Reliability of distal radius fracture classification systems: a CT based study.

Area of Science:

• Orthopedic surgery
• Radiology
• Medical imaging

Background:

• Distal radius fractures are common injuries.
• Accurate classification is crucial for treatment and prognosis.
• Existing classification systems (AO/OTA, Frykman, Fernandez) have varying reliability.

Purpose of the Study:

• To evaluate the reliability and reproducibility of three common classification systems for distal radius fractures using computed tomography (CT).
• To compare the performance of AO/OTA, Frykman, and Fernandez classifications in terms of interobserver and intraobserver agreement.

Main Methods:

• Four radiologists independently classified 115 distal radius fractures on CT scans using AO/OTA, Frykman, and Fernandez systems.
• Interobserver reliability was assessed using intraclass correlation coefficient (ICC) and kappa statistics.
• Intraobserver agreement was evaluated using Cohen's kappa after a six-week interval.

Main Results:

• Interobserver reliability was fair for AO/OTA (ICC 0.447) and Frykman (ICC 0.432), and poor for Fernandez (ICC 0.196).
• Intraobserver agreement was substantial for Frykman (Observer 1: 0.754), moderate for AO/OTA (Observer 1: 0.449) and Fernandez (0.333, 0.320).
• Reliability of AO/OTA decreased with subgroup classification; Fernandez showed the lowest overall reliability.

Conclusions:

• No current classification system demonstrates complete reproducibility for distal radius fractures on CT.
• The Frykman system exhibited superior intraobserver reproducibility.
• AO/OTA and Frykman systems offered better interobserver reliability compared to the Fernandez system.