SpheroSeg: ÚTIA and UCT Prague develop a tool for analysing cell spheroids

A new publication in the journal Computer Methods and Programs in Biomedicine presents the results of a collaboration between researchers from the Institute of Information Theory and Automation of the Czech Academy of Sciences and the Faculty of Food and Biochemical Technology at the University of Chemistry and Technology, Prague. Together, they developed SpheroSeg, a platform using deep learning methods for the automated analysis of cell spheroids.

Cell spheroids are three-dimensional clusters of cells used as models of tumour tissue, for example to study tumour growth or evaluate treatment response. Analysing them using microscopy images can be time-consuming, and the results can be affected by differences in the conditions under which the images were acquired.

SpheroSeg can automatically identify and segment spheroids in bright-field microscopy images. Researchers can then manually adjust the results, obtain measurements of individual spheroid morphology, and export the data for further processing.

An important part of the project is also the SpheroHQ dataset, which contains 22,683 expert-annotated images of spheroids from seven cancer cell lines. In the study, the authors tested and compared eight different deep-learning architectures under various conditions.