TY - JOUR
T1 - Pan-cancer image-based detection of clinically actionable genetic alterations
AU - Kather, Jakob Nikolas
AU - Heij, Lara R.
AU - Grabsch, Heike I.
AU - Loeffler, Chiara
AU - Echle, Amelie
AU - Muti, Hannah Sophie
AU - Krause, Jeremias
AU - Niehues, Jan M.
AU - Sommer, Kai A. J.
AU - Bankhead, Peter
AU - Kooreman, Loes F. S.
AU - Schulte, Jefree J.
AU - Cipriani, Nicole A.
AU - Buelow, Roman D.
AU - Boor, Peter
AU - Ortiz-Bruechle, Nadina
AU - Hanby, Andrew M.
AU - Speirs, Valerie
AU - Kochanny, Sara
AU - Patnaik, Akash
AU - Srisuwananukorn, Andrew
AU - Brenner, Hermann
AU - Hoffmeister, Michael
AU - van den Brandt, Piet A.
AU - Jaeger, Dirk
AU - Trautwein, Christian
AU - Pearson, Alexander T.
AU - Luedde, Tom
PY - 2020/8
Y1 - 2020/8
N2 - Molecular alterations in cancer can cause phenotypic changes in tumor cells and their microenvironment. Routine histopathology tissue slides, which are ubiquitously available, can reflect such morphological changes. Here, we show that deep learning can consistently infer a wide range of genetic mutations, molecular tumor subtypes, gene expression signatures and standard pathology biomarkers directly from routine histology. We developed, optimized, validated and publicly released a one-stop-shop workflow and applied it to tissue slides of more than 5,000 patients across multiple solid tumors. Our findings show that a single deep learning algorithm can be trained to predict a wide range of molecular alterations from routine, paraffin-embedded histology slides stained with hematoxylin and eosin. These predictions generalize to other populations and are spatially resolved. Our method can be implemented on mobile hardware, potentially enabling point-of-care diagnostics for personalized cancer treatment. More generally, this approach could elucidate and quantify genotype-phenotype links in cancer. Two papers by Kather and colleagues and Gerstung and colleagues develop workflows to predict a wide range of molecular alterations from pan-cancer digital pathology slides.
AB - Molecular alterations in cancer can cause phenotypic changes in tumor cells and their microenvironment. Routine histopathology tissue slides, which are ubiquitously available, can reflect such morphological changes. Here, we show that deep learning can consistently infer a wide range of genetic mutations, molecular tumor subtypes, gene expression signatures and standard pathology biomarkers directly from routine histology. We developed, optimized, validated and publicly released a one-stop-shop workflow and applied it to tissue slides of more than 5,000 patients across multiple solid tumors. Our findings show that a single deep learning algorithm can be trained to predict a wide range of molecular alterations from routine, paraffin-embedded histology slides stained with hematoxylin and eosin. These predictions generalize to other populations and are spatially resolved. Our method can be implemented on mobile hardware, potentially enabling point-of-care diagnostics for personalized cancer treatment. More generally, this approach could elucidate and quantify genotype-phenotype links in cancer. Two papers by Kather and colleagues and Gerstung and colleagues develop workflows to predict a wide range of molecular alterations from pan-cancer digital pathology slides.
KW - COMPREHENSIVE MOLECULAR CHARACTERIZATION
KW - GENOMIC CHARACTERIZATION
KW - COLORECTAL-CANCER
KW - MUTATIONS
KW - SUBTYPES
KW - HEAD
U2 - 10.1038/s43018-020-0087-6
DO - 10.1038/s43018-020-0087-6
M3 - Article
C2 - 33763651
VL - 1
SP - 789
EP - 799
JO - Nature Cancer
JF - Nature Cancer
SN - 2662-1347
IS - 8
ER -