Related Experiment Video
Updated: Jun 21, 2026
09:58
DNA-barcode-based Multiplex Immunofluorescence Imaging to Analyze FFPE Specimens from Genetically Reprogrammed Murine Melanoma
Published on: June 6, 2025
Development of a microarray platform for FFPET profiling: application to the classification of human tumors
Sven Duenwald1, Mingjie Zhou, Yanqun Wang
1Translational Sciences, Department of Molecular Profiling, Merck Research Laboratories, Seattle, WA 98109, USA. sven_duenwald@merck.com
Journal of Translational Medicine
|July 30, 2009
Summary
This study optimized a microarray platform for profiling formalin-fixed paraffin-embedded tissues (FFPET). The platform accurately classifies breast cancer samples, enabling prognostic assay development from archival FFPET specimens.
Area of Science:
- Biotechnology
- Genomics
- Cancer Research
Background:
- Messenger RNA (mRNA) profiling is crucial for developing prognostic assays to predict treatment response and patient outcomes.
- Formalin-fixed paraffin-embedded tissues (FFPET) archives offer a valuable resource for retrospective studies.
- There is a need for methods to profile FFPET samples for hypothesis generation and clinical classifier development.
Purpose of the Study:
- To develop and optimize a microarray-based profiling platform for FFPET samples.
- To assess the accuracy of FFPET profiling in predicting clinical outcomes.
- To compare the predictive power of classifiers derived from FFPET data versus fresh-frozen (FF) data.
Main Methods:
- Developed a two-color microarray platform with optimized target amplification, experimental design, quality control, and content.
- Profiled 50 fresh-frozen (FF) breast cancer samples and 50 matched FFPET samples.
- Assigned class labels based on established methods and compared classifier performance between FF and FFPET data.
Main Results:
- Classifiers developed from FF samples accurately assigned FFPET samples to "good" or "poor" outcome classes with low error rates (12-16%) and high Odds Ratios.
- A classifier derived from FFPET data achieved 14% error rate in predicting FFPET sample class labels (p < 3.7 x 10(-7)).
- The FFPET-derived classifier demonstrated 96% accuracy when applied to matched FF samples, highlighting the platform's robustness and the importance of sample quality control.
Conclusions:
- Optimized microarray platform enables accurate expression analysis and classification of FFPET samples.
- FFPET-derived classifiers possess comparable predictive power to those derived from FF samples.
- This technology facilitates hypothesis generation from archival FFPET, potentially leading to prognostic/predictive classifiers for clinical trial enrollment and patient treatment guidance.
