BackgroundThe yield of whole‐body MRI for preventive health screening is currently not completely clear.PurposeTo systematically review the prevalence of whole‐body MRI findings in asymptomatic subjects.Study TypeSystematic review and meta‐analysis.SubjectsMEDLINE and Embase were searched for original studies reporting whole‐body MRI findings in asymptomatic adults without known disease, syndrome, or genetic mutation. Twelve studies, comprising 5373 asymptomatic subjects, were included.Field Strength/Sequence1.5T or 3.0T, whole‐body MRI.AssessmentThe whole‐body MRI literature findings were extracted and reviewed by two radiologists in consensus for designation as either critical or indeterminate incidental finding.Statistical TestsData were pooled using a random effects model on the assumption that most subjects had ≤1 critical or indeterminate incidental finding. Heterogeneity was assessed by the I
2 statistic.ResultsPooled prevalences of critical and indeterminate incidental findings together and separately were 32.1% (95% confidence interval [CI]: 18.3%, 50.1%), 13.4% (95% CI: 9.0%, 19.5%), and 13.9% (95% CI: 5.4%, 31.3%), respectively. There was substantial between‐study heterogeneity (I
2 = 95.6–99.1). Pooled prevalence of critical and indeterminate incidental findings together was significantly higher in studies that included (cardio)vascular and/or colon MRI compared with studies that did not (49.7% [95% CI, 26.7%, 72.9%] vs. 23.0% [95% CI, 5.5%, 60.3%], P < 0.001). Pooled proportion of reported verified critical and indeterminate incidental findings was 12.6% (95% CI: 3.2%, 38.8%). Six studies reported false‐positive findings, yielding a pooled proportion of 16.0% (95% CI: 1.9%, 65.8%). None of the included studies reported long‐term (>5‐year) verification of negative findings. Only one study reported false‐negative findings, with a proportion of 2.0%.Data ConclusionPrevalence of critical and indeterminate incidental whole‐body MRI findings in asymptomatic subjects is overall substantial and with variability dependent to some degree on the protocol. Verification data are lacking. The proportion of false‐positive findings appears to be substantial.
Level of Evidence: 4
Technical Efficacy: Stage 3J. Magn. Reson. Imaging 2019;50:1489–1503.