Denis Baručić presents Characterization of drug effects on cell cultures from phase-contrast microscopy images

In the presented work, we classify images of drugs based on their
effect on cells. We work with phase-contrast microscopy images, which
are faster and easier to obtain than fluorescence images and support
live cell imaging. We employ a convolutional neural network trained
end-to-end directly on the input images. Our method can distinguish
between tested cytotoxic drugs with an accuracy of 98%, provided that
their mechanism of action differs, outperforming previous work. The
results are even better when substance-specific concentrations are
used. Visualization of extracted features reveals clusters which
correspond to known class labels, suggesting the possible use of our
methodology for drug discovery application in analyzing new, unseen
drugs.


This is a presentation of our paper:

Baručić D., Kaushik S., Kybic J., Stanková J., Džubák P. and Hajdúch M..
"Characterization of drug effects from phase-contrast images." Computers in
Medicine and Biology, 2022.
https://www.sciencedirect.com/science/article/pii/S0010482522008794

Highlights

- Three topoisomerase inhibitors can be distinguished from phase-contrast
microscopy images.

- Phase-contrast images are as good for classification as fluorescence images.

- Classification accuracy is better on images taken 72 h after treatment than
24
h.

- Classification is better when substance-specific concentrations are used.