Interobserver variability in the determination of upper lobe-predominant emphysema

Area of Science:

• Pulmonary Medicine
• Radiology
• Medical Imaging Analysis

Background:

• High-resolution CT (HRCT) scans guide lung volume reduction surgery decisions.
• Radiologist and pulmonologist agreement on HRCT interpretation for emphysema is unknown.

Purpose of the Study:

• To assess inter-specialty agreement between radiologists and pulmonologists in evaluating emphysema severity and upper lobe predominance on HRCT.
• To compare subjective assessments with objective density mask analysis.

Main Methods:

• Two radiologists and three pulmonologists reviewed HRCT scans of 30 emphysema patients.
• Readers assigned emphysema severity scores and assessed upper lobe predominance.
• Objective density mask analysis determined the percentage of emphysema at -910 HU.

Main Results:

• Inter-reader correlation for emphysema severity scores ranged from 0.59 to 0.87, with better agreement within specialties.
• Agreement on upper lobe predominance was poor (kappa 0.20–0.60), though radiologists and pulmonologists agreed on distribution in 75% of cases.
• Agreement was higher when upper lobe predominance was clearly indicated by densitometry.

Conclusions:

• Interobserver agreement for determining upper lobe-predominant emphysema on HRCT is generally poor.
• Objective densitometry can enhance agreement, particularly in cases with clear upper lobe predominance.