Faire progresser la découverte de médicaments en combinant des modèles in vitro complexes avec une analyse d'images 3D/4D pilotée par l'IA

Yes, no matter the size, format or source of the image, the ZEISS arivis software can analyze it. The software works with most microscopy file formats from many imaging technologies.

Yes, in fact electron microscopy images (2D or 3D) can really benefit from training a deep learning model on ZEISS arivis Cloud. Then you can set up an AI-driven end-to-end analysis pipeline with ZEISS arivis Pro. If you want to scale up the segmentation and analysis of your EM images, you can run the batch analysis in ZEISS arivis Hub.

It is very simple. When you notice that the model does not perform as expected in a specific data set, you can simply upload the relevant image (or a cropped version of your image if it is very large) into ZEISS arivis Cloud, add some additional annotations where the model failed and continue training the model. After this process the model will also learn to consider the variety (level of uniformity) in your images.

Yes, you can set up image analysis pipelines yourself om ZEISS arivis Pro. The ZEISS arivis software platform contains many “out of the box” tools that do not require any programming knowledge. Even if you want to train AI models there is no need to code.

Yes, ZEISS arivis experts also offer customer support and professional consulting, so we can help you in creating your image analysis assays.

Yes. One of the benefits of arivis software is that you can create any type of image analysis pipelines even for live cells, in which you may want to track their movement and maybe quantify how a drug affects the migration of cells.

ZEISS arivis Hub has data plotting tools and well plate heatmaps. It can also connect with data analytics software solutions such as Genedata and Core Life that might be already available at your company, and which can give you more specific tools for data analytics.

These immortal ovarian cancer cells respond to cisplatin within a similar concentration range. Our in-house and published data report iC50 values between 1 and 3 µM for 2D adherent cells, while 3D culture iC50 values range from 6 to 11 µM depending on lab and method.

The RASTRUM Platform shown during the webinar is a standalone benchtop unit. While the created plates can readily be used with liquid handling and automated imaging platforms, this device itself is not accessible to robotic arms. For automation integration, please reach out to Inventia Life Sciences directly.

This is a very relevant point indeed, especially when imaging across multiple samples. The RASTRUM 3D cultures are created with multiple nanoliter droplets per well to ensure a consistent cell number output. Forming microliter gels with nanoliter droplets in a defined pattern allows furthermore a high control over within-well distribution. This not only makes automated imaging possible, but importantly ensures reproducible cell behaviors across wells.

The diversity of cell behaviors does indeed increase in 3D cultures compared to 2D, and even more so in matrix-based cultures compared to suspension spheroids. While this diversity is desired to represent more of the natural diversity observed in vivo, it does represent a challenge for assays that assume uniformity. To address this, the cell models generated with the RASTRUM platform express a heightened within-well diversity due to the advanced model complexity options, that is reproduced across wells through the precise nanoliters ejection method. This low within-plate variability despite the within-well diversity makes the RASTRUM cultures suitable for standardized assays. We recommend, however, to utilize multiple readout parameters (such as BF, IF, and metabolic activity) to capture the full diversity expressed by each well model.