Discussion
Age, histological grade, myometrial invasion (MI), lymph-vascular space invasion (LVSI), tumor size (TS) have been proved to have value on prediction of lymph node metastasis (LNM) which might help physicians to make better clinical decisions 12-15. How can we achieve the goal of maximum resection of lesions and cause minimum injury if it is not accessible to use SLN. We believe the pathologic grade should be the most reliable and accessible factor. That’s the reason grade 3 was chosen as the indicator to evaluate the possibility of LNM in this study.
Inconsistent standard were published by different institutions: Mayo and FIGO(The International Federation of Gynecology and Obstetrics) guidelines tend to categorize TS > 2cm as high-risk, while for NCCN (The National Comprehensive Cancer Network) and ESMO (Europe’s leading medical oncology society) guidelines, TS was not considered as a high-risk factor5,16,17. In fact, it is uneasy to determinate the size of tumor due to its irregular shape of lesion, and the latest NCCN guideline had deleted this factor16.
It is true that deep myometrial invasion increased the risk of lymph nodes metastasis based on suggestions of almost every international or national guideline and our data (not shown in this study). However, it is not easy to accurately evaluate the deep myometrial invasion before operation. Although the NCCN guidelines recommend the use of pelvic enhanced MRI to determine the depth of myometrial invasion (myoinvasion >50% is considered as a risk factor)16, the accuracy of MRI is about 68% for T2-weithgted imaging18.
For age, it was considered older age was associated with higher chance of LNM, while the age cut-off was still inconsistent. According to the latest NCCN guidelines, “age >= 60” was considered as an age cut-off 16, while in our data >=70-year-old seemed to have significant impact on LNM.
For LVSI, never to be ignored in EC carcinogenesis, which has been emphasized again recently. LVSI is a potential predictor for EC recurrence14,19 and associated with a significant higher rate of paraaortic LNM with OR at five 20. A retrospective study showed no overall survival difference between adjuvant external beam radiation therapy versus vaginal brachytherapy in LVSI-positive patients, while it was associated with increased risk of death (hazard ratio of 1.94) 21. Unfortunately, it is quite difficult to get before operation.
Finally, we selected G3 as our first concern in making decisions before operations of EC patients. We found that EEC G3 patients without any extra risk factor (LVSI, AI or CSI) had an LNM rate of 16.4%, which is remarkably lower than other risk factors, for which LND might not be needed. On the contrary, EEC G3 patients with one or more extra risk factors mentioned above had higher chances of lymph node metastasis, which could be explained by synergetic effects in between these risk factors. Grade 3 EEC patients with only one risk factor presented with an LNM rate ranged at 14.5% to 50% while if with two extra risk factors, LNM rate increased to 60% and 75%. Although grade 3 EEC patients with positive LVSI, AI and CSI, was found to have lower LMN incidence at 33.3%, also known as 1 out of 3. We assume it was due to the limited sample size.
The categorization of three tiers of LVSI was mentioned by pathologists to evaluate its potential risk19. Although our study failed to categorize LVSI into three tiers, our findings are consistent with these opinions. LVSI were independent risk factors of both grade 3 and grade 1&2 subgroups, increased the LNM rates more than 3 folds and 8 folds, respectively. Therefore, we recommend EEC G3 patients with LVSI should undertake lymph node evaluation, EEC G3 patients with no other risk factor could be evaluated only with imaging, while EEC G3 patients with positive LVSI in their final pathological diagnosis, should be evaluated by both imaging and comprehensive staging surgery if not done previously, which is consistent with the latest NCCN guidelines16.
Two randomized trials indicated that systematic pelvic LND had no survival benefit compared with no LND in early EC patients8,9. Our data showed no survival difference between grade 3 and grade 1&2 patients, which indicated that “grade” might not influence patients’ prognosis although grade 3 patients had higher rate of LNM than grade 1&2 patients. This could be due to the small sample size of grade 3 patients. Similarly, patients with pelvic and para-aortic lymph node dissection done had similar OS and PFS than patients with only pelvic lymph node dissection done in grade 3 population at early stage, suggesting a less traumatic treatment could be considered in these patients.
In addition, we found that MELF and MI only increased the LNM rate of grade 1&2 population, while CSI increased the LNM risk of grade 3 patients. This might indicate different metastasis pathways in EC patients with different grades, for which further investigations should be carried out.
In this study, our strength stood on the follow-up duration of almost ten years. In addition, the independent high-risk factors were robust with adjustments. We also provided clinically friendly data by stratifying these risk factors, providing certain evidence for clinical practice especially in areas with poor medical resource. However, this study still had some limitations. Since this is a retrospective analysis, selection bias may have influenced the treatments given to these grade 3 patients. Only 381 (30.9%) patients were found to have complete follow-up data due to technical reasons. Larger sample size and prospective clinical trial design with novel molecular classification warrant more consideration.