SPHG 101 - Biostatistics
Introduction to Biostatistics in Public Health
Current Role of Biostatistics
Discusses the importance of biostatistics in public health.
Originates from the speaker's work at a comprehensive cancer center, particularly at Lineberger focusing on cancer.
Speaker's Background
Personal Journey and Education
Grew up in Maryland.
Internship at NIH during high school ignited interest in cancer research.
Attended Duke University, majoring in biology and math, became interested in bioinformatics and genomics during genomics’ rise in the early 2000s.
Completed training grant in biostatistics at UNC, focusing on cancers.
Further research conducted at Harvard School of Public Health and Dana Farber Cancer Institute on using genomic information for patient prognosis.
Current Position at UNC
Leads cancer clinical trial designs to evaluate new treatments.
Director of the biostatistics core at Lineberger cancer center, aiding faculty in study design and statistical consultations.
Teaches Bio 735: Basic Coding in R, software implementation, and machine learning in biostatistics.
Importance of Biostatistics at UNC
Why Pursue Biostatistics
Definition: Applications of statistics in human health, public health, and biology.
Distinction from traditional statisticians, focusing on health data and analyses.
Career prospects for biostatistics graduates:
Significant placements in pharmaceutical, technology firms, academic positions.
Strong track record of graduates advancing to medical and graduate schools or directly into the workforce.
Research Focus: Cancer Genomics and Precision Medicine
Current Research Areas
Emphasis on precision medicine and cancer subtyping.
Developing biomarkers to subcategorize patients for tailored treatments.
Investigating machine learning methods to handle missing entries in electronic health record datasets.
Funding for the development of large language models to recommend clinical trials for pancreatic cancer patients.
Genomics and Cancer Research
Genomic Insights
Genomics defined: Measuring cellular activities via assays.
DNA sequencing for mutations, RNA sequencing for gene expression, investigation of proteins that regulate gene expression.
Aberrations at DNA/RNA levels lead to cancer through cellular state changes.
Biochemical Examinations in Cancer
During diagnosis, assays help determine tumor characteristics and potential treatments.
Uses genomic data from multiple patients to correlate markers indicating treatment efficacy.
Long-standing Role of UNC Cancer Research
Historical contributions to the subtyping of breast and pancreatic cancers.
Clinical tests developed based on these studies integrated into clinical practice.
Ongoing studies to address cancer disparity based on subtypes and demographics.
Pancreatic Cancer Research Initiatives
Challenges in Pancreatic Cancer
Poor survival rates (5-year survival: 10%; median survival under 11 months) due to late-stage diagnosis.
Silent nature of the disease complicates early detection.
Research Developments
Identification of pancreatic cancer subtypes using gene expression data through non-negative matrix factorization and consensus clustering identified two main subtypes:
"Basal-like" (orange) and "Classical" (blue).
Differences in survival rates observed depending on subtype; basal-like patients show a higher rate of aggressive disease with lower survival.
Statistical Models for Treatment Prediction
Developed the PUREST model for predicting pancreatic cancer subtypes using penalized logistic regression.
Outcome validation: basal-like subtype patients exhibit poor tumor response to standard therapies compared to classical subtype.
Clinical Trial Design and Innovative Approaches
Current Clinical Trial Work
Engagement in designing adaptive trials in response to patient tumor changes during treatment.
Collaboration with physician experts to develop frameworks for real-time monitoring and treatment adaptation.
ARPA-H Program Contribution
Funding of $28-30 million for a clinical trial focused on responding to evolving tumor environments in metastatic cancer.
Trial Design Innovations
Patient samples used to determine treatment eligibility and randomization based on biomarkers; allows real-time treatment adjustments.
Student Involvement and Outcomes
Student Research Opportunities
Examples include a student analyzing mutations in colorectal cancer impacting treatment efficacy and leading to publications and potential industry collaborations.
Statistical Method Development
Students worked on developing new statistical models and software for cancer research outcomes.
Epidemiological and Preventive Measures in Cancer Research
Public Health Perspective
Discussion on preventive measures receiving less funding; need for robust research on early detection methods for various cancers.
National Cancer Institute initiative for early detection to combat late-stage diagnoses and enhance treatment efficiency.
Innovative Screenings
Blood and imaging-based early detection methods being explored.
Conclusion
Overall Contributions of Biostatistics
Empowering research efforts to tackle complex health questions and improve patient outcomes through collaboration between statisticians and subject matter experts.
Critical role of biostatisticians in advancing public health, clinical applications, and the utilization of technology in cancer research.
Q&A Session Insights
Interaction with Patients
Limited direct contact with patients; key insights gained through advisory roles enhancing research focus.
Calls to leverage computational tools to improve patient access to clinical trial information.
Patient-Centric Research
Emphasis on the necessity for public education on cancer prevention as a vital aspect of healthcare.