Segmentation-Free Multi-Fractal and Radiomics Response Profiling of Live Unstained Confluent Microglia In Vitro Using...

This chapter presents a segmentation-free, cell ensemble-level workflow for detecting treatment-induced responses in confluent microglial cultures in vitro, using standard phase-contrast microscopy and Python-based imaging data analysis. Entire fields of view are analyzed after minimal preprocessing, avoiding fixation, staining, and single-cell segmentation. We implement multifractal, lacunarity, and texture analysis (MFTA) to characterize the microglial monolayers. In parallel, we introduce cell radiomics as the adaptation of radiomics principles to label-free microscopy, extracting a high-dimensional panel of intensity- and texture-based features directly from grayscale images of live cells. Together, MFTA and cell radiomics provide complementary, quantitative readouts that are sensitive to subtle changes in ensemble organization. Statistical comparisons employ false discovery rate control to identify robust discriminative features between conditions, with optional validation by conventional morphometry on a subset of segmented cells. The workflow is demonstrated on an experimental dataset by the successful reveal of the effect of a moderate static magnetic field on human C20 microglia cell line.