%0 Journal Article
%A Aldera, Alessandro Pietro
%A Cifci, Didem
%A Veldhuizen, Gregory Patrick
%A Tsai, Wan-Jung
%A Pillay, Komala
%A Boutall, Adam
%A Brenner, Hermann
%A Hoffmeister, Michael
%A Kather, Jakob Nikolas
%A Ramesar, Raj
%T Deep learning predicts microsatellite instability status in colorectal carcinoma in an ethnically heterogeneous population in South Africa.
%J Journal of clinical pathology
%V nn
%@ 0021-9746
%C London
%I BMJ Publ. Group
%M DKFZ-2025-01042
%P nn
%D 2025
%Z epub
%X Deep learning (DL) models are effective pre-screening tools for detecting mismatch repair deficiency (dMMR) in colorectal carcinoma (CRC). These models have been trained and validated on large cohorts from the Northern Hemisphere, without representation of African samples. We sought to determine the performance of a DL model in an ethnically heterogeneous cohort of patients from South Africa.Our cohort comprised 197 CRC resection specimens, with scanned whole slide images tessellated and inputted into a transformer-based DL model trained on large international cohorts. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC), sensitivity and specificity. The maximal Youden's J index was calculated to determine the optimal cut-off threshold for the model prediction score.Our model yielded an AUROC of 0.91 (±0.05). Using a prediction score threshold of 0.620 produced an overall sensitivity of 85.7
%K Artificial Intelligence (Other)
%K Colorectal Neoplasms (Other)
%K Medical Oncology (Other)
%K Neoplastic Syndromes, Hereditary (Other)
%F PUB:(DE-HGF)16
%9 Journal Article
%$ pmid:40393786
%R DOI:10.1136/jcp-2025-210053
%U https://inrepo02.dkfz.de/record/301503