Faculty Bibliography

The associations between covariates and the outcomes often vary over time, regardless of whether the covariate is time-varying or time-invariant. For example, we hypothesize that the impact of chronic diseases, such as diabetes and heart disease, on people's physical functions differ with aging. However, the age-varying effect would be missed if one models the covariate simply as a time-invariant covariate (yes/no) with a time-constant coefficient. We propose a fused lasso-based time-varying linear mixed effect (FTLME) model and an efficient two-stage parameter estimation algorithm to estimate the longitudinal trajectories of fixed-effect coefficients. Simulation studies are presented to demonstrate the efficacy of the method and its computational efficiency in estimating smooth time-varying effects in high dimensional settings. A real data example on the Health and Retirement Study (HRS) analysis is used to demonstrate the practical usage of our method to infer age-varying impact of chronic disease on older people's physical functions.

Intron retention (IR) is the least well-understood alternative splicing type in animals, and its prevalence and function in physiological and pathological processes have long been underestimated. Cellular senescence contributes to individual aging and age-related diseases and can also serve as an important cancer prevention mechanism. Dynamic IR events have been observed in senescence models and aged tissues; however, whether and how IR impacts senescence remain unclear. Through analyzing polyA⁺ RNA-seq data from human replicative senescence models, we found IR was prevalent and dynamically regulated during senescence and IR changes negatively correlated with expression alteration of corresponding genes. We discovered that knocking down (KD) splicing factor U2AF1, which showed higher binding density to retained introns and decreased expression during senescence, led to senescence-associated phenotypes and global IR changes. Intriguingly, U2AF1-KD-induced IR changes also negatively correlated with gene expression. Furthermore, we demonstrated that U2AF1-mediated IR of specific gene (CPNE1 as an example) contributed to cellular senescence. Decreased expression of U2AF1, higher IR of CPNE1, and reduced expression of CPNE1 were also discovered in dermal fibroblasts with age. We discovered prevalent IR could fine-tune gene expression and contribute to senescence-associated phenotypes, largely extending the biological significance of IR.

OBJECTIVE:Electronic health records (EHRs) have become a common data source for clinical risk prediction, offering large sample sizes and frequently sampled metrics. There may be notable differences between hospital-based EHR and traditional cohort samples: EHR data often are not population-representative random samples, even for particular diseases, as they tend to be sicker with higher healthcare utilization, while cohort studies often sample healthier subjects who typically are more likely to participate. We investigate heterogeneities between EHR- and cohort-based inferences including incidence rates, risk factor identifications/quantifications, and absolute risks. MATERIALS AND METHODS/METHODS:This is a retrospective cohort study of older patients with type 2 diabetes using EHR from New York University Langone Health ambulatory care (NYULH-EHR, years 2009-2017) and from the Health and Retirement Survey (HRS, 1995-2014) to study subsequent cardiovascular disease (CVD) risks. We used the same eligibility criteria, outcome definitions, and demographic covariates/biomarkers in both datasets. We compared subsequent CVD incidence rates, hazard ratios (HRs) of risk factors, and discrimination/calibration performances of CVD risk scores. RESULTS:The estimated subsequent total CVD incidence rate was 37.5 and 90.6 per 1000 person-years since T2DM onset in HRS and NYULH-EHR respectively. HR estimates were comparable between the datasets for most demographic covariates/biomarkers. Common CVD risk scores underestimated observed total CVD risks in NYULH-EHR. DISCUSSION AND CONCLUSION/CONCLUSIONS:EHR-estimated HRs of demographic and major clinical risk factors for CVD were mostly consistent with the estimates from a national cohort, despite high incidences and absolute risks of total CVD outcome in the EHR samples.

BACKGROUND:Immune checkpoint inhibition (ICI) improves survival outcomes for patients with several types of cancer including metastatic melanoma (MM), but serious immune-related adverse events requiring intervention with immunosuppressive medications occur in a subset of patients. Skin toxicity (ST) has been reported to be associated with better response to ICI. However, understudied factors, such as ST severity and potential survivor bias, may influence the strength of these observed associations. METHODS:To examine the potential confounding impact of such variables, we analyzed advanced cancer patients enrolled prospectively in a clinicopathological database with protocol-driven follow up and treated with ICI. We tested the associations between developing ST, stratified as no (n = 617), mild (n = 191), and severe (n = 63), and progression-free survival (PFS) and overall survival (OS) in univariable and multivariable analyses. We defined severe ST as a skin event that required treatment with systemic corticosteroids. To account for the possibility of longer survival associating with adverse events instead of the reverse, we treated ST as a time-dependent covariate in an adjusted model. RESULTS:Both mild and severe ST were significantly associated with improved PFS and OS (all P < 0.001). However, when adjusting for the time from treatment initiation to time of skin event, severe ST was not associated with PFS benefit both in univariable and multivariable analyses (P = 0.729 and P = 0.711, respectively). Receiving systemic steroids for ST did not lead to significant differences in PFS or OS compared to patients who did not receive systemic steroids. CONCLUSIONS:Our data reveal the influence of time to event and its severity as covariates in analyzing the relationship between ST and ICI outcomes. These differences in outcomes cannot be solely explained by the use of immunosuppressive medications, and thus highlight the importance of host- and disease-intrinsic factors in determining ICI response and toxicity. TRIAL REGISTRATION:The patient data used in this manuscript come from patients who were prospectively enrolled in two institutional review board-approved databases at NYU Langone Health (institutional review board #10362 and #S16-00122).

