Image entropy-based label-free functional characterization of human induced pluripotent stem cell-derived 3D cardiac spheroids

Human induced pluripotent stem cell-derived cardiac spheroids (iPSC-CSs) in 3D possess tremendous potential for treating heart diseases and screening drugs for their cardiac effect. The beating pattern (including beating frequency and amplitude) of iPSC-CSs is a direct indicator of their health and function. However, detecting the beating pattern of 3D cardiac spheroid is not well studied and the probes commonly used for labeling cardiomyocytes for their beating pattern detection is toxic during long-term culture. Here, we reveal that the beating pattern of 3D iPSC-CSs can be conveniently detected/quantified by calculating the relative change of entropy in all the frames/images of non-fluorescent optical signal without labeling any cells. The entropy rate superpixel segmentation method is used for image segmentation in frames containing multiple or aggregated iPSC-CSs to identify individual iPSC-CSs, enabling rapid detection/quantification of the beating pattern of each iPSC-CS. Moreover, the responses of iPSC-CSs to both anticancer and cardiac drugs can be reliably detected with the image entropy-based label-free method in terms of their beating patterns. This novel label-free approach may be valuable for convenient and efficient functional evaluation of 3D and 2D cardiac constructs, which is important not only for drug screening but also the advancement of manufacturing functional cardiac constructs to treat heart diseases.