Main findings
We found that 18.7% of clinically FIGO stage I-II EOC patients are upstaged based on comprehensive surgical staging. The individual estimates from the included 23 studies varied widely, ranging from 4.5% up to 38.5%, highlighting the need for a pooled approach to increase accuracy. The result of this study enables clinicians to more accurately inform patients on the staging utility and clinical consequences of their planned surgery. This can be further stratified according to individual patient characteristics in FIGO stage, histology and tumor differentiation. The upstaging rate increases to 23.0% in those patients suspected of stage I EOC. The stage II EOC patients seem to reduce the overall risk, however, in this subgroup less tumor positive sites (e.g., uterus involvement) will lead to upstaging that in stage I. Notably, those with a high grade tumor (40.9%), and serous EOC (35.3%), yielded a high probability of upstaging. We have pooled the data of grade II and III tumors into high grade tumors to ensure translation to the clinical practice. In contrast to what we expected, and based on a limited number of cases, clear cell carcinoma of the ovary was not associated with a substantial upstaging risk (11.2%) relative to the overall risk of 18.7%. While certain low-risk groups exist, we could not identify any subgroup in which a staging surgery can be safely omitted based on an absent benefit.
This meta-analysis provides insight into the tumor positivity rate, and the upstaging contribution of each component encompassing surgical staging. The tumor positivity was identified in uterus, cytology, peritoneal biopsies, omentum, and the appendix in 6.2%, 18.4%, 9.7%, 5.2%, and 3.6% of EOC patients, respectively, with corresponding upstaging rates of 5.9%, 8.5%, 3.5%, 3.9%, and 1.6%. Remarkably, results on tumor involvement of the uterus, and the contralateral ovary are only scarcely reported in the included studies. This is a limitation across the retrieved studies, and particularly restricts the ability to preoperatively assess the understaging risks in young women where fertility-sparing surgery is considered.
Lymph node assessment resulted in 8.7% upstaging, whereas this increased to 12.0% when mucinous EOC cases were excluded. This was decided based on previous reports on the rare lymphatic metastatic dissemination of mucinous EOC, at approximately 0.8% found at staging38. When specified upon lymph node level, para-aortic lymph nodes are more often affected with 6.8% upstaging, than pelvic lymph nodes with 4.2%. The impression might be given that the combined upstaging risk for both positive pelvic and para-aortic lymph nodes is higher relative to the calculated overall risk of 8.7%. However, a subgroup of patients (2.3%) had metastases in both their pelvic and para-aortic lymph nodes. Furthermore, the included studies differed between the data analyses, possibly contributing to a variation in results.
Unfortunately, the terms lymph node sampling, dissection, and lymphadenectomy were used interchangeably within the included literature. A metaregression was used to study whether an association between the number of resected lymph nodes – regardless of the name used for the type of structured nodal assessment – and the upstaging rate. This did not yield a significant association, though this can possibly be attributed to insufficient statistical power as it was based on ten included studies. When multivariately adjusted for the study period of each individual study, assuming that recent studies would routinely use a CT based nodal assessment preoperatively, the results remained not significant.