Following unenhanced MRI assessment for local recurrence after surgical resection of mesenchymal soft tissue tumors, do additional gadolinium-enhanced images change reader confidence or diagnosis?

Abstract

Purpose: Evaluate if gadolinium enhanced MR imaging (GeMRI) improves confidence, changes the final diagnosis, or improves accuracy in the assessment of musculoskeletal (MSK) tumor residual or recurrence following surgical resection. We also assess if different experience levels change the above results.

Methods and materials: Initially, pre-contrast images were independently reviewed by two radiologists, one with 25 years of experience (R1) and one undergoing MSK specialty training (R2). Two questions were answered: (1) Mass present? and (2) Likelihood of malignancy? Subsequently, both pre-contrast and post-contrast images were independently reviewed. The same questions were again answered plus four others including if GeMRI changed mass characterization, better defined cystic versus solid, better defined tumor extent, or improved conspicuity. Lastly, the readers answered whether GeMRI changed confidence, and changed their final diagnosis. Histologic diagnoses were available in 43 cases, with the remaining 44 cases based upon clinical and/or imaging follow-up.

Results: GeMRI definitely improved confidence in 8/7 cases, and slightly improved confidence in 20/29 cases and changed the final diagnosis in 11/8 cases for R1 and R2 respectively. Positive and negative predictive values statistically improved for R2 (positive predictive value 36.4% versus 50%, p=0.02; negative predictive value 75.4% versus 79.1%, p=0.04) but not for R1. Reader concordance for malignancy improved with GeMRI (κ=0.44 pre-contrast and κ=0.71 post-contrast).

Conclusion: GeMRI improved reader confidence, improved reader concordance and modestly improved accuracy for the less experienced reader. Where possible, GeMRI should be used in the assessment of MSK tumor residual or recurrence.