Susanna Fletcher Greer, Chief Scientific Officer at The V Foundation, shared on LinkedIn:
“One of the things I love most about cancer research is that innovation does not always mean inventing something entirely new. Sometimes, it means learning how to see more in what is already in front of us. A new study funded in part by the the V Foundation from grantee Dr. Michael Haffner and lab Fred Hutch and published in JCI Insightdoes exactly that.
Dr. Haffner and team developed an artificial intelligence model that can extract biological information from a routine pathology slide and use it to help identify a particularly aggressive form of metastatic prostate cancer. I love this application because most prostate cancers rely heavily on androgen-receptor signaling to grow, which is why therapies that block this pathway are so important. These treatments can be highly effective, until they are not.
We know cancer isn’t static. Under the pressure of therapy, tumors can adapt and change their identity. In some cases, prostate cancer can transform into neuroendocrine prostate cancer, which is an aggressive variant that grows rapidly, responds poorly to standard hormone therapies, and often requires different treatment strategies.
The challenge is that this transformation is not obvious. Only parts of a tumor may develop neuroendocrine features while other regions may resemble conventional prostate cancer, making early transformation difficult to identify. So the Haffner team asked a powerful question: could artificial intelligence detect these subtle changes directly from the same H&E-stained slides pathologists already use every day?
The team’s answer was to develop NEURAL-PC, an AI model trained not only to recognize visual patterns but also to focus on biologically meaningful characteristics of cells and nuclei, including their size, shape, texture, and organization.
When tested on biopsies, NEURAL-PC distinguished neuroendocrine prostate cancer from conventional prostate cancer with strong accuracy. Importantly, it could also identify small regions of neuroendocrine transformation within mixed tumors, suggesting the model may detect changes that could otherwise be difficult to recognize.
Even more compelling, the model’s predictions aligned with the team’s molecular data. Meaning tumors classified as neuroendocrine showed gene-expression patterns consistent with neuroendocrine prostate cancer biology. And, features detected by the model were also associated with differences in patient outcomes.
This is absolutely not a replacement for expert pathology or for molecular testing. Instead, this study points to an exciting possibility of AI acting as a second layer of vision, helping pathologists extract additional clinically meaningful information from tissue they are already examining.
The bottom line is that a routine pathology slide may contain far more information about how a cancer is evolving than we have previously been able to see. Sometimes progress in cancer research comes from developing a completely new treatment. And sometimes it comes from learning to see the disease more clearly. I love this so much.
Find the Haffner lab at Haffner Lab and read their paper at JCI Insight.”
Title: Deep learning–based histologic classifiers enable molecular subtyping of metastatic prostate cancer
Authors: Zhijun Chen, Erolcan Sayar, Daniela Guevara, Helen Richards, Haoyue Zhang, Radhika A. Patel, Agnes C. Gawne, Lucas J. Liu, Ilsa Coleman, Ruth Dumpit, Colm Morrissey, Michael T. Schweizer, Ruben Raychaudhuri, Laura S. Graham, Evan Y. Yu, Heather H. Cheng, Chien-Kuang C. Ding, Yuzhuo Wang, Peter Choyke, Baris Turkbey, Chantal Chanel-Vos, Christina Fedorov, John R. Otilano III, Troy Kane, Jyothi Manohar, Michael Sigouros, Jones T. Nauseef, Ana Molina, David Nanus, Scott T. Tagawa, Juan Miguel Mosquera, Himisha P. Beltran, Ruth Etzioni, Peter S. Nelson, Rama Soundararajan, Ana M. Aparicio, Cora N. Sternberg, Michael C. Haffner, and Stephanie A. Harmon.
Read the Full Article on JCI Insight.
Susanna Fletcher Greer was recognized among OncoDaily’s 100 Influential Women in Oncology in 2025, highlighting her contributions to the field of oncology.
Nominations for 100 Influential People in Oncology 2026 are now open, recognizing leaders, researchers, clinicians, innovators, advocates, and other individuals making a meaningful impact on oncology and cancer care.
Don’t miss the opportunity to nominate yourself, a colleague, or someone whose contributions deserve to be recognized.
Submit your nomination by September 15.
