A Multimodal Dataset for Survival Prediction in Resected Pancreatic Ductal Adenocarcinoma
Ranking
No observed public metrics; popularity remains neutral/archived.
Merged summary
TL;DR - Researchers released a multimodal dataset linking pancreatic cancer histology, clinical variables, targeted sequencing, and long-term survival outcomes for 302 resected PDAC patients. Initial benchmarks show clinical variables outperform image-only and neural multimodal models, providing a baseline for future externally validated survival prediction.
- The dataset includes 446 H&E whole-slide images, clinicopathological data, survival outcomes, and targeted sequencing for 154 patients.
- Ridge Cox regression achieved mean concordance of 0.649, rising slightly to 0.652 with KRAS and TP53 mutation status.
- The image-only attention model achieved 0.603 concordance; multimodal fusion led the neural models at 0.619.
- The retrospective, single-centre cohort requires external validation before broader clinical applicability can be established.
Sources (1)
A Multimodal Dataset for Survival Prediction in Resected Pancreatic Ductal Adenocarcinoma
TL;DR - Researchers released a multimodal dataset linking pancreatic cancer histology, clinical variables, targeted sequencing, and long-term survival outcomes for 302 resected PDAC patients. Initial benchmarks show clinical variables outperform image-only and neural multimodal models, providing a baseline for future externally validated survival prediction.
- The dataset includes 446 H&E whole-slide images, clinicopathological data, survival outcomes, and targeted sequencing for 154 patients.
- Ridge Cox regression achieved mean concordance of 0.649, rising slightly to 0.652 with KRAS and TP53 mutation status.
- The image-only attention model achieved 0.603 concordance; multimodal fusion led the neural models at 0.619.
- The retrospective, single-centre cohort requires external validation before broader clinical applicability can be established.