Susan Halabi

Overview:

Design and analysis of clinical trials, statistical analysis of biomarker and high dimensional data, development and validation of prognostic and predictive models.

Positions:

Professor of Biostatistics and Bioinformatics

Biostatistics & Bioinformatics
School of Medicine

Chief, Division of Biostatistics

Biostatistics & Bioinformatics
School of Medicine

Member of the Duke Cancer Institute

Duke Cancer Institute
School of Medicine

Education:

Ph.D. 1994

University of Texas Health Sciences Center, Houston

Grants:

PCRP Clinical Consortium: Duke University Clinical Research Site

Administered By
Medicine, Medical Oncology
Awarded By
Department of Defense
Role
Co Investigator
Start Date
End Date

Developing and Validating Prognostic Models of Clinical Outcomes In Men With Castration Resistant Prostate Cancer

Administered By
Biostatistics & Bioinformatics
Awarded By
National Institutes of Health
Role
Principal Investigator
Start Date
End Date

Surrogate Endpoints of Overall Survival in Men with Metastatic Hormone Sensitive Prostate Cancer

Administered By
Biostatistics & Bioinformatics
Awarded By
Prostate Cancer Foundation
Role
Principal Investigator
Start Date
End Date

Precision Medicine in Platinum-treated Lethal Bladder Cancer

Administered By
Biostatistics & Bioinformatics
Awarded By
Memorial Sloan Kettering Cancer Center
Role
Principal Investigator
Start Date
End Date

Serum Androgens and Survival in CRPC

Administered By
Duke Cancer Institute
Awarded By
University of California - San Francisco
Role
Principal Investigator
Start Date
End Date

Publications:

Comparison of regression imputation methods of baseline covariates that predict survival outcomes

<jats:title>Abstract</jats:title> <jats:sec id="S2059866120005336_as1"> <jats:title>Introduction:</jats:title> <jats:p>Missing data are inevitable in medical research and appropriate handling of missing data is critical for statistical estimation and making inferences. Imputation is often employed in order to maximize the amount of data available for statistical analysis and is preferred over the typically biased output of complete case analysis. This article examines several types of regression imputation of missing covariates in the prediction of time-to-event outcomes subject to right censoring.</jats:p> </jats:sec> <jats:sec id="S2059866120005336_as2"> <jats:title>Methods:</jats:title> <jats:p>We evaluated the performance of five regression methods in the imputation of missing covariates for the proportional hazards model via summary statistics, including proportional bias and proportional mean squared error. The primary objective was to determine which among the parametric generalized linear models (GLMs) and least absolute shrinkage and selection operator (LASSO), and nonparametric multivariate adaptive regression splines (MARS), support vector machine (SVM), and random forest (RF), provides the “best” imputation model for baseline missing covariates in predicting a survival outcome.</jats:p> </jats:sec> <jats:sec id="S2059866120005336_as3"> <jats:title>Results:</jats:title> <jats:p>LASSO on an average observed the smallest bias, mean square error, mean square prediction error, and median absolute deviation (MAD) of the final analysis model’s parameters among all five methods considered. SVM performed the second best while GLM and MARS exhibited the lowest relative performances.</jats:p> </jats:sec> <jats:sec id="S2059866120005336_as4"> <jats:title>Conclusion:</jats:title> <jats:p>LASSO and SVM outperform GLM, MARS, and RF in the context of regression imputation for prediction of a time-to-event outcome.</jats:p> </jats:sec>
Authors
Solomon, N; Lokhnygina, Y; Halabi, S
MLA Citation
Solomon, Nicole, et al. “Comparison of regression imputation methods of baseline covariates that predict survival outcomes.” Journal of Clinical and Translational Science, vol. 5, no. 1, Cambridge University Press (CUP), 2021. Crossref, doi:10.1017/cts.2020.533.
URI
https://scholars.duke.edu/individual/pub1459269
Source
crossref
Published In
Journal of Clinical and Translational Science
Volume
5
Published Date
DOI
10.1017/cts.2020.533

Prostate Cancer Foundation Hormone-Sensitive Prostate Cancer Biomarker Working Group Meeting Summary.

