From Image to Diagnosis: Characterizing Sources of Error in Histopathologic Interpretation

Area of Science:

• Pathology
• Medical Education
• Digital Health

Background:

• Accurate histopathologic diagnosis is vital for patient care, research, and public health.
• Diagnostic errors in pathology can have significant clinical implications.

Purpose of the Study:

• To identify sources of diagnostic error in digitized breast biopsy interpretation.
• To analyze interpretive and descriptive errors made by pathology residents and attending physicians.

Main Methods:

• A mixed-methods approach was used with 90 participants (residents and attending pathologists).
• Participants interpreted digitized whole-slide images of breast biopsies, with their review behavior tracked via eye-tracking and case viewer software.
• Diagnostic accuracy was assessed against a consensus diagnosis, analyzing errors in 4 interpretive phases.

Main Results:

• Pathologists consistently detected the critical region of interest (cROI) (∼94% of the time).
• Pathology trainees were less likely than attending pathologists to consider the cROI relevant and more likely to use incorrect terminology to describe features (41% vs. 21%).
• Failure to accurately describe histopathologic features was strongly associated with incorrect diagnoses.

Conclusions:

• Diagnostic errors in breast biopsy interpretation often stem from descriptive inaccuracies rather than detection failures.
• Improving trainees' ability to recognize and verbally describe histopathologic features is critical for diagnostic competency.