PhenoImager® Fusion

We are developing CNN model to segment neurons and astrocytes, etc

Review Date: 26 Sept 2022 | Akoya Biosciences

Fluorescence based multispectral imaging can be performed with Vectra. We have used Vectra for more than 3 years, and analyzed any tissue samples stained for multiplex IHC. This imaging platform can be a good solution for detecting up to 7 markers. It can also scan slides stained with H&E or chromogenic IHC. We have tested detection limit for fluorescence signal. If the signal is weak (e.g. immunofluorescence without signal amplification like immunohistochemistry), it may or may not detect the signal because of the limited exposure time for scanning. Immunohistochemically amplified signals usually do not have problem being detected. However, when the signal is super high on one marker, it may disrupt unmixing procedure on inForm software, so the staining conditions need to be optimized to acquire signals from multiple markers. Scanning slides is easy to perform, and it does not need to be in the dark room. Scanning speed is quite fast to acquire entire tissue image (x4 and x10) and multi-spectral images (x20, x40). Analysis for scanned images is easily performed on inForm software. The inForm software can do: (1) unmix the each fluorescence marker signal from multispectral image, (2) perform tissue segmentation (if tissue marker like cytokeratin is stained), (3) cell segmentation to locate each cell location on the image and fractionate subcellular compartment (nucleus, cytoplasm and membrane), (4) score marker levels and (5) summarize and convert dataset into tissue and cell related parameters and different image format. The parameters indicating the levels of each marker can be used for FACS like analysis. Converted images into multilayered TIFF file can be opened with other imaging software. This AI supported imaging analysis procedure with inForm is getting easier in newer version. However, if tissue architecture is unique (e.g. cells are elongated, distance of nucleus to membrane is big, the tissue has various size of nuclei), it may take much time to train the AI or it may not clearly show each cell membrane. The maintenance is minimal like a conventional microscope. Since a light bulb for detecting fluorescence needs to be replaced after a certain number of hours, we have upgraded the light source to LED. The worst thing on Vectra is actually when the light bulb is reaching a certain time. It will keep beeping, so no one wants to be in the room! Customer care for technical service is very quick and user friendly. As a scientist who has been working with various microscopes, it is not good enough to acquire molecular localization on a particular organelles but this Vectra system is good for identifying cell types in tissue sample and it is also good for quantifying the levels of markers in each cell in the tissue sample.

Review Date: 3 Dec 2020 | Akoya Biosciences