Physician interpretation of ultrasound in the evaluation of ankle edema

The aim of this study was to assess inter‑ and intra‑observer agreement of detection and grading of oedema measured by vascular physicians using ultrasonography.

In this observational study, three investigators read 113 videos of ultrasonography of the medial lower leg of 77 patients using a 15-18 MHz linear probe and reported whether oedema was present or not and, if present, to what extent. The videos were selected by the coordinator and sent every two months for a period of six months to the three investigators. Intra‑ and interobserver agreements were calculated using the Fleiss Kappa coefficient, reported with their 95% confidence interval and interpreted using the Landis and Koch values.

Inter-observer agreement as to whether oedema was present or not was 0.88 (0.77-0.98), 0.96 (0.89-1.0) and 0.91 (0.80-1.0) for the first, second and third readings, respectively. The concordance was considered to be "excellent." Inter-observer agreement as to the severity of oedema was 0.52 (0.38-0.65), 0.53 (0.39-0.66) and 0.61 (0.47-0.75) for the first, second and third readings, respectively. The concordance was "moderate". Intra-observer reliability for the diagnosis of oedema was only 0.89 (0.70-1.0), 0.93 (0.75-1.0) and 0.92 (0.74-1.0) for the first, second and third reader, respectively, which was "excellent".

The inter-observer agreement interpreting ultrasound videos using standard probes was excellent. Only moderate agreement in grading the severity of the edema using our arbitrary criteria was shown.