3D cell shape analysis in cancer cell xenografts in zebrafish larvae. We present an experimental and analytical pipeline to identify single-cell morphologies within tissues of zebrafish larvae and identify differences in morphological cell states, referred to as morphotypes, by computer-vision analysis of 3D cell shape, as follows. (1) In order to distinguish morphotypes of individual cells, we prepared a mosaic cell mixture consisting of 90–99% of cells labeled with a green-emitting fluorophore and 1–10% of cells labeled with a red-emitting fluorophore, or vice versa. We then injected the cell mixture into the PVS or HBV of 2-d-old zebrafish larvae. For PVS injections, we selected specifically for injected larvae which also had xenografted cells in the CHT immediately following injection. Fish larvae were imaged live or fixed 24 h after injection using an ASLM, which is optimized for maximum throughput, resolution, and flexibility of sample orientation. High axial resolution is obtained by combining a thin, moving light-sheet with a synchronized rolling shutter (active pixels) that blocks out of focus fluorescence on the image plane on an sCMOS camera. (2) Individual cells were computationally extracted from the image data of sparsely labeled cells, and after image processing were segmented into 3D volumes. We then measured 12 global 3D geometric features as listed, and each cell was indexed according to fish, location, and cell type. (3) Using PCA we captured the heterogeneity of morphotypes (N = number of fish, n = number of cells). PC1 captured the protrusive state of the cell, while PC2 was influenced largely by volume. Insets show 3D renderings of segmented cell volumes indicated in the PC plot. We evaluated our analytical pipeline and used several visualization and classification techniques to capture differences among heterogeneous single-cell data. Scale bar, 20 µm.