Fig. 6: Comparison of 2D and 3D features.

a Heatmap representation of correlations between features. Correlation between the selected features was tested using Spearman’s correlation method. Purple colour represents the monoculture dataset (n = 223 for 2D and n = 110 for 3D) and the orange colour represents the features collected from the co-culture dataset (n = 203 for 2D and n = 114 for 3D). b Boxplot representation of the 2D and 3D features. Thirty-seven spheroids were selected and verified that the 2D measurements correspond to the same objects in 3D. Features such as Volume 3D (µm³), Solidity (−), Sphericity (−), and Extent (−) were extracted from both the 2D and 3D images and used for comparison. Results are displayed as Monoculture 2D (light purple, M-2D), Monoculture 3D (dark purple, M-3D), Co-culture 2D (light orange, C-2D), and Co-culture 3D (dark orange, C-3D). The Kolmogorov-Smirnov test was used. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. Each group showed a statistically significant difference (p < 0.001). c Percentage difference from the average values between the 2D and 3D features are shown. Purple colour represents the monoculture dataset, orange colour represents the co-culture dataset (n = 37). d Boxplot representation of the number of nuclei in the monoculture dataset for each experiment (n ≥ 7). Significant differences were not included. Correlation was measured between the number of nuclei, Volume 3D, and Area. e Boxplot representation of the number of nuclei in the co-culture dataset for each experiment (n ≥ 7). Significant differences were not included. Correlation is measured between the number of nuclei, Volume 3D, and Area. Box plots were visualised using Tukey’s method, with medians shown as central lines, boxes representing interquartile ranges, and whiskers extending to 1.5× interquartile range. All scatter plots were generated using a fitted regression line, with shaded bands indicating the standard error of the mean (SEM). Source data are provided as a Source Data file.