PRS and MBL risk, where 93% of MBLs were low-count MBL. Here we evaluate the PRS in an independent sample ascertained from the Mayo Clinic Biobank.MethodsThe Mayo Clinic Biobank is a large-scale bio-repository of adult patients assembled to provide a wide array of health-related research studies. Biobank participants (N=2530) were screened for MBL using a highly-sensitive, 8-color (CD38, CD45, Kappa, Lambda, CD19, CD23, CD5 and CD20) flow-cytometry assay on stored cryopreserved peripheral blood mononucleotide cells. Low-count and high-count MBL were defined as those individuals who had <85% and ≥85 clonal B-cells out of total B-cells, respectively. Individuals without MBL (“controls”) were frequency matched to MBLs based on age and sex, after excluding non-Caucasian and non-interpretable flow results. Participants were genotyped using the Illumina OmniExpress array. Standard genotyping quality-control metrics were applied. We computed the PRS as previously published and categorized it by quintiles based on values previously used with InterLymph controls. The middle quintile served as the reference category. We used logistic regression to estimate odds ratios (ORs) and 95% confidence intervals (CI), adjusting for age at consent, sex, and socioeconomic status. In addition, we compared the median PRS by low-count MBL and high count MBL using the Mann-Whitney U test.ResultsWe identified 276 MBLs overall and 2109 controls; 97% (N=267) of MBLs had low-count MBL and only 3% (N=9) had high-count MBL. We then selected 844 controls frequency matched to the MBL cases on age and sex. Family history of leukemia/lymphoma was 11.3% among MBLs overall and 8.8% among controls. The median PRS was 7.88 and 7.46 (Figure) among MBLs overall and controls, respectively. The continuous PRS had a 1.75-fold increased risk for MBL overall (CI:1.5-2.1, P=2.65x10-12). Compared to the middle quintile, the highest quintile had 2.2-fold increased risk for MBL overall (CI:1.5-3.3, P=1.46x10-4) and the lowest quintile had 0.6-fold decreased risk (CI:0.4-0.96, P=0.03). The median PRS for low-count and high count MBL were 7.86 and 8.04, respectively. The continuous PRS had a 1.75-fold increased risk for low-count MBL (CI:1.5-2.1, P=4.81x10-12) and a 1.89-fold increased risk for high-count MBL (CI:0.8-4.2, P=0.13), however, the effect is not statistically significant between high-count and low-count MBL (P=0.89).ConclusionIn this independent sample, we validated our previous findings that the CLL PRS is associated with MBL risk. Although the sample size of high-count MBL was limited (N=9), there is evidence of a weak trend towards a higher PRS compared to low-count MBL and controls Larger sample size of high-count MBL is warranted in order to stratify MBL risk by low-count and high-count MBL. These results may help identify individuals at higher risk of MBL including low-count and high-count MBL, beyond the risk associated with age and family history of CLL.Figure: Polygenic risk score distribution by control and MBL status DisclosuresParikh:Janssen: Research Funding; Pharmacyclics: Honoraria, Research Funding; Gilead: Honoraria; Abbvie: Honoraria, Research Funding; AstraZeneca: Honoraria, Research Funding; MorphoSys: Research Funding. Cerhan:Nanostring: Research Funding; Celgene: Research Funding; Jannsen: Other: Scientific Advisory Board. Kay:Janssen: Membership on an entity's Board of Directors or advisory committees; Acerta: Research Funding; Infinity Pharm: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees; Agios Pharm: Membership on an entity's Board of Directors or advisory committees; Pharmacyclics: Membership on an entity's Board of Directors or advisory committees, Research Funding; Cytomx Therapeutics: Membership on an entity's Board of Directors or advisory committees; Gilead: Membership on an entity's Board of Directors or advisory committees; Tolero Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees, Research Funding; Morpho-sys: Membership on an entity's Board of Directors or advisory committees. Shanafelt:Genentech: Research Funding; Pharmacyclics: Research Funding; GlaxoSmithKline: Research Funding; Jansen: Research Funding.

