Methodology to standardize heterogeneous statistical data presentations for combining time-to-event oncologic...

Area of Science:

• Oncology
• Biostatistics
• Medical Informatics

Background:

• Survival analysis in oncology requires specialized techniques due to data complexities like censoring and loss to follow-up.
• Standardization of oncologic outcomes (overall survival, disease-free survival, cancer recurrence) is challenging as hazard ratios are not always reported.
• Existing literature lacks a comprehensive framework for maximizing publication inclusion in meta-analyses of long-term oncologic outcomes.

Purpose of the Study:

• To propose a hierarchical methodology for extracting or estimating hazard ratios.
• To standardize time-to-event outcomes for enhanced meta-analysis inclusion.
• To facilitate more robust synthesis of oncologic outcomes from diverse statistical data presentations.

Main Methods:

• Developed a hierarchical approach to extract or estimate hazard ratios from published oncologic studies.
• Applied the method to compare overall survival, disease-free survival, and cancer recurrence for different surgical approaches.
• Utilized statistical analysis to quantify the increase in data inclusion potential.

Main Results:

• The proposed methodology significantly increases the potential for data inclusion in meta-analyses.
• In a proof-of-concept analysis, data inclusion rose from 108 to 240 hazard ratios (a 122% increase).
• Demonstrated the method's utility in comparing oncologic outcomes across robotic, open, and minimally invasive surgeries.

Conclusions:

• The proposed framework effectively standardizes time-to-event data for meta-analysis.
• This approach maximizes the inclusion of studies reporting long-term oncologic outcomes.
• Enhances the ability to conduct comprehensive meta-analyses in cancer research by overcoming data reporting heterogeneity.