Coronal displacement in proximal humeral fractures: correlation between shoulder radiographic and computed tomography scan measurements.

Abstract

The purpose of this study was to analyze if there is a correlation in displacement measurements in coronal projections between shoulder radiographs and computed tomography (CT) scans in patients with proximal humeral fractures (PHFs). A comparative, retrospective analysis of images from our database was performed. Ninety-seven cases with PHFs with radiographs and CT scans were included. Four evaluators measured PHF displacement using the following: metaphyseal extension (ME), medial calcar (MC) medialization, apex-tuberosity distance (ATD), and cervicodiaphyseal (CD) angle. Measurement reliability was evaluated with a pilot sample by performing intraclass coefficient correlation analysis. Surgery indication according to displacement (CD angle <105° or >155° and ATD <3 mm) and agreement analysis were assessed by κ tests. All evaluated parameters presented correlations among methods for intrarater and inter-rater reliability. All measurements showed significant differences (ME of 1.2 ± 6 mm, P = .034; MC of 1 ± 5 mm, P = .041; ATD of 2.6 ± 5 mm, P = .001; and CD angle of 9° ± 16°, P = .001). Regarding indications for treatment type relating to ATD and CD angle, there was agreement between CT scans and radiographs in 66 of 97 cases (κ = 0.351, P < .001). Twelve cases with surgical indications by CT scans had conservative indications by radiographs, whereas in 19 cases with conservative indications by CT scans, radiographic measurements suggested surgical treatment. This study documented regular concordance between radiographs and CT scans for coronal displacement measurements in PHFs. Statistical differences were documented for all measurements. Webelieve that ATD and CD angle differences are clinically relevant (mean, 3 mm and 9°, respectively) because these differences might change the type of treatment.