Identifying risk factors for the development of sepsis during adult severe malaria

BackgroundSevere falciparum malaria can be compounded by bacterial sepsis, necessitating antibiotics in addition to anti-malarial treatment. The objective of this analysis was to develop a prognostic model to identify patients admitted with severe malaria at higher risk of developing bacterial sepsis.MethodsA retrospective data analysis using trial data from the South East Asian Quinine Artesunate Malaria Trial. Variables correlating with development of clinically defined sepsis were identified by univariable analysis, and subsequently included into a multivariable logistic regression model. Internal validation was performed by bootstrapping. Discrimination and goodness-of-fit were assessed using the area under the curve (AUC) and a calibration plot, respectively.ResultsOf the 1187 adults with severe malaria, 86 (7.3%) developed clinical sepsis during admission. Predictors for developing sepsis were: female sex, high blood urea nitrogen, high plasma anion gap, respiratory distress, shock on admission, high parasitaemia, coma and jaundice. The AUC of the model was 0.789, signifying modest differentiation for identifying patients developing sepsis. The model was well-calibrated (Hosmer–Lemeshow Chi squared = 1.02). The 25th percentile of the distribution of risk scores among those who developed sepsis could identify a high-risk group with a sensitivity and specificity of 70.0 and 69.4%, respectively.ConclusionsThe proposed model identifies patients with severe malaria at risk of developing clinical sepsis, potentially benefiting from antibiotic treatment in addition to anti-malarials. The model will need further evaluation with more strictly defined bacterial sepsis as outcome measure.

Abstract 227: Germline variation at 8q24 and prostate cancer risk in men of European ancestry

Abstract We performed an in-depth and well-powered investigation of genetic variation across the cancer susceptibility region at chromosome 8q24 (127.6-129.0 Mb) to search for novel risk variants associated with prostate cancer (PCa) risk in the European ancestry population. We combined genotyped and imputed data from the PRACTICAL/ELLIPSE OncoArray and iCOGS consortia consisting of 71,535 PrCa cases and 52,935 controls of European ancestry. Variants with high imputation quality score (&amp;gt;0.8) were retained for a total of 5,600 overlapping variants between the two datasets. Associations of genetic variants with PCa risk were evaluated using unconditional logistic regression with adjustment for country and ten principal components. The marginal risk estimates for the 5,600 variants that passed quality control were combined by a fixed effects meta-analysis. A meta-stepwise selection was performed on variants marginally associated with PCa risk from the meta results (P&amp;lt;0.05). A polygenic risk score and the contribution to the familial relative risk of PCa were estimated for variants from the final model. Of the 5,600 variants at 8q24 retained for analysis, 1,268 (23%) were associated with PCa risk at P&amp;lt;5x10-8 while 2,772 (49%) were marginally associated at P&amp;lt;0.05. In the stepwise model, 12 variants remained statistically significantly associated with PCa risk with conditional meta p-values between 2.93x10-137 and 4.28x10-15. The independent stepwise signals were confirmed by Joint Analysis of Marginal (JAM) summary statistics, which defined the credible sets of variants driving those signals. Three of the variants (rs1914295, rs190257175, rs12549761) were weakly correlated (r2≤0.17) with any known PCa risk marker, and may define novel association signals. Men in the top 1% of the polygenic risk score distribution had a 3.97-fold relative risk (95%CI=3.87-4.07) compared to men with "average risk" (25th-75th percentiles). The 12 independent signals at 8q24 capture 11.54% (95%CI=9.86-13.65) of the familial relative risk of PCa, which is approximately one quarter of the total PCa familial relative risk explained by known genetic risk factors. Most of the independently associated signals have good evidence for biologic functionality; in particular, many reside within putative transcriptional enhancers and/or binding sites for AR and FOXA1 transcription factors in prostate cell lines. In summary, we defined 12 independent association signals among men of European ancestry, with three of the risk variants representing putative novel association signals. Whereas the individual associations of these variants with PCa risk are relatively modest (ORs&amp;lt;2.0), their cumulative effects are substantial, and their contribution to the overall familial relative risk of PCa is substantially greater than any other known prostate cancer risk locus. Citation Format: Marco Matejcic, Edward J. Saunders, Tokhir Dadaev, Mark Brook, Ali Amin Al Olama, Fredrick R. Schumacher, Sonja I. Berndt, Sara Benlloch, Kenneth Muir, Koveela Govindasami, Victoria L. Stevens, Susan M. Gapstur, Catherine M. Tangen, Jyotsna Batra, Judith Clements, APCB (Australian Prostate Cancer Bio Resource), Henrik Gronberg, Nora Pashayan, Johanna Schleutker, Demetrius Albanes, Alicja Wolk, Catharine West, Lorelei Mucci, Peter Kraft, Géraldine Cancel-Tassin, Stella Koutros, Karina Dalsgaard Sorensen, Lovise Maehle, Eli Marie Grindedal, Sara Strom, David E. Neal, Freddie C. Hamdy, Jenny L. Donovan, Ruth C. Travis, Robert J. Hamilton, Sue Ann Ingles, Barry Rosenstein, Yong-Jie Lu, Graham G. Giles, Adam S. Kibel, Ana Vega, Jeanette Bensen, Manolis Kogevinas, Fredrik Wiklund, Stephen Chanock, Douglas F. Easton, Rosalind A. Eeles, Zsofia Kote-Jarai, David V. Conti, Christopher A. Haiman. Germline variation at 8q24 and prostate cancer risk in men of European ancestry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 227.

