Inter- and Intraobserver Agreement in Measuring Urolithiasis Density on Nonenhanced Computed Tomography

Area of Science:

• Radiology
• Urology
• Medical Imaging

Background:

• Treatment decisions for urolithiasis rely partly on kidney stone density measured in Hounsfield units (HU) via nonenhanced computed tomography (NECT).
• Variability among observers in HU measurements can impact treatment choices.
• This study investigates observer agreement in HU measurements and the effect of a standardized protocol.

Purpose of the Study:

• To assess inter- and intraobserver agreement for HU measurements in NECT of urolithiasis.
• To determine if a specific measurement protocol improves agreement.
• To identify the optimal protocol for consistent HU density assessment.

Main Methods:

• Retrospective analysis of 155 NECT scans from patients with kidney stones (≥4 mm).
• Five observers (radiologists, urologists, and residents) performed measurements four times each.
• Measurements were taken without a protocol and with two different protocols (A and B) involving bone setting, zoom, and specific plane selection.

Main Results:

• Initial interobserver agreement without a protocol was good (ICC=0.84).
• Protocol A decreased agreement (ICC=0.62), while Protocol B significantly improved it (ICC=0.90).
• Protocol B also enhanced intraobserver agreement (ICC=0.93) and showed the largest improvement for urologists.

Conclusions:

• Good observer agreement for HU measurements in urolithiasis exists without a protocol.
• A protocol using zoom, bone setting, and measurement in a representative plane (Protocol B) is recommended for improved consistency.
• Measuring in a single representative plane is effective, and measuring in three planes does not further increase agreement.