In this meta-analysis of patients with FIGO stage III endometrial cancer who underwent radical surgery, we compared cancer recurrence rates and survival between those who received AC or ACR after surgery. The two adjuvant treatments were associated with similar distant recurrence, while ACR was associated with significantly lower risk of total recurrence and local recurrence. These outcomes were associated with low heterogeneity. ACR was also associated with significantly higher OS than AC, although this outcome was associated with high heterogeneity and we were unable to identify its sources.
The results from our meta-analysis are consistent with several previous studies. For example, analysis of data in the US National Cancer Data Base suggests better survival with ACR than AC [18, 28, 42].
We found evidence of better survival with ACR than AC specifically in patients with endometrial cancer in stage IIIA or IIIC. These results suggest that ACR is more effective at preventing cancer recurrence and improving OS in the presence of pelvic lymph node involvement (substage IIIC1) or para-aortic lymph node metastasis (substage IIIC2).
We found evidence of better survival with ACR than AC specifically in patients with either endometroid or non-endometroid carcinoma. This is an important finding because the non-endometroid type is associated with worse prognosis. In this sense, our finding is consistent with reports that ACR provides overall survival benefit over AC for serous carcinoma, which is also associated with worse prognosis [43, 44]. This literature and our meta-analysis suggest that ACR may lead to better prognosis in patients with severe or advanced forms of endometrial cancer.
While our findings may differ from those in the large RCT GOG-258, some work suggests that this trial may underestimate the ability of ACR to reduce recurrence rate because the radiation dose in the combination regime may have been insufficient [12, 18]. Consistent with our meta-analysis, the multi-center PORTEC-3 trial showed that ACR led to better 5-year rates of DFS and OS than adjuvant radiotherapy alone, especially among patients with high-risk histological types of endometrial cancer[45].
Our meta-analysis presents several limitations. First, pooled OS data showed high heterogeneity, the sources of which we were unable to identify. Nevertheless, our OS meta-analysis seems reliable because we obtained similar results in subgroup analyses, which involved fewer confounding factors. Second, we were unable to take into account factors that might affect survival, such as residual tumor volume after surgery, differences in chemo- or radiotherapy regimens, and lymph node dissection. Few studies reported data on such factors. Third, all but one study in our meta-analysis were retrospective studies, which may increase the risk of selection bias. In addition, the one RCT in our meta-analysis contained 22 patients in stages I, II or IV, whom we could not exclude. Fourth, the radiotherapy in the ACR regimen varied across the included studies (e.g. external pelvic radiotherapy, vaginal brachytherapy). This may have contributed to the observed heterogeneity. Fifth, we had to extract HRs from Kaplan-Meier plots in nearly half the studies[17, 26, 27, 30, 35, 38, 39, 41], which may have introduced error. Finally, we found evidence of publication bias toward greater clinical benefits of ACR than AC among the studies in our meta-analysis.
Despite these limitations, our meta-analysis provides strong evidence that ACR is superior to AC for preventing recurrence and improving survival of patients with stage III endometrial cancer. To the best of our best knowledge, this is the first meta-analysis to compare the two adjuvant approaches after primary radical surgery. Our findings suggest recommending ACR to patients with stage III disease, especially those at greater risk of recurrence.