BEAMSTER: Brain mEtAstases segMentation for STEreotactic Radiotherapy, A Retrospective MRI Dataset with Expert...

Area of Science:

• Radiology
• Medical Imaging
• Radiation Oncology

Background:

• Brain metastases are a significant clinical challenge.
• Accurate segmentation of metastatic lesions is crucial for effective radiotherapy planning.
• Existing datasets may lack sufficient detail for advanced computational analysis.

Purpose of the Study:

• To present a de-identified dataset of contrast-enhanced T1-weighted MRI scans from patients with brain metastases.
• To provide annotated metastatic lesions for developing and validating computer-aided detection and segmentation algorithms.
• To facilitate reproducible research in radiotherapy planning and clinical translation of AI tools.

Main Methods:

• Retrospective collection of 260 contrast-enhanced T1-weighted MRI scans from 140 patients with brain metastases.
• Annotation of metastatic lesions by an experienced radiation oncologist using a radiotherapy planning system.
• Conversion of binary segmentation masks to NIfTI format and de-identification of all patient data.

Main Results:

• A comprehensive dataset of brain metastasis MRI scans is now available.
• The dataset includes annotations for 260 metastatic lesions, with a focus on small, clinically relevant lesions.
• All data is de-identified and privacy-preserving, including removal of facial features.

Conclusions:

• This dataset is a valuable resource for advancing computer-aided detection and segmentation methods in neuro-oncology.
• The inclusion of small lesions addresses a key challenge in automated analysis.
• The dataset will support the development of robust algorithms for clinical radiotherapy planning.