Automated Sorting of 3D Liver Spheroids Using Deep Learning
Author Information
Author(s): Sampaio da Silva Claudia, Boos Julia Alicia, Goldowsky Jonas, Blache Manon, Schmid Noa, Heinemann Tim, Netsch Christoph, Luongo Francesca, Boder-Pasche Stéphanie, Weder Gilles, Pueyo Moliner Alba, Samsom Roos-Anne, Marsee Ary, Schneeberger Kerstin, Mirsaidi Ali, Spee Bart, Valentin Thomas, Hierlemann Andreas, Revol Vincent
Primary Institution: CSEM SA Centre Suisse d’Electronique et de Microtechnique, Neuchâtel, Switzerland
Hypothesis
Can a fully automated platform enhance the sorting of 3D spheroids for bioprinting applications?
Conclusion
The developed sorting platform effectively sorts liver spheroids while preserving their viability and functionality.
Supporting Evidence
- The platform achieved a picking efficiency of 91% for low-quality spheroids.
- High-quality spheroids were harvested with an efficiency of 92% within 20 minutes.
- Deep learning models improved classification accuracy for spheroids with limited datasets.
Takeaway
Scientists created a robot that can quickly sort tiny liver balls to help make new organs for people who need transplants.
Methodology
The study developed a fully automated sorting platform that integrates image acquisition, analysis, and sorting of liver spheroids using deep learning.
Potential Biases
Potential bias in labeling spheroid quality due to reliance on expert judgment.
Limitations
The study's findings may be limited by the small dataset of tri-cellular spheroids used for training the deep learning model.
Participant Demographics
The study utilized human hepatocellular carcinoma (HepG2) cells and tri-cellular liver spheroids composed of multiple cell types.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website