Accuracy and precision of two computer-assisted methods of radiographic wear measurement in total hip arthroplasty.

Abstract

Roentgen Monographic Analysis Tool (ROMAN) and Hip Analysis Suite (HAS) were used to analyze radiographs of a phantom hip model. Displacements of known magnitude and direction were produced using dial micrometers. Differences between the known displacement and the programs' reported displacement were compared. Hip Analysis Suite was superior with a median error of 0.075 mm (range, 0.019-0.205 mm) compared with 0.137 mm (range, 0.008-0.389 mm) for ROMAN (P = .002). Hip Analysis Suite was also more precise when evaluating intraobserver variability, with a standard deviation between radiographs of 0.007 mm (range, 0.002-0.009 mm), whereas ROMAN's standard deviation was 0.117 mm (range, 0.007-0.153 mm). Repeatability for HAS was 0.019 mm and 0.325 mm for ROMAN. Hip Analysis Suite was more accurate and precise than ROMAN under experimental conditions with digital radiographs.