Back

Label-free classification of cell death pathways via holotomography-based deep learning framework

kim, m.; park, w. s.; kim, g.; oh, s.; do, j.; park, j.; Park, Y.

2025-06-08 cell biology
10.1101/2025.06.06.658404 bioRxiv
Show abstract

Accurate classification of cell death pathways is critical in understanding disease mechanisms and evaluating therapeutic responses, as dysregulated cell death underlies a wide range of pathological conditions including cancer and therapy resistance. Conventional imaging methods such as fluorescence and bright-field microscopy, or 2D phase imaging, often suffer from phototoxicity, labeling artifacts, or limited morphological contrast. Here, we present a real-time, label-free platform for classifying cell death phenotypes--apoptosis, necroptosis, and necrosis--by combining three-dimensional holotomography with deep learning. Our convolutional neural network, trained on refractive index (RI)-based features from HeLa cells, achieved high classification accuracy (97.2 {+/-} 2.8%) under varying cell densities. Notably, the model identified early RI changes during necroptosis several hours prior to fluorescence-based markers. These findings demonstrate the potential of holotomography-based AI for high-resolution, label-free cell death profiling.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.