BACKGROUND:Recent research suggests that baseline body mass index (BMI) is associated with response to immunotherapy. In this study, we test the hypothesis that worsening nutritional status prior to the start of immunotherapy, rather than baseline BMI, negatively impacts immunotherapy response. METHODS:We studied 629 patients with advanced cancer who received immune checkpoint blockade at New York University. Patients had melanoma (n=268), lung cancer (n=128) or other primary malignancies (n=233). We tested the association between BMI changes prior to the start of treatment, baseline prognostic nutritional index (PNI), baseline BMI category and multiple clinical end points including best overall response (BOR), objective response rate (ORR), disease control rate (DCR), progression-free survival (PFS) and overall survival (OS). RESULTS:0.001 and p<0.001). Baseline BMI category was not significantly associated with any treatment outcomes. CONCLUSION/CONCLUSIONS:Standard of care measures of worsening nutritional status more accurately associate with immunotherapy outcomes than static measurements of BMI. Future studies should focus on determining whether optimizing pretreatment nutritional status, a modifiable variable, leads to improvement in immunotherapy response.

INTRODUCTION/BACKGROUND:A grading system for pulmonary adenocarcinoma has not been established. The IASLC pathology panel evaluated a set of histological criteria associated with prognosis aimed at establishing a grading system for invasive pulmonary adenocarcinoma. DESIGN/METHODS:A multi-institutional study involving multiple cohorts of invasive pulmonary adenocarcinomas was conducted. A cohort of 284 stage I pulmonary adenocarcinomas was used as a training set to identify histological features associated with patient outcomes (recurrence-free survival [RFS] and overall survival [OS]). ROC curve analysis was used to select the best model, which was validated (n=212) and tested (n=300, including Stage I-III) in independent cohorts. Reproducibility of the model was assessed using kappa statistics. RESULTS:The best model (AUC= 0.749 for RFS and 0.787 for OS) was composed of a combination of predominant plus high-grade histological pattern with a cut off of 20% for the latter. The model consists of: Grade 1: lepidic predominant tumor; Grade 2: acinar or papillary predominant tumor, both with no or less than 20% of high-grade patterns; and Grade 3: any tumor with 20% or more of high-grade patterns (solid, micropapillary and or complex gland). Similar results were seen in the validation (AUC = 0.732 for RFS and 0.787 for OS) and test cohorts (AUC of 0.690 for RFS and 0.743 for OS), confirming the predictive value of the model. Inter-observer reproducibility showed good agreement (k=0.617). CONCLUSION/CONCLUSIONS:A grading system based on the predominant and high-grade patterns is practical and prognostic for invasive pulmonary adenocarcinoma. ACKNOWLEDGEMENT/UNASSIGNED:This project was supported by grants from the International Association for the study of Lung Cancer (IASLC), and NIH/NCI Early Detection Research Network 1U01CA214195-01 to HP used for biospecimen collection and salary support (DFL). We would like to thank the staff of the Center of Biospecimen Research and Development (CBRD) from the New York University for their help in digital pathology and histology for this project. CBRD is partially supported by the Cancer Center Support Grant P30CA016087 at the Laura and Isaac Perlmutter Cancer Center.

Melanoma disseminates to the skeletal system where it is then difficult to treat. Yet, there remains limited research investigating metastatic bone disease (MBD) in melanoma. Here, we evaluate whether there are distinct clinicopathologic variables at the time of primary melanoma diagnosis that predispose metastases to engraft bone, and we test the hypothesis that patients with MBD have different responses to treatment. Cutaneous melanoma patients enrolled in a prospective database were studied. Individuals with metastatic melanoma and bone metastases (M-Bone) were compared to those with metastatic disease but no M-Bone. Of the 463 (42.7%) patients, 198 with unresectable metastatic melanoma had M-Bone and 98 developed bone metastasis (bone mets) as first site. Progression-free survival and overall survival were significantly worse in patients with M-Bone compared to those without M-Bone (P < 0.001) independent of treatment modalities, and in patients whose melanoma spread to bone first, compared to those who developed first mets elsewhere (P < 0.001). Interestingly, patients with bone mets presented with primary tumors that had more tumor infiltrating lymphocytes (P < 0.001) and less often a nodular histologic subtype compared to patients without M-Bone (P < 0.001). Our data suggest that melanoma bone metastasis is a distinct clinical and biological entity that cannot be explained by generalized metastatic phenotype in all patients. The observed dichotomy between more favorable primary histopathologic characteristics and a grave overall prognosis requires more studies to elucidate the molecular processes by which melanomas infiltrate bone and to build a mechanistic understanding of how melanoma bone metastases yield such detrimental outcomes.

