LEOPARD: Prostate Cancer Recurrence Prediction from Histopathology Slides

A challenge dataset from Radboud University Medical Center for predicting biochemical recurrence in prostate cancer patients. The data likely contains H&E-stained histopathology slides and associated clinical outcomes, such as time to recurrence and risk factors like ISUP grade. It is designed to develop deep learning models that can identify prognostic morphological features from tissue sections.

Use Cases

• Predict time to biochemical recurrence based on morphological features in H&E-stained slides.
• Develop deep learning models to uncover fine-grained prognostic tissue patterns beyond ISUP grade.
• Benchmark AI solutions for prostate cancer prognosis against established clinical risk factors.
• Study the relationship between histopathology image features and post-prostatectomy PSA levels.

Strengths

• Addresses a clinically significant problem, with prostate cancer impacting 1.4 million men annually.
• Focuses on predicting biochemical recurrence, which affects approximately 30% of post-prostatectomy patients.
• Built for a specific challenge, suggesting a curated and task-oriented structure.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Radboud University Medical Center

Collection Method

Likely gathered from clinical pathology workflows and patient records for a research challenge.

Time Range

null

Freshness

null

Geography

null