LazySlide Connects Histopathology to Omics Data and Operations

Microscopic images of human tissues are essential in biomedical research and clinical care, yet they remain difficult to analyze systematically or connect to other biological data sources. A study led by André Rendeiro at CeMM, published in Nature Methods, introduces LazySlide, an open-source software that connects histopathology to omics data and computational operations. This tool brings large-scale image analysis closer to the integrated workflows already used in genomics and single-cell biology. 

Digital microscopy now enables researchers to scan an entire tissue sample into a detailed whole-slide image, capturing structures from the organ level down to individual cells. These data-rich images hold tremendous potential, but their size, complex formats, and lack of compatible tools have limited their use in combined datasets. Unlike genetics, which has standardized methods for data sharing, digital pathology often operates in isolation, preventing tissue-level insights from being linked to molecular profiles such as RNA sequencing.

LazySlide addresses this disconnect by dividing slide images into smaller regions and analyzing them with advanced AI models. These models can identify tissue patterns, classify cell types, and quantify subtle changes in structure—tasks that previously required extensive manual effort. Rendeiro’s team demonstrated that LazySlide could distinguish healthy from calcified artery tissues using only image data, and that integrating these visual features with gene expression profiles revealed underlying biological pathways, including inflammatory signaling, associated with disease.

“Histology contains an enormous amount of biological information, but it is often difficult to access computationally,” notes first author Yimin Zheng. “With LazySlide, we wanted to provide a tool that allows researchers to explore tissue images in a systematic, quantitative way and to connect what they see under the microscope with underlying molecular processes.”

LazySlide also enables language-based searches within images. By linking image features to text concepts, the software allows users to locate visual patterns—such as “calcification”—and return measurable results. This capability supports “zero-shot” analysis, meaning LazySlide can identify tissues or disease states without task-specific training, making automated image analysis more accessible to a wider range of researchers.