Validity of intraoperative gross examination of myometrial invasion in patients with endometrial cancer: a meta-analysis

Objective. The current recommended endometrial cancer surgical treatment is abdominal extrafascial total hysterectomy with bilateral salpingo-oophorectomy followed by pelvic lymphadenectomy if >50% myometrial invasion is estimated by intraoperative gross examination (IGE). This meta-analysis aims to quantify evidence regarding the validity/predictive value of IGE staging compared with final histology. Design. Meta-analysis of studies published until October 2011. Setting. Systematic search, according to PRISMA guidelines, of the six major medical literature databases – Medline, Scopus, EMBASE, Google Scholar, Ovid, Cochrane. Population. Sixteen eligible studies including 2567 endometrial cancer patients. Methods. Pooled sensitivity/specificity, accuracy, negative/positive predictive value (NPV/PPV) and diagnostic odds ratio (DOR) of IGE were calculated and the summary receiver operator characteristic (sROC) curve was constructed. A meta-regression analysis was used to explore the role of potential modifiers of sensitivity and specificity. Main outcome measures. Pooled diagnostic measures of IGE indices. Results. Sixteen studies (15 retrospective, one prospective) meeting the inclusion criteria were qualitatively analyzed. Pooled IGE estimates were: sensitivity = 0.75 (95%CI: 0.72–0.78), specificity = 0.92 (95%CI 0.90–0.94), accuracy = 0.87 (95%CI 0.86–0.88), NPV = 0.89 (95%CI 0.87–0.92), PPV = 0.80 (95%CI 0.76–0.84) and DOR = 36.9 (95%CI 28.7–47.4). No significant modifiers were identified for sensitivity or specificity. Conclusions. The synthesized measures presented here for the first time showed that accuracy, sensitivity and specificity of IGE were 87, 75 and 92%, respectively, which indicates that IGE is useful for estimating depth of myometrial invasion and staging of endometrial cancer in clinical practice. The degree to which the relatively low values of some of its performance indicators could be improved remains to be elucidated in order for the values to be comparable with those from frozen section biopsies.