Quantitative and qualitative assessment of a coding system for reporting CT colonography

Area of Science:

• Radiology
• Gastroenterology
• Medical Coding

Background:

• Computed tomography colonography (CTC) is a key tool for colorectal cancer (CRC) screening.
• Standardized coding systems are crucial for accurate and reproducible interpretation of CTC findings.
• Assessing interobserver variability is essential for validating such coding systems.

Purpose of the Study:

• To validate a coding system for summarizing CTC findings in suspected CRC detection.
• To evaluate interobserver agreement and identify weaknesses in the coding system through qualitative analysis.

Main Methods:

• Retrospective analysis of 1,321 CTC procedures performed between July and December 2016.
• Dual reading of studies by an advanced practitioner radiographer and a consultant gastrointestinal radiologist.
• Quantitative assessment of interobserver agreement using the kappa statistic, followed by qualitative analysis of disagreements.

Main Results:

• High agreement (90% percentage agreement, kappa=0.69) for intra-colonic findings.
• Fair agreement (47% percentage agreement, kappa=0.22) for extra-colonic findings.
• Identified reasons for disagreement including perceived importance of incidental findings and skill set differences.

Conclusions:

• The coding system demonstrates substantial agreement for intra-colonic findings, indicating reproducibility.
• Refinements are recommended for the coding system to improve classification of extra-colonic findings.
• The study provides insights for improving the coding system's accuracy and consistency.