Computed tomography reduces treatment uncertainty without improving classification reliability in distal radius fractures

Abstract Purpose To determine whether adding computed tomography (CT) to plain radiographs reduces uncertainty in surgical decision-making for distal radius fractures and improves the reliability of four classification systems. Methods Fifty acute displaced distal radius fractures were assessed by four orthopaedic surgeons, first with radiographs alone and, after a four-week washout, with radiographs and CT. At each session, observers classified fractures using the AO/OTA, Fernandez, Frykman and Universal systems and recorded a treatment decision (operative, conservative or undecided); for undecided cases, they selected the factor most likely to resolve their uncertainty. Agreement was assessed with kappa statistics, and changes between sessions with clustered McNemar tests. Results Undecided responses decreased from 29.0% with radiographs alone to 3.5% with CT, and operative decisions increased from 55.5% to 73.0% (both p < 0.001). Interobserver agreement on treatment remained fair (Fleiss' κ, 0.303 vs. 0.294). Interobserver reliability of classification was slight to moderate on radiographs (κ, 0.184–0.461) and did not improve with CT (κ, 0.169–0.386; differences, − 0.075 to + 0.021), although CT changed the assigned category in 23.0–51.0% of assessments and reclassified one third of apparently extra-articular fractures as articular. Radiological parameters were the leading cause of indecision with radiographs alone (65.5%) but were not cited for any of the seven assessments that remained undecided after CT. Conclusion CT did not make fracture classification more reproducible or bring surgeons to the same treatment choice, but it substantially reduced individual decisional uncertainty. Once imaging uncertainty was resolved, residual uncertainty related to patient factors rather than radiological findings.