BT-20 TNBC Spheroid Brightfield Images with Segmentation Outputs

Keith Ncube published a dataset of brightfield images of BT-20 triple-negative breast cancer spheroids on 2026-04-12. It includes 11.9 MB of images in JPG, TIF, and PNG formats, representing untreated and drug-treated conditions. The dataset also contains segmentation outputs from three analytical tools: SpheroidSizer, OrganoSeg2, and Ilastik.

Use Cases

• Benchmarking segmentation algorithm performance based on provided outputs from SpheroidSizer, OrganoSeg2, and Ilastik.
• Training or validating machine learning models for spheroid boundary detection based on brightfield images.
• Illustrating segmentation challenges in post-treatment imaging conditions based on the included drug-treated spheroid images.
• Reproducing analysis workflows using the provided ImageJ macro for processing Ilastik probability maps.

Strengths

• Includes segmentation outputs from three distinct analytical tools (SpheroidSizer, OrganoSeg2, Ilastik) for direct comparison.
• Provides images under two distinct biological conditions (untreated and drug-treated) to illustrate segmentation challenges.
• Dataset size is 11.9 MB, making it quick to download and inspect.

Limitations

• Row count and exact number of images are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect bias inherent to the specific cell line (BT-20) and imaging setup used.

Provenance

Source

Keith Ncube via figshare.

Collection Method

Created to illustrate segmentation behavior for a review article; method of original image capture is not specified.

Freshness

Last updated 2026-04-12 06:19:43; freshness should be verified.