Faculty Bibliography

Breast cancer screening using Mammography serves as the earliest defense against breast cancer, revealing anomalous tissue years before it can be detected through physical screening. Despite the use of high resolution radiography, the presence of densely overlapping patterns challenges the consistency of human-driven diagnosis and drives interest in leveraging state-of-art localization ability of deep convolutional neural networks (DCNN). The growing availability of digitized clinical archives enables the training of deep segmentation models, but training using the most widely available form of coarse hand-drawn annotations works against learning the precise boundary of cancerous tissue in evaluation, while producing results that are more aligned with the annotations rather than the underlying lesions. The expense of collecting high quality pixel-level data in the field of medical science makes this even more difficult. To surmount this fundamental challenge, we propose LatentCADx, a deep learning segmentation model capable of precisely annotating cancer lesions underlying hand-drawn annotations, which we procedurally obtain using joint classification training and a strict segmentation penalty. We demonstrate the capability of LatentCADx on a publicly available dataset of 2,620 Mammogram case files, where LatentCADx obtains classification ROC of 0.97, AP of 0.87, and segmentation AP of 0.75 (IOU = 0.5), giving comparable or better performance than other models. Qualitative and precision evaluation of LatentCADx annotations on validation samples reveals that LatentCADx increases the specificity of segmentations beyond that of existing models trained on hand-drawn annotations, with pixel level specificity reaching a staggering value of 0.90. It also obtains sharp boundary around lesions unlike other methods, reducing the confused pixels in the output by more than 60%.

Introduction/UNASSIGNED:Smoking cessation after a cancer diagnosis can significantly improve treatment outcomes and reduce the risk of cancer recurrence and all-cause mortality. Aim/UNASSIGNED:We sought to measure the association between cancer diagnosis and subsequent smoking cessation. Methods/UNASSIGNED:. Our sample was composed of 7,286 adult smokers at the baseline representing an estimated 40.9 million persons. Smoking cessation rates after a diagnosis differed after a tobacco-related cancer (25.9%), a nontobacco-related cancer (8.9%), and no cancer diagnosis (17.9%). After adjustment, diagnosis with a tobacco-related cancer was associated with a higher odds of smoking cessation (OR 1.83, 95% CI 1.00-3.33) compared to no cancer diagnosis. Diagnosis with a nontobacco-related cancer was not significantly linked to smoking cessation (OR 0.52, 95% CI 0.48-1.45). Conclusion/UNASSIGNED:Diagnosis with a tobacco-related cancer is associated with greater odds of subsequent smoking cessation compared to no cancer diagnosis, suggesting that significant behavioral change may occur in this setting.

Although racial/ethnic disparities in police contact are well documented, less is known about other dimensions of inequity in policing. Sexual minority groups may face disproportionate police contact. We used data from the P18 Cohort Study (Version 2), a study conducted to measure determinants of inequity in STI/HIV risk among young sexual minority men (YSMM) in New York City, to measure across-time trends, racial/ethnic disparities, and correlates of self-reported stop-and-frisk experience over the cohort follow-up (2014-2019). Over the study period, 43% reported stop-and-frisk with higher levels reported among Black (47%) and Hispanic/Latinx (45%) than White (38%) participants. Stop-and-frisk levels declined over follow-up for each racial/ethnic group. The per capita rates among P18 participants calculated based on self-reported stop-and-frisk were much higher than rates calculated based on New York City Police Department official counts. We stratified respondents' ZIP codes of residence into tertiles of per capita stop rates and observed pronounced disparities in Black versus White stop-and-frisk rates, particularly in neighborhoods with low or moderate levels of stop-and-frisk activity. YSMM facing the greatest economic vulnerability and mental disorder symptoms were most likely to report stop-and-frisk. Among White respondents levels of past year stop-and-frisk were markedly higher among those who reported past 30 day marijuana use (41%) versus those reporting no use (17%) while among Black and Hispanic/Latinx respondents stop-and-frisk levels were comparable among those reporting marijuana use (38%) versus those reporting no use (31%). These findings suggest inequity in policing is observed not only among racial/ethnic but also sexual minority groups and that racial/ethnic YSMM, who are at the intersection of multiple minority statuses, face disproportionate risk. Because the most socially vulnerable experience disproportionate stop-and-frisk risk, we need to reach YSMM with community resources to promote health and wellbeing as an alternative to targeting this group with stressful and stigmatizing police exposure.

