Imaging · dataset · 2026
<p>Workflow of the AI-mrEMVI segmentation model from manual annotation to clinical validation.</p>
Listed in UCL Research Data Repository
<p><b>(A)</b> Slice-by-slice manual annotation of multiplanar T2-weighted MR images performed by a radiologist, with annotated regions including the primary tumor (pink), intravascular tumor signals (green), and dilated vascular structures (purple). <b>(B)</b> Model Training and Testing Pipeline.
Description
The segmentation training dataset was divided into training and validation subsets at the plane-specific volume level using five-fold cross-validation.
The trained five-fold ensemble model was then applied to two external test cohorts recruited from 4 independent centers (n = 671). <b>(C)</b> Pairwise inter-reader consistency analysis presented as a Kappa heatmap across six radiologists of varying experience levels and the AI model. <b>(D)</b> Kaplan-Meier curves illustrating the prognostic value of AI-mrEMVI status for disease-free survival. AI, artificial intelligence; MRI, magnetic resonance imaging; EMVI, extramural vascular invasion; R, reader; Val, validation fold.</p>
Links
Where it is published
- DOI doi.org/10.1371/journal.pdig.0001763.g001 ↗
DOI / persistent id · from rdr ucl ac uk
Catalogue records · 1
- OAI-PMH record api.figshare.com/v2/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Af… ↗
metadata API · from rdr ucl ac uk
Topics
Provenance · 1 source records, 14 field assertions
| Source | Key | Last seen | Raw |
|---|---|---|---|
| UCL Research Data Repository | oai:figshare.com:article/34048943 | 5 d ago | JSON v1 |
| Field | Assertion | Extractor | Evidence |
|---|---|---|---|
| access_level | source · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | |
| concepts[disease].local:disease:cancer | mapping · rdr ucl ac uk | vocabulary-mapper@1.0.0 | keywords['Cancer'] |
| concepts[field].anzsrc:group:3101 | mapping · rdr ucl ac uk | vocabulary-mapper@1.0.0 | keywords['Cell Biology'] |
| concepts[field].local:field:earth-environmental | mapping · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | |
| concepts[field].local:field:life-sciences | mapping · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | |
| concepts[field].local:field:mathematics-statistics | mapping · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | |
| concepts[field].local:field:medicine-health | mapping · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | |
| concepts[modality].local:modality:image | enrichment · rdr ucl ac uk | keyword-concept-rules@1.0.0 | title+description (65%) |
| concepts[modality].local:modality:imaging | enrichment · rdr ucl ac uk | keyword-concept-rules@1.0.0 | title+description (75%) |
| concepts[modality].local:modality:mri | mapping · rdr ucl ac uk | vocabulary-mapper@1.0.0 | keywords['magnetic resonance imaging'] |
| description | source · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | /metadata/dc/description |
| license | source · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | /metadata/dc/rights |
| publication_date | source · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | |
| title | source · rdr ucl ac uk | connector:rdr_ucl_ac_uk@1.0.0 | /metadata/dc/title |