Researchers at the University of California, Los Angeles (UCLA), in collaboration with pathologists from Hadassah Hebrew University Medical Center and the University of Southern California, have developed a deep learning–based technique that enables the digital generation of multiple immunohistochemical (IHC) stains from a single, unstained tissue section. This breakthrough approach, known as virtual multiplexed immunostaining, leverages artificial intelligence to predict the appearance of various protein markers on tissue samples without the need for physical staining. Traditionally, diagnosing cancer involves applying multiple IHC stains to different sections of a tissue sample to visualize specific proteins associated with disease progression, treatment response, and prognosis. This process is time-consuming, requires large amounts of tissue, and can be costly. The new method overcomes these limitations by using a deep neural network trained on paired data—unstained tissue images and their corresponding actual IHC stains—to learn the complex relationships between tissue morphology and protein expression. The model can then simulate the results of multiple stains from a single slide, effectively creating a virtual multiplexed image. This not only preserves precious tissue for further analysis but also allows for more comprehensive and consistent profiling of tumor characteristics. The technique has been validated on a range of cancer types, including breast, lung, and prostate cancers, and has shown high accuracy in predicting the spatial distribution of key biomarkers. The researchers emphasize that the method enhances diagnostic precision, supports personalized treatment planning, and could streamline workflows in clinical pathology. By reducing the need for multiple physical slides and reagents, the technology also lowers costs and resource use. The team is now working on integrating the tool into clinical environments and validating its performance across diverse patient populations and laboratory settings.