English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Definitive cancer diagnosis and management depend upon the extraction ofinformation from microscopy images by pathologists. These images containcomplex information requiring time-consuming expert human interpretation thatis prone to human bias. Supervised deep learning approaches have provenpowerful for classification tasks, but they are inherently limited by the costand quality of annotations used for training these models. To address thislimitation of supervised methods, we developed Histomorphological PhenotypeLearning (HPL), a fully blue{self-}supervised methodology that requires noexpert labels or annotations and operates via the automatic discovery ofdiscriminatory image features in small image tiles. Tiles are grouped intomorphologically similar clusters which constitute a library ofhistomorphological phenotypes, revealing trajectories from benign to malignanttissue via inflammatory and reactive phenotypes. These clusters have distinctfeatures which can be identified using orthogonal methods, linking histologic,molecular and clinical phenotypes. Applied to lung cancer tissues, we show thatthey align closely with patient survival, with histopathologically recognisedtumor types and growth patterns, and with transcriptomic measures ofimmunophenotype. We then demonstrate that these properties are maintained in amulti-cancer study. These results show the clusters represent recurrent hostresponses and modes of tumor growth emerging under natural selection. Code,pre-trained models, learned embeddings, and documentation are available to thecommunity athttps://github.com/AdalbertoCq/Histomorphological-Phenotype-Learning

Mapping the landscape of histomorphological cancer phenotypes using self-supervised learning on unlabeled, unannotated pathology slides