Abstract P1-06-09: Correlation of breast cancer index (BCI) results to lymphovascular invasion in early stage HR+ breast cancer

Abstract Background: Positive lymphovascular invasion (LVI) is a negative prognostic factor for women with early-stage ER+ breast cancer. LVI, along with other clinicopathologic factors such as larger tumor size, higher grade and positive nodal status, increase a patient's risk of late (post-5y) recurrence. Breast Cancer Index (BCI) is a validated gene expression-based assay for patients with early-stage HR+ breast cancer that reports an individualized risk of late distant recurrence based on a combination of the HOXB13/I17BR ratio and the molecular grade index (MGI). The correlation of LVI and individualized risk stratification by genomic analysis is not well characterized. Therefore, this study evaluated risk stratification by BCI based on the presence or absence of LVI. Methods: A population of 2,613 patients with known LVI status were identified in the Breast Cancer Index Clinical Database for Correlative Studies, an IRB-approved de-identified database which contains clinicopathologic and molecular variables of more than 19,000 clinical cases submitted for BCI testing. LVI was recorded as either positive or negative based on pathology report review. BCI results based on LVI status from LN- (n=2035) and LN+ patients (n=578) were evaluated separately. Chi-squared tests were used to compare BCI results between LVI groups. Results: In analyses of 2,613 patients with LVI data available (median age 59.1 y; range 28-89y; 74% ≥50y), 18.3% of patients showed evidence of LVI (LVI-pos). In comparison to the LVI-neg tumors submitted for BCI testing, the LVI-pos tumors had a higher proportion of grade 3 tumors (33% vs 16%, p&amp;lt;0.0001), more LVI-pos tumors were 2.0 cm or greater (45% vs 23%, p&amp;lt;0.0001), a higher percentage LVI-pos patients had node-positive disease (51% vs 16%, p&amp;lt;0.0001), and a higher proportion of LVI-pos tumors showed high Ki67 (Ki67 ≥14%; 64% vs 51%, p=0.004). A correlation between LVI positivity and high BCI prognostic risk was observed, with a higher proportion of LVI-pos patients classified as high risk of late distant recurrence in both the LN- (68% vs 49%, p&amp;lt;0.0001) and LN+ subsets (84% vs 70%, p&amp;lt;0.0001) compared to LVI-neg patients. LVI-pos patients had a higher median molecular proliferative status (MGI) compared to LVI-neg patients regardless of nodal status (p&amp;lt;0.0001 for both). In contrast to the categorical LVI prognostic factor, the wide distribution of BCI individual risk scores provides additional resolution that identifies a substantial subset of LVI positive tumors (32%) that that have a low risk of late recurrence by genomic analysis. Conclusion: While BCI Prognostic stratification correlated with LVI status, BCI identified a subset of patients with LVI positivity as having a low risk of late distant recurrence that otherwise would have an unfavorable prognosis based on LVI and/or LN positivity. These findings help to characterize differential patient stratification based on an individualized assessment of tumor biology versus LVI for patients considering EET. Citation Format: Vulchi M, Sagalnik M, Schnabel CA, Abraham J. Correlation of breast cancer index (BCI) results to lymphovascular invasion in early stage HR+ breast cancer [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P1-06-09.