Background/UNASSIGNED:Patients with head and neck cancer (HNC) experience painful, debilitating symptoms and functional limitations that can interrupt cancer treatment, and decrease their health-related quality of life (HRQoL). The Electronic Patient Visit Assessment (ePVA) for head and neck is a web-based mHealth patient-reported measure that asks questions about 21 categories of symptoms and functional limitations common to HNC. This article presents the development and usefulness of the ePVA as a clinical support tool for real-time interventions for patient-reported symptoms and functional limitations in HNC. Methods/UNASSIGNED:Between January 2018 and August 2019, 75 participants were enrolled in a clinical usefulness study of the ePVA. Upon signing informed consent, participants completed the ePVA and the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire (QLQ) general (C30) questionnaire v3.0 (scores range from 0 to 100 with 100 representing best HRQoL). Clinical usefulness of the ePVA was defined as demonstration of reliability, convergent validity with HRQoL, and acceptability of the ePVA (i.e., >70% of eligible participants complete the ePVA at two or more visits and >70% of ePVA reports are read by providers). Formal focus group discussions with the interdisciplinary team that cared for patients with HNC guided the development of the ePVA as a clinical support tool. Qualitative and quantitative methods were used throughout the study. Descriptive statistics consisting of means and frequencies, Pearson correlation coefficient, and Student's t-tests were calculated using SAS 9.4 and STATA. Results/UNASSIGNED:The participants were primarily male (71%), White (76%), diagnosed with oropharyngeal or oral cavity cancers (53%), and undergoing treatment for HNC (69%). Data analyses supported the reliability (alpha =0.85), convergent validity with HRQoL scores, and acceptability of the ePVA. Participants with the highest number of symptoms and functional limitations reported significantly worse HRQoL (sum of symptoms: r=-0.50, P<0.0001; sum of function limitations: r=-0.56, P<0.0001). Ninety-two percent of participants (59 of 64) who had follow-up visits within the 6-month study period completed the ePVA at two or more visits and providers read 89% (169 of 189) of automated ePVA reports. The use of the ePVA as a clinical support tool for real-time interventions for symptoms and functional limitations reported by patients is described in a clinical exemplar. Conclusions/UNASSIGNED:This research indicates that the ePVA may be a useful mHealth tool as a clinical support tool for real-time interventions for patient-reported symptoms and functional limitations in HNC. The study findings support future translational research to enhance the usefulness of the ePVA in real world settings for early interventions that decrease symptom burden and improve the QoL of patients with HNC.

BACKGROUND:Women who have experienced adolescent pregnancy and early motherhood are disproportionately affected in terms of their health and parenting capabilities, as well as their offspring's health. Guided by Stress Process and Social Determinants of Health (SDH) Frameworks, which posit that multiple sources of stressors and structural determinants of adolescent pregnancy influence adolescent mothers' subsequent health and quality of parenting (Xavier et al 2018, McLoyd 1998, Conger et al 2010, Gipson et al 2008). These dynamics then further impact offspring's outcomes. Using an Integrated Stress-SDH Process for Health Disparities model and we test on whether early motherhood is associated with and subsequent maternal and child health from two informal settlements in Nairobi. METHODS:A cross-sectional design with 394 mothers of 2-16 years old children who sought maternal and child health services at Kariobangi and Kangemi public health centers between October 2015 to April 2016 were recruited. Participating mothers were asked questions related to their adolescent pregnancy history, their current health, wellbeing, and parenting practices, and their child's health. Structural equation modeling (SEM) was utilized to examine hypothesized mediational pathways that adolescent pregnancy history has negative influences on women's health and parenting during adulthood, which also influence their child's health and development. RESULTS:Our study supports that women with a history of adolescent motherhood have poor physical and mental health outcomes as adults after adjusting for demographic confounders. SEM results partially support the Stress-SDH Process model that history of adolescent pregnancy had negative consequences on women's adulthood health, which also negatively impacted offspring's physical and mental health. CONCLUSION/CONCLUSIONS:Consistent with the Stress Process and SDH literature, we found consistent cross-cultural literature that adolescent pregnancy set the stage for, subsequent poor maternal health and child outcomes. Although history of adolescent pregnancy and motherhood was not necessarily associated with negative parenting, consistent with parenting literature, negative parenting was associated with poor child mental health. Findings suggest importance of providing integrated care that address health and parenting needs to optimize offspring's development in instances of early motherhood.

