Summary

Metal Artifact Reduction Sequence Magnetic Resonance Imaging (MARS-MRI) study demonstrates a high accuracy and reliability for the making the diagnosis of PSI

Abstract

Background

The diagnosis of peri-prosthetic shoulder infection (PSI) in patients with painful arthroplasty is challenging. The use of Metal Artifact Reduction Sequence Magnetic Resonance Imaging (MARS-MRI) in diagnosing PSI has been reported but never studied. The goal of this study was to determine its accuracy for making the diagnosis of PSI.

Methods

Patients suspected to have PSI were prospectively recruited from one institution between 2015 and 2019. The inclusion criteria were a minimum of one year follow up and the availability of blood studies, radiographs and MARS MRI. Patients were categorized according to the International Consensus Meeting (ICM) 2018 criteria for PSI. Patients were considered “not infected” if they did not require surgery at last follow up (47/89, 53%)53 or if they did not meet criteria for infection using the criteria of the ICM. The diagnostic accuracy of MARS MRI findings for PSI was quantified as sensitivity, specificity, and accuracy using receiver operator curve (ROC) analysis with area under the curve (AUC).

Results

Of 130 patients who underwent MARS MRI in the study period, 89 (68%) patients met inclusion criteria. The MARS-MARI findings of axillary lymphadenopathy (AUC=0.94, Sn=95%, Sp=92%, OR=4) and edematous synovitis (AUC=0.94, Sn=91%, Sp=97%, OR=10) were highly accurate with very high specificity and sensitivity for the diagnosis of PSI (). Complex joint effusion (AUC=0.86, Sn=86%, Sp=86%, OR=5.7) was found to be accurate with high specificity and sensitivity. Rotator cuff muscle edema (AUC=0.75,OR=3.6) and extraarticular fluid collection (AUC=0.71, OR=5.0) were both moderately accurate and while these criteria had very high specificity (> 90%) they had low sensitivity for diagnosing PSI. While the finding of a sinus tract finding in MARS MRI had a very high specificity (99%) with odds ratio of 9.3, the accuracy (AUC=0.63) and sensitivity (28%) were low. Periprosthetic edema and periprosthetic resorption or osteolysis had low accuracy, sensitivity, and specificity for diagnosis of PSI

Conclusion

This study demonstrates a high accuracy and reliability of MARS MRI for the making the diagnosis of PSI. This diagnostic test should be considered used when evaluating patients for PSI. These conclusions are based upon a specific MRI protocol with experienced musculoskeletal radiologists.