OBJECTIVES/OBJECTIVE:Recent investigations have shown strong correlations between cytology and surgical non-small cell lung carcinoma (NSCLC) specimens in programmed death-ligand 1 (PD-L1) immunohistochemical (IHC) evaluations. Our study aims to evaluate the reproducibility of PD-L1 IHC scoring in NSCLC cytology cell blocks (CBs) and to assess the impact of CB cellularity, method of sample collection, and observer subspecialty on scoring agreement. METHODS:PD-L1 IHC was performed on 54 NSCLC cytology CBs and was scored independently by seven cytopathologists (three of seven with expertise in pulmonary pathology). Three-tier scoring of negative (<1%), low positive (1%-49%), and high positive (≥50%) and interrater agreement were assessed. RESULTS:Total and majority agreement among cytopathologists was achieved in 48% and 98% of cases, respectively, with κ = 0.608 (substantial agreement; 95% confidence interval, 0.50-0.72). Cytopathologists with pulmonary pathology expertise agreed in 67% of cases (κ = 0.633, substantial agreement), whereas the remaining cytopathologists agreed in 56% of cases (κ = 0.62, substantial agreement). CB cellularity (P = .36) and sample collection type (P = .59) had no statistically significant difference between raters. CONCLUSIONS:There is substantial agreement in PD-L1 IHC scoring in cytology CB specimens among cytopathologists. Additional expertise in pulmonary pathology, sample collection type, and CB cellularity have no statistically significant impact on interobserver agreement.

BACKGROUND:The American Joint Committee on Cancer (AJCC) maintains that the eighth edition of its Staging Manual (AJCC8) has improved accuracy compared to the seventh (AJCC7). However, there are concerns that implementation may disrupt analysis of active clinical trials for stage III patients. We used an independent cohort of melanoma patients to test the extent to which AJCC8 has improved prognostic accuracy compared to AJCC7. METHODS:We analyzed a cohort of 1,315 prospectively enrolled patients. We assessed primary tumor and nodal classification of stage I-III patients using AJCC7 and AJCC8 to assign disease stages at diagnosis. We compared recurrence-free (RFS) and overall survival (OS) using Kaplan-Meier curves and log-rank tests. We then compared concordance indices of discriminatory prognostic ability and area under the curve (AUC) of 5-year survival to predict RFS/OS. All statistical tests were two-sided. RESULTS:Stage IIC continued to have worse outcomes than those for stage IIIA patients, with 5-year RFS of 26.5% (95%CI=12.8-55.1%) vs. 56.0% (95%CI=37.0-84.7%) by AJCC8 (P = 0.002). For stage I, removing mitotic index as T classification factor decreased its prognostic value, although not statistically significantly (RFS C-index=0.63 [95%CI=0.56-0.69] to 0.56 [95%CI=0.49-0.63], P = 0.07; OS C-index=0.48 [95%CI=0.38-0.58] to 0.48 [95%CI=0.41-0.56], P = 0.90). For stage II, prognostication remained constant (RFS C-index=0.65 [95%CI=0.57-0.72]; OS C-index=0.61 [95%CI=0.50-0.72]), and. For stage III, AJCC8 yielded statistically significantly enhanced prognostication for RFS (C-index=0.65 [95%CI=0.60-0.70] to 0.70 [95%CI=0.66-0.75], P = 0.01). CONCLUSIONS:Compared with AJCC7, we demonstrate that AJCC8 enables more accurate prognosis for patients with stage III melanoma. Restaging a large cohort of patients can enhance the analysis of active clinical trials.

PURPOSE/OBJECTIVE:For more than half of Crohn's disease patients, strictures will cause bowel obstructions that require surgery within 10 years of their initial diagnosis. This study utilizes computed tomography imaging and clinical data obtained at the initial emergency room visit to create a prediction model for progression to surgery in Crohn's disease patients with acute small bowel obstructions. METHODS:A retrospective chart review was performed for patients who presented to the emergency room with an ICD-10 diagnosis for Crohn's disease and visit diagnosis of small bowel obstruction. Two expert abdominal radiologists evaluated the CT scans for bowel wall thickness, maximal and minimal luminal diameters, length of diseased segment, passage of oral contrast, evidence of penetrating disease, bowel wall hyperenhancement or stratification, presence of a comb sign, fat hypertrophy, and small bowel feces sign. The primary outcome was progression to surgery within 6 months of presentation. The secondary outcome was time to readmission. RESULTS:Forty patients met the inclusion criteria, with 78% receiving medical treatment alone and 22% undergoing surgery within 6 months of presentation to the emergency room. Multivariable analysis produced a model with an AUC of 92% (95% CI 0.82-1.00), 78% sensitivity, and 97% specificity, using gender, body mass index, and the radiographic features of segment length, penetrating disease, and bowel wall hyperenhancement. CONCLUSIONS:The model demonstrates that routine clinical and radiographic data from an emergency room visit can predict progression to surgery, and has the potential to risk stratify patients, guide management in the acute setting, and predict readmission.