Androgen deprivation therapy remains the backbone therapy for the treatment of metastatic hormone-sensitive prostate cancer (mHSPC). In recent years, several treatments, including docetaxel, abiraterone + prednisone, enzalutamide, and apalutamide, have each been shown to demonstrate survival benefit when used upfront along with androgen deprivation therapy. However, treatment selection for an individual patient remains a challenge. There is no high level clinical evidence for treatment selection among these choices based on biological drivers of clinical disease. In August 2020, the Prostate Cancer Foundation convened a working group to meet and discuss biomarkers for hormone-sensitive prostate cancer, the proceedings of which are summarized here. This meeting covered the state of clinical and biological evidence for systemic therapies in the mHSPC space, with emphasis on charting a course for the generation, interrogation, and clinical implementation of biomarkers for treatment selection.
Authors
Hofmann, MR; Hussain, M; Dehm, SM; Beltran, H; Wyatt, AW; Halabi, S; Sweeney, C; Scher, HI; Ryan, CJ; Feng, FY; Attard, G; Klein, E; Miyahira, AK; Soule, HR; Sharifi, N
MLA Citation
Hofmann, Martin R., et al. “Prostate Cancer Foundation Hormone-Sensitive Prostate Cancer Biomarker Working Group Meeting Summary.Urology, Dec. 2020. Pubmed, doi:10.1016/j.urology.2020.12.021.
URI
https://scholars.duke.edu/individual/pub1469924
PMID
33373705
Source
pubmed
Published In
Urology
Published Date
DOI
10.1016/j.urology.2020.12.021

Sunitinib in Patients with Metastatic Colorectal Cancer (mCRC) with FLT-3 Amplification: Results from the Targeted Agent and Profiling Utilization Registry (TAPUR) Study.

BACKGROUND: TAPUR is a pragmatic, phase II basket study evaluating the antitumor activity of commercially available targeted agents in patients with advanced cancers harboring genomic alterations known to be drug targets. Sunitinib is an oral multikinase inhibitor of FMS-like tyrosine kinase-3 (FLT-3), among other targets. Results from a cohort of patients with metastatic colorectal cancer (mCRC) with FLT-3 amplification treated with sunitinib are reported. OBJECTIVE: This study aimed to investigate whether patients with mCRC with FLT-3 amplification would be responsive to sunitinib, an oral multikinase inhibitor. METHODS: Eligible patients received a standard sunitinib dose of 50 mg orally for 4 weeks followed by 2 weeks off. Simon's two-stage design was used with the primary study endpoint of objective response (OR) or stable disease (SD) at 16 weeks based on Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. Secondary endpoints were progression-free survival, overall survival, and safety. RESULTS: Ten patients were enrolled from November 2016 to April 2018. All patients had mCRC with FLT-3 amplification. No ORs were observed. Although two patients had SD at 16 weeks, one died because of disease progression shortly thereafter and the cohort was closed. A single grade 3 adverse event of diarrhea was reported as possibly related to sunitinib. CONCLUSIONS: Monotherapy with sunitinib does not have clinical activity in patients with mCRC with FLT-3 amplification and should not be prescribed for off-label use. Other treatments should be considered for these patients, including treatments offered in clinical trials. CLINICAL TRIAL REGISTRATION: NCT02693535 (26 February 2016).
Authors
Al Baghdadi, T; Garrett-Mayer, E; Halabi, S; Mangat, PK; Rich, P; Ahn, ER; Chai, S; Rygiel, AL; Osayameh, O; Antonelli, KR; Islam, S; Bruinooge, SS; Schilsky, RL
MLA Citation
Al Baghdadi, Tareq, et al. “Sunitinib in Patients with Metastatic Colorectal Cancer (mCRC) with FLT-3 Amplification: Results from the Targeted Agent and Profiling Utilization Registry (TAPUR) Study.Target Oncol, vol. 15, no. 6, Dec. 2020, pp. 743–50. Pubmed, doi:10.1007/s11523-020-00752-8.
URI
https://scholars.duke.edu/individual/pub1463079
PMID
33068284
Source
pubmed
Published In
Target Oncol
Volume
15
Published Date
Start Page
743
End Page
750
DOI
10.1007/s11523-020-00752-8

Cetuximab in Patients with Breast Cancer, Non-Small Cell Lung Cancer, and Ovarian Cancer Without KRAS, NRAS, or BRAF Mutations: Results from the Targeted Agent and Profiling Utilization Registry (TAPUR) Study.

