TCGA-ESCA Dataset: Whole Slide Images for Esophageal Cancer TNM Staging

A dataset of whole slide images (WSIs) for esophageal cancer TNM staging, evaluated using a novel deep learning model. The data likely originates from the TCGA-ESCA dataset. The model's performance was validated via three-fold cross-validation.

Use Cases

• Training automated TNM staging models based on whole slide images.
• Evaluating feature attention mechanisms for tumor region localization.
• Benchmarking multi-class classification methods for composite pathological stage prediction.
• Studying methods to address class imbalance in medical image datasets.

Strengths

• Model performance was rigorously evaluated using three-fold cross-validation.
• The approach includes a feature attention mechanism to pinpoint critical tumor areas.
• The study addresses pronounced TNM stage imbalance with a class-weighted loss function.

Limitations

• The dataset size is only 54.0 KB, indicating a very limited scope.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

TCGA-ESCA dataset

Collection Method

Whole slide images analyzed via a fine-tuned pathology foundation model within a Multiple Instance Learning framework.

Freshness

Last updated 2026-05-29 04:45:02; freshness should be verified.