In cervical cancer, the diagnostic value of SLN technique has been verified by previous studies, with a meta-analyzed sensitivity of 91.4% in early-stage patients.15 In the latest FIGO staging system for cervical cancer, patients with lymphatic metastasis are classified as stage IIIC and should be treated with concurrent chemoradiotherapy.16 In many institutions, gynecologists are accustomed to replacing lymphadenectomy with SLN biopsy and waiting for final pathology only. This policy was based on the hypothesis that chemoradiotherapy alone is equally efficient as lymphadenectomy combined with chemoradiotherapy when nodal metastasis occurs in early-stage patients. However, this hypothesis has not been verified in a randomized controlled trial.
An obvious concern is, if the patients have chemoradiotherapy resistance, the undetected metastatic nodes may survive and become the sources of recurrence. On the other hand, the extent of metastases is usually beyond the level that positive SLNs locate, so it will be difficult to formulate precise radiation coverage without the information from lymphadenectomy. In our study, the risk of having undetected metastasis after positive SLN biopsy was 63.6%, and it was 66.7% if positive SLNs were found on frozen section examination. Therefore, additional lymphadenectomy based on frozen section examination of SLN might be necessary.
In our study, frozen section examination successfully detected all macrometastases, however, it missed all micrometastases and isolated tumor cells. This result accorded with the findings from previous studies. However, there is no consensus on the clinical significance of SLN micrometastases and whether this indicates the necessity of pelvic lymphadenectomy remains unanswered. In a histological study by Barranger et al.,17 non-SLNs were also examined by ultrastaging techniques including serial sectioning and immunohistochemistry, however, none of the 106 non-SLNs was found to be metastatic. A similar finding was confirmed in the study by Okamoto et al., suggesting that the non-SLNs were seldom involved if the SLNs harbor merely micrometastases.18 In our cohort, none of the non-SLN metastases occurred on the same lateral of SLN micrometastasis or isolated tumor cells, and no pelvic nodal recurrence occurred during the follow-up period. Taken together, these evidences imply that the micrometastasis and isolated tumor cells in SLNs may represent the very beginning of lymphatic spread. For these patients, SLN biopsy is not just diagnostic but also therapeutic, thus lymphadenectomy can be omitted if SLNs had been dissected without macrometastasis.
If frozen section examination has enough accuracy to detect SLN macrometastases, a selective lymphadenectomy policy can be established based on SLN biopsy, which may provide an option to reduce the potential risks. In a prospective study containing 35 patients, the pelvic lymphadenectomy was omitted in patients with FS-negative SLNs and none of these patients experienced pelvic recurrence in a median follow-up period of 49 months.19 This study well supported our viewpoint and suggested an important role for frozen section examination in SLN biopsy. However, data on this issue is limited and this viewpoint needs further validation.
As an exploratory study, we included a series of patients that were conventionally recognized not as candidates for SLNB, such as patients with tumor larger than 2 cm or neoadjuvant chemotherapy. The bilateral detection rate was relatively low, which should be partly attributed to our single tracer method. However, we found a significant difference in bilateral detection rates between patients with and without lymphovascular invasion, while no difference was revealed in the comparisons by other factors. This finding suggested that the lymphovascular obstruction by tumorous embolus might be the real reason behind the failure of SLN detection. Besides, we found that the lymphatic drainage from cervix to SLNs was usually multidirectional as 37.7% patients simultaneously had multiple groups of SLNs detected in their hemipelvis. Accordingly, there exists a possibility that lymphovascular embolus blocked the drainage toward metastatic nodes, whereas the normal nodes were successfully detected. This selective “shielding effect” by lymphovascular invasion may lead to the omission of metastatic nodes and produce false-negative results. This may also explain why the rate of non-SLN metastasis was so high in the cases with positive SLN. Therefore, in cases with extensive lymphovascular invasion, the pathological results should be paid special attention.
Understanding the particular risk associated with lymphovascular invasion in SLN biopsy may help to optimize the treatment for these patients. Although it is difficult to identify them before radical hysterectomy, it is possible to establish a clinic-pathological model to predict lymphovascular invasion and guide the performance of SLN techniques. Furthermore, we recommend performing multipoint injections on normal cervical area to reduce the influence of lymphovascular invasion, and that patients whose cervix is completely occupied by tumor should not be considered for SLN technique.