BACKGROUND: The Targeted Agent and Profiling Utilization Registry (TAPUR) Study, a phase II basket study, evaluates anti-tumor activity of commercially available targeted agents in patients with advanced cancers harboring genomic alterations known as drug targets. OBJECTIVE: With no known genomic targets predictive of sensitivity to cetuximab, cetuximab was evaluated in patients with breast cancer (BC), non-small cell lung cancer (NSCLC), and ovarian cancer (OC), without KRAS, NRAS, or BRAF mutations. PATIENTS AND METHODS: Eligible patients with advanced BC, NSCLC, and OC received a cetuximab loading dose, then weekly infusions (250 mg/m2 over 60 min). A Simon two-stage design, requiring ten patients in stage I, was employed per each disease-specific cohort. The primary endpoint was disease control (objective response or stable disease for at least 16 weeks). If two or more patients in stage I achieved disease control, the cohort would enroll 18 more patients in stage II. Power and alpha of the design are 85% and 10%, respectively. Secondary endpoints included progression-free survival, overall survival, and safety. RESULTS: Patients with BC (n = 10), NSCLC (n = 10), and OC (n = 29) were enrolled between June 2016 and September 2018. No objective responses or stable disease for at least 16 weeks were observed in the BC and NSCLC cohorts. No objective responses and four patients with stable disease for at least 16 weeks were observed in the OC cohort. Six of 49 patients reported grade 3 or higher adverse events or serious adverse events at least possibly related to cetuximab. CONCLUSIONS: Cetuximab does not have clinical activity in patients with advanced BC, NSCLC, and OC without KRAS, NRAS, or BRAF mutations. CLINICAL TRIAL REGISTRATION: NCT02693535 (26 February, 2016).
Authors
Fisher, JG; Tait, D; Garrett-Mayer, E; Halabi, S; Mangat, PK; Schink, JC; Alvarez, RH; Veljovich, D; Cannon, TL; Crilley, PA; Pollock, T; Calfa, CJ; Al Baghdadi, T; Thota, R; Fleming, N; Cotta, JA; Rygiel, AL; Warren, SL; Schilsky, RL
MLA Citation
URI
https://scholars.duke.edu/individual/pub1463080
PMID
33090333
Source
pubmed
Published In
Target Oncol
Volume
15
Published Date
Start Page
733
End Page
741
DOI
10.1007/s11523-020-00753-7

Pan-cancer prognostic models of clinical outcomes: statistical exercise or clinical tools?

Authors
MLA Citation
Halabi, S. “Pan-cancer prognostic models of clinical outcomes: statistical exercise or clinical tools?Ann Oncol, vol. 31, no. 11, Nov. 2020, pp. 1427–29. Pubmed, doi:10.1016/j.annonc.2020.08.2233.
URI
https://scholars.duke.edu/individual/pub1460004
PMID
32891792
Source
pubmed
Published In
Ann Oncol
Volume
31
Published Date
Start Page
1427
End Page
1429
DOI
10.1016/j.annonc.2020.08.2233

Research Areas:

Adenocarcinoma
Adenocarcinoma, Clear Cell
African Americans
Age Factors
Aged, 80 and over
Alkaline Phosphatase
Alleles
Arab countries
Area Under Curve
Biological Markers
Biomarkers, Pharmacological
Breast Neoplasms
Cancer Vaccines
Carcinoma
Carcinoma, Renal Cell
Case-Control Studies
Chemoprevention
Chemotherapy
Chi-Square Distribution
Clinical Trials, Phase II as Topic
Clinical trials
Cohort Studies
Computer Simulation
Confidence Intervals
Construction Materials
Contraceptives, Oral
DNA Damage
DNA Primers
DNA Repair
DNA, Neoplasm
Data Interpretation, Statistical
Decision Making
Decision Support Techniques
Diagnostic Imaging
Disease Progression
Disease-Free Survival
Drug Design
Dust
Efficiency, Organizational
Endpoint Determination
Equipment Design
Factor Analysis, Statistical
Family relationships
Gels
Gene Expression
Genes, Immunoglobulin
Genetic Predisposition to Disease
Genetics, Medical
Genotype
Germany
Graft vs Host Disease
HIV Infections
Hispanic Americans
Individualized Medicine
Kaplan-Meier Estimate
Ketoconazole
Lasso
Logistic Models
Lymphokines
Mining
Models, Biological
Models, Statistical
Models, Theoretical
Molecular Sequence Data
Multiprotein Complexes
Multivariate Analysis
Mutation
Neoplasms, Hormone-Dependent
Nomograms
Odds Ratio
Outcome Assessment (Health Care)
Ovarian Neoplasms
Personalized medicine
Population
Population Surveillance
Precision Medicine
Predictive Value of Tests
Pregnancy
Probability
Prognosis
Proportional Hazards Models
Prospective Studies
ROC Curve
Randomized Controlled Trials as Topic
Receptors, Progesterone
Registries
Reproducibility of Results
Research Design
Residence Characteristics
Retrospective Studies
Ribosomal Protein S6 Kinases
Risk
Risk Assessment
Risk Factors
Sample Size
Selective Estrogen Receptor Modulators
Sensitivity and Specificity
Statistics as Topic
Survival
Survival Analysis
Survival Rate
Tamoxifen
Translocation, Genetic
Treatment Failure
Treatment Outcome
Tumor Markers, Biological
United States
Urologic Neoplasms
Validation Studies as Topic
Vascular Endothelial Growth Factors