Fig. 2: LeGO metastasis and vessel analysis pipeline.
From: LeGO-3D: 3D imaging of lung metastases and vascularisation using light sheet fluorescence microscopy
a Analysis workflow using semi-automatic and automatic Fiji Macro and Python scripts. Created in BioRender. Lewis, S. (2025) https://BioRender.com/g0il646. b Each image channel is segmented using Ilastik. Manual training is performed before automatic interpolation across the entire lung is completed. Representative image of a metastasis-containing lung, acquired on a light sheet microscope. TSaphhire (cyan), Venus (yellow), TdTomato (magenta) and blood vessel (grey) channels, as well as overlay image is shown. Original and segmented image is shown. Scale bar 1000 μm. c LeGO BPs are automatically classified across thousands of metastases in lungs. Representation SVT metastasis (left), LeGO BPs classification of SVT metastasis (middle) and LeGO BPs correction of SVT metastasis (right). Scale bar, 50 μm. d, e Automatic measurement of the metastasis’ nearest vessel and diameter of that vessel is performed. d Representative local thickness map shows diameter of vessels in lung. Brightest pixels represent larger vessel diameters, whereas dimmer pixels represent smaller vessel diameters (μm). White pixels represent vessels that are 150 μm or larger. Scale bar, 1000 μm. e Representative local distance map shows distance to vessels in lung. Brightest pixels represent nearer vessels, whereas dimmer pixels represent further vessels. White pixels represent distances that are 180 μm or further. Scale bar, 1000 μm.
