Advancing Pancreatic Cancer Prognosis with AI-Driven Spatial Analysis of Pathology Slides

ROCHESTER, Minn. — Researchers at Mayo Clinic have unveiled a groundbreaking use of artificial intelligence (AI) in evaluating pathology slides, aiming to refine the prognostic capabilities for patients battling pancreatic cancer. This approach is designed to discern patterns that could pinpoint patients at heightened risk for recurrence following surgical intervention and treatment.

A study detailed in Clinical Cancer Research illuminates the significance of analyzing the organization of remaining cancerous cells rather than merely their quantity. Remarkably, patients exhibiting a fragmented interplay between tumor and surrounding tissue tended to experience quicker recurrence, regardless of the overall mass of residual cancer.

According to Ryan Carr, M.D., Ph.D., an oncologist at Mayo Clinic and senior author of the study, "Current pathology assessments largely tell us how much tumor is left after treatment. We wanted to know whether the geography of that remaining cancer could reveal additional biology about recurrence risk."

Analyzing Cancer Geography 

The study evaluated tissue from 203 patients diagnosed with pancreatic ductal adenocarcinoma. These patients underwent treatment prior to surgery but exhibited limited pathologic responses. Researchers integrated an AI-powered digital pathology system with landscape ecology principles to scrutinize standard hematoxylin and eosin (H&E) slides. Their analysis focused on tissue morphology, fragmentation, and the interaction between cancer cells and the stroma.

This initiative builds upon Dr. Carr's expansive research into applying ecological theories to cancer studies. His team employs machine learning and spatial assessments to explore the complex environment of pancreatic cancer, examining the relationships between malignant and surrounding cells and how these dynamics might affect treatment responses and recurrence.

Two identified spatial patterns have been associated with disease-free survival, independent of conventional risk factors like tumor stage and lymph node involvement. For instance, one of the models indicated that high-risk patients faced a 71% higher adjusted likelihood of recurrence, while another model suggested that high-risk patients had more than double the adjusted risk. This spatial analysis provided insights beyond what traditional metrics could reveal.

Significantly, the methodology leverages routine pathology slides generated as part of patient care, offering a potential advantage for clinicians aiming to assess recurrence risk without necessitating additional tissue sampling.

Connecting Patterns to Immune Response 

"What is exciting is that this information is already present in the tissue," Dr. Carr notes, emphasizing that AI's analytical power can unveil features that are traditionally challenging to observe, thus enhancing accuracy in recurrence risk assessments.

The findings also showed that high-risk spatial configurations harbored fewer immune cells within the tumor, with these cells tending to localize around rather than infiltrate the tumor. This observation highlights the tumor microenvironment's pivotal role in influencing treatment resistance and cancer behavior.

The research aligns with Mayo Clinic's initiative to harness data and technology for early risk prediction, aimed at preemptively addressing severe diseases. Dr. Carr remarks, "Our long-term goal is to better identify which patients remain at greatest risk and use that knowledge to refine individualized treatment strategies."

While the findings are promising, they necessitate further validation in prospective studies before becoming part of routine clinical practice. Funding for the research was partly provided by various Mayo Clinic awards and the ARPA-H ADAPT program. For a complete list of authors and funding disclosures, consult the study.

About Mayo Clinic 
Mayo Clinic is dedicated to advancing clinical practice, education, and research, focusing on compassion and expertise in patient care. Visit the Mayo Clinic News Network for more updates.

Media contact: 

• Julie Ferris-Tillman, Ph.D., Mayo Clinic Communications, [email protected]

The post Advancing Pancreatic Cancer Prognosis with AI-Driven Spatial Analysis of Pathology Slides appeared first on Mayo Clinic News Network.