Is pancreatogram interpretation reliable?--a study of observer variation and error

Area of Science:

• Gastroenterology
• Radiology
• Oncology
• Hepatology

Background:

• Endoscopic pancreatograms are crucial for diagnosing pancreatic and biliary duct diseases.
• Radiographic interpretation can be subjective, leading to observer variation.
• Clinical details may influence diagnostic accuracy in endoscopic retrograde cholangiopancreatography (ERCP).

Purpose of the Study:

• To evaluate observer variation in interpreting endoscopic pancreatograms.
• To assess the impact of clinical information on diagnostic accuracy and consistency.
• To determine the reliability of ERCP findings in diagnosing pancreatic cancer and pancreatitis.

Main Methods:

• Four experienced observers interpreted 40 sets of endoscopic pancreatograms (20 cancer, 20 pancreatitis) on three separate occasions.
• Interpretations were performed both "blind" (without clinical details) and with clinical information.
• Observer consistency and diagnostic accuracy were analyzed.

Main Results:

• Individual observer consistency for "blind" diagnoses ranged from 61% to 78%.
• Clinical information significantly improved diagnostic consistency and accuracy for both cancer (52%-83%) and pancreatitis (87%-95%).
• Unanimous correct diagnoses were achieved in only 53% of cases, even with clinical details; 43% of interpretations changed with clinical data, improving accuracy in 83% of these instances.

Conclusions:

• While clinical information enhances diagnostic accuracy of endoscopic pancreatograms, observer variation persists.
• ERCP provides valuable direct information, but caution is advised when relying solely on radiological appearances.
• Multidisciplinary discussion is essential for accurate diagnosis and patient management.