Dataset Open Access

Trained deep neural networks for MSI/dMMR detection in colorectal cancer histology

Kather, Jakob Nikolas; Ghaffari Laleh, Narmin

These are trained neural network models in PyTorch format to process tessellated images of colorectal cancer histology samples. The input is expected to be 224x224 px RGB image tiles normalized with the Macenko method. The output is a probability of the image tile for being MSI/dMMR or MSS/pMMR. 

The models have been trained on eight cohorts but not on the validation cohort. The validation cohorts are:

Ex_0  : DACHS

Ex_1  :  DUSSEL

Ex_2  :  MECC

Ex_3  : QUASAR

Ex_4  : RAINBOW

Ex_5  : TCGA

Ex_6  : UMM

Ex_7  : YORKSHIRE

Ex_8  : MUNICH

The models can be loaded in Python with 

>>> model = torch.load(path, map_location=torch.device('cpu'))

Further details are given in the manuscript.

Files (403.2 MB)
Name Size
Exp0_MODEL_Full
md5:05406ba26560f1d2a13b5bb9732ab1c8
44.8 MB Download
Exp1_MODEL_Full
md5:32467671864bd0fb8cfc1d6f537c74e8
44.8 MB Download
Exp2_MODEL_Full
md5:7632b4c1cde18262e0add152c85e4128
44.8 MB Download
Exp3_MODEL_Full
md5:213dde6c5fbc0f98a223f77a9fe3c04a
44.8 MB Download
Exp4_MODEL_Full
md5:3d0871a001c0820ed10851df204da1bd
44.8 MB Download
Exp5_MODEL_Full
md5:4dfb35dc9a8dd8c32a750da1c6a145d4
44.8 MB Download
Exp6_MODEL_Full
md5:01a9e874ef25f97b18a7de456a3bb8ef
44.8 MB Download
Exp7_MODEL_Full
md5:fa3472ba06b701e76c4fd74207f064f7
44.8 MB Download
Exp8_MODEL_Full
md5:77941b7ab9713fbb9c62fdebb0b14e16
44.8 MB Download
164
84
views
downloads
All versions This version
Views 164164
Downloads 8484
Data volume 3.8 GB3.8 GB
Unique views 137137
Unique downloads 2828

Share

Cite as