Search for: All records

Breast carcinoma is the most common cancer among women worldwide that consists of a heterogeneous group of subtype diseases. The whole-slide images (WSIs) can capture the cell-level heterogeneity, and are routinely used for cancer diagnosis by pathologists. However, key driver genetic mutations related to targeted therapies are identified by genomic analysis like high-throughput molecular profiling. In this study, we develop a deep-learning model to predict the genetic mutations and biological pathway activities directly from WSIs. Our study offers unique insights into WSI visual interactions between mutation and its related pathway, enabling a head-to-head comparison to reinforce our major findings. Using the histopathology images from the Genomic Data Commons Database, our model can predict the point mutations of six important genes (AUC 0.68–0.85) and copy number alteration of another six genes (AUC 0.69–0.79). Additionally, the trained models can predict the activities of three out of ten canonical pathways (AUC 0.65–0.79). Next, we visualized the weight maps of tumor tiles in WSI to understand the decision-making process of deep-learning models via a self-attention mechanism. We further validated our models on liver and lung cancers that are related to metastatic breast cancer. Our results provide insights into the association between pathological image features, molecular outcomes, and targeted therapies for breast cancer patients.

 

The structure mapping task is a simple method to test people’s mental representations of spatial relationships, and has recently been particularly useful in the study of volumetric spatial cognition such as the spatial memory for locations in multilevel buildings. However, there does not exist a standardised method to analyse such data and structure mapping tasks are typically analysed by human raters, based on criteria defined by the researchers. In this article, we introduce a computational method to assess spatial relationships of objects in the vertical and horizontal domains, which are realized through the structure mapping task. Here, we reanalyse participants’ digitised structure maps from an earlier study (N=41) using the proposed computational methodology. Our results show that the new method successfully distinguishes between different types of structure map representations, and is sensitive to learning order effects. This method can be useful to advance the study of volumetric spatial cognition.