AI-Assisted High-Throughput Tissue Microarray Workflow
Author Information
Author(s): Konrad Kurowski, Sylvia Timme, Melanie Christine Föll, Clara Backhaus, Philipp Anton Holzner, Bertram Bengsch, Oliver Schilling, Martin Werner, Peter Bronsert
Primary Institution: University of Freiburg
Hypothesis
The study aimed to create a high-throughput workflow using modern technologies to facilitate IHC biomarker studies on large patient groups.
Conclusion
The AI-assisted high-throughput TMA workflow significantly reduces labor and material resource consumption while maintaining research quality.
Supporting Evidence
- The workflow reduced labor time to 7.7% compared to traditional whole-slide studies.
- Only 18 unstained slides were required, taking 15.25 hours to process versus 347.49 hours for whole-slide studies.
- Antibody use was significantly lower, with 5.4 µg required compared to 271.3 µg in whole-slide studies.
- Core loss was minimized through a refined melting process, reducing loss rates to 1.25%.
Takeaway
This study created a faster way to analyze cancer samples using AI, which helps researchers save time and resources while still getting good results.
Methodology
The study involved creating semi-automated tissue microarrays from FFPE samples, followed by AI-assisted analysis of IHC staining and statistical correlation with clinicopathological data.
Potential Biases
Potential variability in results due to differences in tissue processing and analysis methods.
Limitations
The workflow's effectiveness may vary based on specific institutional factors such as tissue fixation methods and antigen retrieval techniques.
Participant Demographics
Cohort A included 49 patients with intrahepatic cholangiocellular carcinoma, and Cohort B included 142 patients with non-small cell lung cancer.
Statistical Information
P-Value
p<0.001
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website