Usefulness of MRI-Based Local Surveillance After Surgical Treatment of Musculoskeletal Soft-Tissue Sarcomas: A Systematic Review and Meta-Analysis

BACKGROUND. The value of routine MRI follow-up after surgical treatment of musculoskeletal soft-tissue sarcoma (STS) is controversial. OBJECTIVE. The purpose of this study was to evaluate the usefulness of MRI-based surveillance for musculoskeletal STS represented by the proportion of local recurrences (LRs) discovered by MRI versus clinically, stratified by imaging surveillance intensity; the characteristics of LRs detected on imaging versus clinically; and the impact of imaging surveillance on survival. EVIDENCE ACQUISITION. Multiple electronic databases were searched systematically for articles published through November 28, 2022, about controlled trials and cohort studies on the usefulness of MRI-based surveillance for musculoskeletal STS. The risk of bias was assessed using an adapted Newcastle-Ottawa scale. Random-effects meta-analyses of the proportion of LRs discovered by MRI as opposed to clinically were conducted. The association of low-versus high-intensity surveillance with the proportion of LR detected on MRI was assessed with a chi-square test of subgroup differences; for this latter assessment, high intensity was defined as at least one local surveillance imaging examination for low-risk tumors and at least three imaging examinations for highrisk tumors during the first 2 posttreatment years. EVIDENCE SYNTHESIS. A total of 4821 titles and abstracts were identified, and 19 studies were included. All studies were retrospective cohorts. There was substantial variability in follow-up approaches. The risk of bias was moderate in 32% and high in 68% of studies. The pooled proportion of LRs detected on MRI was 53% (95% CI, 36-71%) with high-intensity surveillance and 6% (95% CI, 3-9%) with low-intensity surveillance (p <.01). Comparison of LR characteristics (LR size, depth, grade, location, resection margins) detected on imaging versus clinically identified inconsistent results between studies. Trends toward better survival for imaging-detected LRs or more frequent imaging use were noted in four studies. CONCLUSION. When used at a high intensity, MRI-based surveillance can detect many clinically occult LRs, although the studies are small, occasionally yielded conflicting results, and are often of poor quality. A survival benefit could be associated with imaging use, but further research is needed to evaluate the causality of any observed survival differences. CLINICAL IMPACT. MRI-based surveillance after surgical treatment of musculoskeletal STS is useful to detect clinically occult LRs and could improve patient outcomes.