Validation and reproducibility of a microarray-based gene expression test for tumor identification in formalin-fixed,...

Area of Science:

• Oncology
• Molecular Diagnostics
• Genomics

Background:

• Diagnosing the primary site of metastatic or undifferentiated tumors is clinically challenging.
• Accurate tumor origin identification is crucial for effective treatment selection and patient management.
• Existing diagnostic methods may be limited for poorly characterized or metastatic cancers.

Purpose of the Study:

• To validate a gene expression-based diagnostic test for determining tumor tissue of origin.
• To assess the test's performance using formalin-fixed, paraffin-embedded (FFPE) tumor specimens.
• To evaluate the test's accuracy, reproducibility, and ability to rule out potential origins.

Main Methods:

• Generated microarray data from 462 FFPE tumor specimens with known reference diagnoses.
• Analyzed gene expression data using a 2000-gene classification model.
• Quantified similarity between specimen RNA expression patterns and 15 tissue types.

Main Results:

• Achieved 89% overall agreement with reference diagnoses across 462 FFPE specimens.
• Demonstrated high specificity, ruling out an average of 12 tissues with >99% probability per specimen.
• Showcased multisite reproducibility with 89.3% concordance between laboratories.

Conclusions:

• The gene expression-based Tissue of Origin Test is a validated and accurate tool for FFPE specimens.
• The test aids in diagnosing tumors with challenging primary sites, improving diagnostic confidence.
• Microarray-based gene expression analysis is feasible and beneficial for routine FFPE tumor diagnostics.