Johns Hopkins University astrophysicist Alexander Szalay, Ph.D., and Johns Hopkins Medicine Department of Dermatology and Sidney Kimmel Comprehensive Cancer Center pathologist Janis Taube, M.D., M.Sc., were chosen for a research project in Anthropic's AI for Science program. They were selected for their pioneering AstroPath technology, which combines astronomy, pathology and artificial intelligence (AI) to create detailed maps of tumors, revealing how cancer cells interact with surrounding cells, particularly immune cells.

AstroPath was initially developed with seed funding from the Johns Hopkins Sidney Kimmel Comprehensive Cancer Center's Bloomberg~Kimmel Institute for Cancer Immunotherapy, followed by significant multiyear support from the Mark Foundation for Cancer Research that helped expand the technology and its applications in cancer research.

The Anthropic AI for Science program supports high-impact scientific research with the potential to accelerate discovery through advanced AI technologies. The initiative provides selected researchers with free credits to access Anthropic’s Claude AI models, which can help scientists analyze complex data, identify patterns, develop hypotheses and accelerate scientific discovery. The program initially focused on biology and life sciences but has expanded to other scientific disciplines.

“We are incredibly fortunate to have pathbreaking researchers like Drs. Szalay and Taube at Johns Hopkins who are harnessing the power of data science and artificial intelligence to advance our founding mission of improving and saving lives,” says Johns Hopkins University President Ronald J. Daniels, who launched the university’s Data Science and AI Institute in 2023. “This award duly recognizes their extraordinary achievements, including effectively and responsibly harnessing the profound capacity of AI to analyze vast amounts of information and help shape the future of scientific discovery.”

AstroPath is an innovative cancer imaging and mapping platform developed by Szalay and Taube that applies technology originally designed to map the universe to the microscopic world of cancer. Basing the data infrastructure on the one developed for the Sloan Digital Sky Survey, which mapped billions of stars and galaxies, AstroPath maps the locations and interactions of cancer cells, immune cells and other cells within tumors. These detailed maps help researchers better understand how cancer interacts with the immune system, identify biomarkers that predict which patients are most likely to benefit from immunotherapy, and uncover new opportunities for cancer treatment.

“Much like earlier waves of technological innovation transformed medical practice, AI is poised to redefine how we deliver care, advance research and educate the next generation of health care leaders,” says Theodore DeWeese, M.D., the Frances Watt Baker, M.D., and Lenox D. Baker, Jr., M.D., dean of the medical faculty and CEO, Johns Hopkins Medicine. “We are applying AI across all three pillars of our mission at Johns Hopkins Medicine to push the boundaries of what’s possible. Tools like these are just the beginning.”

The Mark Foundation's multiyear investment of more than $10 million, together with continued support from the Bloomberg~Kimmel Institute, led to the establishment of the Mark Foundation Center for Advanced Genomics and Imaging at Johns Hopkins in 2020. This support allowed the team to substantially expand AstroPath and undertake the large-scale collection of detailed tumor data. These rich datasets, capturing the complex interactions between cancer cells and their surrounding environment, are now central to the research that attracted Anthropic's interest.

“It’s a convergence of scientific technology, big data and astronomy, aimed at cancer,” says Szalay, Bloomberg Distinguished Professor of Data, director of the Institute for Data-Intensive Engineering in Science at The Johns Hopkins University, and co-investigator at the Mark Foundation Center for Advanced Genomics and Imaging at Johns Hopkins.

The AstroPath team is now using AI to accelerate the creation and analysis of these detailed cancer maps. Tasks that once required researchers to painstakingly identify and outline tumor cells by hand can now be completed in minutes using AI-assisted tools. The result is massive, high-quality datasets that allow researchers to study the complex relationships between cancer cells and the immune system with greater speed and precision.

“These are supervised tools,” says Taube, the George W. Hambrick, Jr., M.D. and Thomas G. Olsen, M.D. professor of dermatology and pathology at the Johns Hopkins University School of Medicine, co-director of the Johns Hopkins Sidney Kimmel Comprehensive Cancer Center’s Bloomberg~Kimmel Institute for Cancer Immunotherapy Tumor Microenvironment Program and the Mark Foundation Center for Advanced Genomics and Imaging. “We still oversee the process, but AI allows us to generate datasets much faster and with far more detail than we could by hand.”

The researchers are also using AI to combine spatial maps of tumors with genetic and molecular information, creating a more comprehensive picture of the biological processes that underlie response to immunotherapy.

“Technologies like AstroPath have the potential to revolutionize how we diagnose and treat cancers, and in this case, help us figure out why immunotherapy works remarkably well for some patients but not others,” says William Nelson, M.D., Ph.D., D.Sc., the Marion I. Knott Professor and director of the Sidney Kimmel Comprehensive Cancer Center. “Anthropic’s recognition of this pioneering work will help accelerate these discoveries. We have already surpassed what was possible just a decade ago. Imagine what the next decade could bring.”

The Anthropic selection provides an opportunity to expand this work, harnessing AI to accelerate the analysis of complex cancer data and potentially transform the discovery of new biomarkers, treatments and more personalized approaches to cancer care. Building on the extensive datasets generated through years of interdisciplinary research, the team is pursuing new questions about the tumor microenvironment and the biological mechanisms that influence treatment response.

“The scale and richness of the data are what make this next phase of AstroPath so exciting,” says Taube. “The Mark Foundation’s support, along with continued investment from the Bloomberg~Kimmel Institute, allowed us to expand our work and generate datasets that simply wouldn’t have been possible otherwise. Now, with the addition of powerful AI tools, we have an opportunity to extract insights from those data at a scale and speed that could fundamentally change how we study cancer.”

Szalay and Taube are also exploring opportunities to extend AstroPath’s impact beyond academic research, including the potential formation of a new company focused on AI-powered tumor analysis and biomarker development. The envisioned effort would combine AstroPath’s computational platform with the development and performance of biomarker assays, with plans to offer these assays to the clinical community while also providing pharmaceutical and biotechnology companies with new ways to understand tumor biology, evaluate treatment response and support the development of precision cancer therapies.

Anthropic launched its AI for Science program in May 2025 to support researchers conducting potentially transformative scientific investigations. Projects are selected based on their scientific contributions, potential impact and the ability of AI to meaningfully accelerate their research.