INTRODUCTION:Hemostasis depends on the delicate balance between coagulants and anticoagulants. Higher levels of circulating coagulants have been associated with higher risk of cerebral infarctions and dementia. In contrast, higher levels of circulating protein C, an endogenous anticoagulant, have been associated with lower risk of cerebral infarctions, and the association between protein C levels and the risk of dementia is unknown. The goal of this study was to evaluate the association of circulating protein C levels in midlife and late life with incident dementia. METHODS:Circulating protein C levels were measured using blood samples collected at the midlife baseline (1987-1989) and the late-life baseline (2011-2013) among 14,462 and 3,614 participants, respectively, in the Atherosclerosis Risk in Communities study. Protein C levels were measured using enzyme-linked immunosorbent assay at midlife and a modified aptamer-based assay at late life. Participants were followed up to 2013 from midlife and up to 2017 from late life. Incident dementia was ascertained during the follow-up periods using in-person cognitive and functional assessment, informant interviews, and International Classification of Diseases codes at hospitalization discharge and on death certificates. Cause-specific Cox regression models were used to evaluate the association between quintiles of circulating protein C and incident dementia. RESULTS:From midlife (mean age of 54), 1,389 incident dementia events were observed over a median follow-up of 23 years. From late life (mean age of 75), 353 incident dementia events were observed over a median follow-up of 4.9 years. At both midlife and late life, circulating protein C had an inverse association with incident dementia after adjusting for demographic, vascular, and hemostatic risk factors, incident stroke as time-dependent covariate, and incorporating stabilized weights based on propensity scores (quintile 5 vs. quintile 1 as the reference, midlife hazard ratio 0.80, 95% confidence interval 0.66-0.96, p value for trend 0.04; late-life hazard ratio 0.84, 95% confidence interval: 0.55-1.28, p value for trend 0.04). DISCUSSION/CONCLUSION:Circulating protein C has an inverse association with incident dementia independent of established risk factors, including stroke. Our results suggest studying anticoagulants in addition to coagulants can increase our understanding on the relationship between hemostasis and dementia.

Introduction/UNASSIGNED:Periodontal disease is a chronic, inflammatory bacterial dysbiosis that is associated with both Alzheimer's disease (AD) and Down syndrome. Methods/UNASSIGNED:A total of 48 elderly cognitively normal subjects were evaluated for differences in subgingival periodontal bacteria (assayed by 16S rRNA sequencing) between cerebrospinal fluid (CSF) biomarker groups of amyloid and neurofibrillary pathology. A dysbiotic index (DI) was defined at the genus level as the abundance ratio of known periodontal bacteria to healthy bacteria. Analysis of variance/analysis of covariance (ANOVA/ANCOVA), linear discriminant effect-size analyses (LEfSe) were used to determine the bacterial genera and species differences between the CSF biomarker groups. Results/UNASSIGNED: = 0.02 and 0.01) but not with P-tau. Discussion/UNASSIGNED:We show a selective relationship between periodontal disease bacterial dysbiosis and CSF biomarkers of amyloidosis, but not for tau. Further modeling is needed to establish the direct link between oral bacteria and Aβ.

Background and Objective/UNASSIGNED:Posttraumatic stress disorder (PTSD) is a serious and frequently debilitating psychiatric condition that can occur in people who have experienced traumatic stessors, such as war, violence, sexual assault and other life-threatening events. Treatment of PTSD and traumatic brain injury (TBI) in veterans is challenged by diagnostic complexity, partially due to PTSD and TBI symptom overlap and to the fact that subjective self-report assessments may be influenced by a patient's willingness to share their traumatic experiences and resulting symptoms. Corticotropin-releasing factor (CRF) is one of the main mediators of hypothalamic pituitary adrenal (HPA)-axis responses in stress and anxiety. Methods and Results/UNASSIGNED:We analyzed serum CRF levels in 230 participants including heathy controls (64), and individuals with PTSD (53), TBI (70) or PTSD+TBI (43) by enzyme immunoassay (EIA). Significantly lower CRF levels were found in both the PTSD and PTSD+TBI groups compared to healthy control (PTSD vs Controls: P=0.0014, PTSD + TBI vs Controls: P=0.0011) and chronic TBI participants (PTSD vs TBI: P<0.0001PTSD + TBI vs TBI: P<0.0001) , suggesting a PTSD-related mechanism independent from TBI and associated with CRF reduction. CRF levels negatively correlated with PTSD severity on the CAPS-5 scale in the whole study group. Conclusions/UNASSIGNED:Hyperactivation of the HPA axis has been classically identified in acute stress. However, the recognized enhanced feedback inhibition of the HPA axis in chronic stress supports our findings of lower CRF in PTSD patients. This study suggests that reduced serum CRF in PTSD should be further investigated. Future validation studies will establish if CRF is a possible blood biomarker for PTSD and/or for differentiating PTSD and chronic TBI symptomatology.