Therapeutic response prediction

DiMAGE model for intrahepatic cholangiocarcinoma immunotherapy

Public preview
upload_file

1. Whole-slide image

Original application layout in read-only mode.

Whole-slide images are opened in the local application.

vaccines

2. Run prediction pipeline

Only ICI treatment response is enabled.

Upload0%
DiMAGE inference0%

Efficacy Profile Result

Sample name, response probability and final prediction.

Waiting

Sample name

—

P(Response)

—

P(Non-response)

—

Predicted ICI outcome

—

Visualization board

WSI, multi-scale heatmaps and immune marker maps

The layout is organized around the DiMAGE workflow: original H&E, 10×/20× attention, fused response heatmap, cell-density map, and PD-1 / PD-L1 virtual immunofluorescence outputs.

image Original H&E

Original thumbnail

Low-magnification overview of the intrahepatic cholangiocarcinoma WSI.

layers Fusion heatmap

DiMAGE fused response map

Integrated response-relevant spatial signal after multi-scale fusion.

zoom_in 20× attention

High-resolution local tumor-region attention.

travel_explore 10× attention

Broader tissue-context attention map.

scatter_plot Cell density

Cell-level spatial density distribution.

science PD-1

Virtual mIF PD-1 expression map.

science PD-L1

Virtual mIF PD-L1 expression map.

Core ICI biomarkers

Calculated in the local application.

analytics

Biomarker metrics appear after local analysis.