Computational methods have made substantial progress in improving the accuracy and throughput of pathology workflows for diagnostic, prognostic, and genomic prediction. Still, lack of interpretability remains a significant barrier to clinical integration. We present an approach for predicting clinically-relevant molecular phenotypes from whole-slide histopathology images using human-interpretable image features (HIFs). Our method leverages >1.6 million annotations from board-certified pathologists across >5700 samples to train deep learning models for cell and tissue classification that can exhaustively map whole-slide images at two and four micron-resolution. Cell- and tissue-type model outputs are combined into 607 HIFs that quantify specific and biologically-relevant characteristics across five cancer types. We demonstrate that these HIFs correlate with well-known markers of the tumor microenvironment and can predict diverse molecular signatures (AUROC 0.601–0.864), including expression of four immune checkpoint proteins and homologous recombination deficiency, with performance comparable to ‘black-box’ methods. Our HIF-based approach provides a comprehensive, quantitative, and interpretable window into the composition and spatial architecture of the tumor microenvironment.
An increasingly large body of work suggests that atypical adenomatous hyperplasia (AAH) of the lung may be a forerunner of pulmonary adenocarcinoma. Recognizing this fact, the World Health Organization now acknowledges the existence of AAH while noting difficulties that may be encountered in distinguishing AAH from the nonmucinous variant of bronchioloalveolar carcinoma. Regrettably, a universally acceptable definition of morphologic criteria for the diagnosis of AAH has not been achieved. This review of the literature examines the epidemiology, gross appearance, light microscopic findings, morphometry, immunohistochemistry, and molecular features of AAH and suggests a set of histopathologic features that may help the practicing pathologist identify this intriguing lesion. These features include the following: irregularly bordered focal proliferations of atypical cells spreading along the preexisting alveolar framework; prominent cuboidal to low columnar alveolar epithelial cells with variable degree of atypia but less than that seen in adenocarcinoma; increased cell size and nuclearcytoplasmic ratio with hyperchromasia and prominent nucleoli, generally intact intercellular attachment of atypical cells with occasional empty-looking spaces between them without high cellularity and without tufting or papillary structures; and slight thickening of the alveolar walls on which the AAH cells have spread, with some fibrosis but without scar formation or significant chronic inflammation of the surrounding lung tissue. Several lines of evidence indicate that AAH is a lesion closely associated with adenocarcinoma of the lung, suggesting AAH may be involved in the early stage of a complex multistep carcinogenesis of pulmonary adenocarcinoma.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.