CAP launches a new podcast, CIPI CONNECTIONS.
###
 |
| CIPI Cup |
One of the several Councils of the College of American Pathologists (CAP) is CIPI - The College on Informatics and Pathology Innovation. Home page here. They have five committees:
- Artificial Intelligence
- Cancer
- Digital and Computational Pathology
- Informatics
- Pathology Electronic Reporting (PERT0
- Welcome to the debut episode of CIPI Connections, the official podcast of the College of American Pathologists' Council on Informatics and Pathology Innovation (CIPI). Co-hosts Dr. Giovanni Lujan and Dr. M.E. de Baca introduce the podcast’s mission and discuss Dr. de Baca’s inspiring journey from ophthalmology in Germany to leading innovation in pathology. Learn how CIPI and its five committees are shaping the future of digital pathology, AI, cancer reporting, and more—and what listeners can expect from future episodes.
CIPI Connections Podcast 1 — “Pathology Meets Artificial Intelligence”
This episode of CIPI Connections from the College of American Pathologists explores how existing laboratory validation frameworks can guide the safe implementation of artificial intelligence in pathology. Hosted by Dr. M.E. de Baca, the discussion features Drs. Matthew Hanna, Nick Spies, and Larissa Furtado from the CAP AI Committee, who examine parallels between AI model validation and familiar laboratory processes such as immunostain verification and molecular assay quality control. They emphasize that AI should be treated as another diagnostic test, requiring analytical and clinical validation, documentation, and continuous performance monitoring. The speakers identify gaps in current CAP checklists—particularly around data quality, explainability, and model updating—and highlight the emerging role of AI implementation specialists to ensure safe, effective deployment. Their consensus: existing laboratory rigor provides a strong foundation for AI adoption, but evolving guidance and education will be essential to maintain patient safety and trust as laboratories integrate machine learning tools into clinical workflows.
CIPI Connections Podcast 2 — “Horror Stories in Pathology Informatics: Unflagged and Overlooked”
In this inaugural episode of the Horror Stories in Pathology Informatics series, Drs. Alexis Carter and Omar Baba dissect a real-world case involving a missed low cortisol result that went unflagged in an electronic health record, delaying diagnosis of adrenal insufficiency. The case—stripped of identifiers but based on actual events—illustrates how informatics design decisions can compromise patient safety. Because cortisol has time-dependent reference ranges, the lab displayed both ranges in a free-text comment, preventing the system from automatically flagging the abnormal value. The physicians discuss how structured data fields, discrete order codes for morning and afternoon tests, or “ask-on-order-entry” prompts could have prevented the oversight. They also caution against overreliance on visual flags as shortcuts for reviewing lab results, emphasizing the need for clinician education, structured data standards, and interdepartmental communication. The episode concludes with practical takeaways: build structured informatics solutions wherever possible, ensure reference ranges are machine-readable, and reinforce the clinician’s responsibility to review all results—not just those flagged as abnormal.