Predicting nosocomial lower respiratory tract infections by a risk index based system

Although belonging to one of the most common type of nosocomial infection, there was currently no simple prediction model for lower respiratory tract infections (LRTIs). This study aims to develop a risk index based system for predicting nosocomial LRTIs based on data from a large point-prevalence survey. Among the 49328 patients included, the prevalence of nosocomial LRTIs was 1.70% (95% confidence interval [CI], 1.64% to 1.76%). The areas under the receiver operating characteristic (ROC) curve for logistic regression and fisher discriminant analysis were 0.907 (95% CI, 0.897 to 0.917) and 0.902 (95% CI, 0.892 to 0.912), respectively. The constructed risk index based system also displayed excellent discrimination (area under the ROC curve: 0.905 [95% CI, 0.895 to 0.915]) to identify LRTI in internal validation. Six risk levels were generated according to the risk score distribution of study population, ranging from 0 to 5, the corresponding prevalence of nosocomial LRTIs were 0.00%, 0.39%, 3.86%, 12.38%, 28.79% and 44.83%, respectively. The sensitivity and specificity of prediction were 0.87 and 0.79, respectively, when the best cut-off point of risk score was set to 14. Our study suggested that this newly constructed risk index based system might be applied to boost more rational infection control programs in clinical settings.

Risk Factors for Nocturnal Hypoglycemia in Insulin-treated Patients With Type 2 Diabetes: A Secondary Analysis of Observational Data Derived From an Integrated Clinical Trial Database

PurposeA trade-off exists in most diabetes therapies between the benefits of good glycemic control and the morbidity of hypoglycemia. Balancing these factors to achieve desired outcomes is a key consideration for personalized diabetes therapy. Hypoglycemia at night (nocturnal hypoglycemia [NH]) is a common but often under-reported problem in insulin-treated patients with type 2 diabetes. To better understand the risk for NH, we pooled data from multiple clinical trials of insulin treatment and specifically examined NH risk factors in relation to glycemic goals. MethodsOf 53 randomized trials involving insulin treatment, 18 trials that collected NH data were included. Risk factors associated with NH were identified by using gradient-boosting methods. A proportional hazards model was used to quantify the hazard ratio (HR) for risk factors. By modeling with individual patient data, a patient-level NH risk score distribution was created. Finally, results of the model were used to quantify an adjustment to the glycemic goal that would fully offset each risk factor, all other factors being equal. FindingsData pooling resulted in the inclusion of 7341 patients with type 2 diabetes from 18 randomized clinical trials. In the mean 6-month treatment period, 43% of patients experienced at least 1 episode of NH (mean [SD], 1.1 [1.5] events/month). Reduction of glycosylated hemoglobin (HbA1c) levels during the trial was a risk factor for NH (HR, 1.40 [95% CI, 1.38-1.43] per –1% of HbA1c). Higher baseline HbA1c level was a protective factor against NH (HR, 0.76 [95% CI, 0.74–0.77] per +1% of HbA1c); and the adjustment to HbA1c goal required to offset 1% higher baseline HbA1c was –0.825%. Patient characteristics for risk of NH included older age (HR, 1.02 [95% CI, 1.01–1.02]) per 1-year increase), female sex (HR, 1.18 [95% CI, 1.15–1.22]), black or African-American race (HR, 1.41 [95% CI, 1.33–1.50] vs white race), longer diabetes duration (HR, 1.02 [95% CI, 1.01–1.02] per 1-year increase), diabetic nephropathy (HR, 1.40 [95% CI, 1.27–1.54]), and concomitant sulfonylurea use (HR, 1.10 [95% CI, 1.05–1.15]). Asian race was associated with a lower risk of NH (HR, 0.50 [95% CI, 0.48–0.53] vs white race); this finding could be offset with a 2.03% adjustment to the HbA1c goal. ImplicationsData on NH are scarce. By pooling multiple clinical trials, this study was able to evaluate patient-level data. A quantitative understanding of the trade-off between individual risk factors for NH and glycemic reduction may help clinicians to personalize patients’ glycemic goals, while effectively managing NH risk. Limitations of the study include that patients were selected through inclusion/exclusion criteria and that patient compliance may be better in a trial setting. Validating the findings in the real world will be helpful.

Quality of Life Scores Improve with Increasing Hemoglobin but Optimal Thresholds Vary According to Transfusion Dependence and Clinical Risk Scores: A Canadian Cross Sectional Study of 689 Patients with 2969 Measurements

Quality of Life Scores Improve with Increasing Hemoglobin but Optimal Thresholds Vary According to Transfusion Dependence and Clinical Risk Scores: A Canadian Cross Sectional Study of 689 Patients with 2969 Measurements