Related Experiment Video
Updated: May 13, 2026
A Robust Method for the Large-Scale Production of Spheroids for High-Content Screening and Analysis Applications
Published on: December 28, 2021
Near-infrared imaging-based high-content analysis for label-free assessment of internal cellular heterogeneity in
Ren Sakai1, Kenjiro Tanaka1, Miya Kawasaki1
1Department of Basic Medicinal Sciences, Graduate School of Pharmaceutical Sciences, Nagoya University, Tokai National Higher Education and Research System, Furocho, Chikusa-ku, Nagoya, Aichi 464-8601, Japan.
Abstract:
Non-invasive analysis of spheroid quality was essential because spheroids better recapitulated in vivo-like cell-cell and cell-extracellular matrix (ECM) interactions than two-dimensional cultures. However, routine non-invasive methods to assess internal structural heterogeneity remained limited. Here, we developed a near-infrared imaging-based high-content analysis (NIR-HCA) approach to non-destructively quantify internal cellular heterogeneity at the level of individual spheroids. A custom-built imaging system operating at wavelengths above 1500 nm was designed to acquire 65 wavelength-resolved transmittance images per spheroid in a single shot. From each image, transmittance-based 38 descriptors were extracted, including global features (whole-spheroid statistics, intensity-gradient features, and gray-level co-occurrence matrix textures) and regional features (partition-wise statistics), yielding 2470 features (38 × 65) per spheroid. Using human fibroblast spheroids as a model system, we evaluated internal cellular quality differences by NIR-HCA under two complementary scenarios: an ECM-supplemented model and an ECM-overproducing model. In the ECM-supplemented model, NIR-HCA non-destructively detected pronounced heterogeneous cellular organization in real time, including localized cell-dense and cell-sparse regions, and a classification model for these internal structural differences achieved a macro F1-score of 0.96. In the ECM-overproducing model, NIR-HCA sensitively discriminated spheroids with higher myofibroblast mixing ratios despite identical total cell numbers, and a classification model for these differences achieved a macro F1-score of 0.82. In contrast, conventional microscopy or histological staining could not reliably resolve these differences. Collectively, these results demonstrated that NIR-HCA enabled quantitative image-based, in-process monitoring of spheroid quality at the individual-spheroid level.
