Diagnostic accuracy of intelligent quick MR reconstruction for grading posterior cruciate ligament injuries

Vollständiger Abstract

Worum geht es in dieser Arbeit?

This retrospective diagnostic accuracy study aimed to evaluate the image quality and diagnostic accuracy of intelligent quick magnetic resonance(IQMR) reconstruction for grading posterior cruciate ligament (PCL) injuries, using arthroscopy as the reference standard. A total of 110 patients with arthroscopically confirmed PCL injuries were included, and their preoperative magnetic resonance images were processed with both standard and IQMR reconstructions. Two independent radiologists assessed subjective image quality using a 4-point Likert scale and objective metrics including signal-to-noise ratio and contrast-to-noise ratio. Diagnostic performance for detecting complete (grade 3) and partial (grade 1 and 2) tears was analyzed using receiver operating characteristic curves, and interobserver agreement was evaluated with weighted kappa statistics. Compared with standard reconstruction, IQMR reconstruction yielded significantly superior subjective and objective image quality (all P < 0.005) and excellent interobserver agreement (weighted kappa = 0.63–0.97). Diagnostic correlation with arthroscopy showed robust performance for complete tears (area under the curve = 0.79, P < 0.05) but limited efficacy for partial tears (area under the curve = 0.54 and 0.17 for grade 1 and 2, respectively), with overall moderate agreement between magnetic resonance imaging and arthroscopic grading (kappa = 0.57). In conclusion, IQMR reconstruction significantly improves image quality for PCL assessment, leading to excellent diagnostic reliability for complete tears, while its performance for partial tears remains limited. Further prospective multi-center studies are warranted.