Prostate MRI (ภาพการตรวจคลื่นแม่เหล็กไฟฟ้าของต่อมลูกหมาก) (Public Sharing)

"The Prostate MRI dataset was developed through a structured pipeline involving data extraction, cleansing/ verification, and annotation. Imaging data were retrospectively collected from the hospital’s PACS system. Initial steps included de-identification to ensure patient privacy, followed by image quality assessment to exclude non-diagnostic or corrupted scans.

Annotation was performed on a subset of studies by expert radiologists, including PI-RADS scoring and manual segmentation of prostate zonal anatomy, seminal vesicle, urinary bladder, and prostate lesions. Annotations adhered to international standards such as SNOMED-CT, RadLex, PI-RADS, and DICOM protocols.

The data are categorized by modality, imaging sequence (e.g., T2WI, DWI, ADC, DCE), and contain linked clinical metadata where applicable.

This curated dataset is optimized for clinical research, education, and AI model development (e.g., classification and segmentation) in prostate cancer diagnostics and treatment planning, as well as image quality or other purposes. "