Diagnostic performance of MRI in detecting residual soft tissue sarcoma after unplanned excision: Systematic review and meta-analysis

Objective: To investigate, in a systematic review and meta-analysis, the diagnostic performance of MRI in detecting residual disease after unplanned excision of soft tissue sarcoma.

Methods: Medline and Embase were searched for original studies on the diagnostic performance of MRI detecting residual soft tissue sarcoma after unplanned excision. Study quality was assessed using QUADAS-2. Sensitivity and specificity were pooled using a bivariate random-effects model. A Chi-squared test was used to test for heterogeneity.

Results: Ten studies were included, comprising a total of 469 patients. Median frequency of residual soft tissue sarcoma was 54.6% (range 31.6–73.1%). There was high risk of bias with respect to flow and timing in one study. There were applicability concerns with respect to patient selection in four studies. Nine studies reported to use the presence of a mass as a diagnostic criterion for residual soft tissue sarcoma. Sensitivities of included studies ranged between 36.4% and 86.7%, and specificities ranged between 77.8% and 100%. Pooled sensitivity was 65.9% (95% confidence interval [CI]: 55.5–74.9%) and pooled specificity was 85.1% (95% CI: 79.1–89.6%). The area under the summary receiver operating characteristic curve was 0.852. The included studies were statistically heterogeneous in their estimates of sensitivity (P = 0.016) and statistically homogeneous in their estimates of specificity (P = 0.793).

Conclusion: The presence of a mass is the most commonly reported diagnostic criterion to diagnose residual soft tissue sarcoma after unplanned resection. MRI achieves moderate sensitivity and fairly high specificity. Pooled estimate of sensitivity was subject to heterogeneity, which needs further exploration.