Rollins School of Public Health | Faculty Profile
Emory Rollins School of Public Health

Zhengjia "Nelson"  Chen

Research Associate Professor

Faculty, Biostatistics and Bioinformatics

  • Designs and some critical issues of Phase I, II, and III clinical trials
  • Dose-Toxicity Bayesian models in cancer Phase I clinical trials
  • A novel toxicity score system to measure the multiple toxicities per patient quantitatively and comprehensively in Phase I clinical trials
  • Methods treating disease response as a continuous variable instead of using RECIST in Phase II clinical trials
View Resume

Contact Information

1518 Clifton Road GCR#342

Atlanta , GA 30322

Phone: (404) 712-8278

Fax: 404-778-5016

Email: zchen38@emory.edu

Update Profile Here

Areas of Interest

  • Bioinformatics
  • Cancer Prevention
  • HIV/AIDS Prevention
  • Biomarkers
  • Biostatistics
  • Clinical Trials
  • Bayesian Analysis

Education

  • PhD 2008, University of Southern California, Los Angeles, CA
  • MS 2001, University of Southern California, Los Angeles, CA
  • MS 1998, Peking University, Beijing, China
  • BS 1995, Peking University, Beijing, China

Courses Taught

  • BIOS 520: Clinical Trials

Affiliations & Activities

SOFTWARE DEVELOPED

1. ID-NETS (Isotonic Design using Normalized Equivalence Toxicity Score).

ID-NETS is a novel semi-parametric Phase I design which is developed by integration of original isotonic design and my NETS system. ID-NETS can substantially improve the accuracy of MTD determination and trial efficiency by fully utilizing all toxicities of each patient and treating toxicity response as a quasi-continuous variable instead of a binary indicator of DLT. ID-NETS is also very robust because it only assumes a monotonically increasing dose toxicity relationship instead of any parametric model. ID-NETS is particularly appropriate for clinical trials which have no reliable parametric model for dose toxicity relationship and the target toxicity level is outside the range of 18-33% of DLT. ID-NETS©TM is an interactive and user-friendly statistical software to implement the design ID-NETS in Phase I clinical trials. The standalone version of the software can be downloaded freely from the website: https://winshipbbisr.emory.edu/IDNETS.html

2. EWOC-NETS (Escalation with Overdose Control using Normalized Equivalence Toxicity Score).

EWOC-NETS is a novel Bayesian adaptive design which is a combination of EWOC and my NETS system. EWOC-NETS not only retains all the advantages of EWOC, but also treats toxicity response as a quasi-continuous variable instead of a binary indicator of DLT, fully utilizes all toxicity information, and improves the accuracy of MTD estimation and the efficiency of Phase I trials. EWOC-NETS©TM is an interactive and user-friendly statistical software to implement the design EWOC-NETS in Phase I clinical trials. The standalone version of the software can be downloaded freely from the website: https://winshipbbisr.emory.edu/EWOCNETS.html

3. Standard 3+3 Design simulator

Standard 3+3 with or without dose de-escalation designs are still most widely used in cancer Phase I clinical trials for their simplicity and robustness. In clinical practice and grant applications, it is necessary to provide the operating characteristics of Phase I clinical trials at the planning stage. Therefore I have developed an interactive and user-friendly software called Standard-3+3-design©TM to estimate the operating characteristics of Phase I clinical trials using Standard 3+3 designs. The standalone version of the software is available for free download and the web-based version can be directly used online at the website: https://winshipbbisr.emory.edu/3_3.html

4.  Matlab Codes (http://web1.sph.emory.edu/users/zchen38/MATLABCode.pdf ) for the Publication

        (Chen Z, Chen X. Exact Calculation of Minimum Sample Size for Estimating a Possion Parameter,  Communication in Statistics - Theory and Methods. 2014, In press)

5. Development of an interactive statistical software for calculation of the operating characteristics of Phase I clinical trials using standard 3+3 design. (Using R shiny).

6. Development of an interactive statistical software for Phase I clinical trials using EWOC-NETS design. (Using R shiny).

Publications