A meta-analysis of all studies (from 1998 to the present) on the effect of SL/LUS in patients with potentially resectable CRLM with respect to alteration in surgical management was performed.
Twelve studies satisfied the inclusion criteria. A total of 1,047 patients underwent SL/LUS. The true yield of SL/LUS for CRLM was 19%(95%confidence interval [CI], 16%-22%), with a diagnostic odds ratio of 132 (95%CI, 56-310) and an overall sensitivity of 59%(95%CI, 53%-65%). Subgroup analysis for detection of other liver and peritoneal lesions showed a sensitivity of 59%(95%CI, 49%-67%) and 75%(95%CI, 63%-85%) respectively. There was major between-study heterogeneity for all analyses, with no obvious cause revealed by meta-regression.
The true benefit of using SL/LUS universally seems limited. It appears more useful as an adjunct in patients when peritoneal disease is suspected.