Main findings and Interpretation
Patients with the 2009 FIGO stage of IB1-IIA2 CC will undergo pLN resection and impair the patient’s quality of life [8,40].Begin with this study, our TCGA data originated from bioinformatics analysis demonstrated that LN metastasis of CESC may be related to immune infiltration. Results stressed that the major immune infiltration cell types were natural killer (NK) cells, macrophages and T cells.
In addition, traditional procedures have been used to elucidate the mechanism by regularly exploring such molecular pathways which would destroy the spatial structure[41,42]. By mIF and HALO system, their spatial orientation interrelation of immune cells and immune markers would be preserved [19,43]. Our results further revealed that immune infiltration, including CD68, PD-1 and PD-L1 were significantly up-regulated and CD8 was significantly down-regulated with pLN metastasis of CC. This can be explained as macrophages and the immune molecule PD-1, PD-L1 may promote pLN metastasis of CC.
Thirdly, by Pearson correlation coefficients (R) and linear regression, we further found a significant positive correlation between CD3 and CD8, CD3 and CD56, CD3 and CD68, CD8 and CD56, CD8 and PD-1, CD8 and PD-L1, CD56 and CD68, CD56 and PD-1, cd56 and PD-L1, PD-1 and PD-L1 in CC without pLN metastasis. With pLN metastasis, these correlations were enhanced between CD3 and CD8, CD8 and CD56, CD8 and CD68. While these correlations became weaken between CD3 and CD68, CD8 and PD-1, CD8 and PD-L1, CD56 and PD-1, CD56 and PD-L1, PD-1 and PD-L1. These results indicated that a large number of T cells are activated, and enhanced communication with macrophages and NK cells may be involved in pLN metastasis of CC.
Fourthly, by spatial proximity analysis, the average distance (um) between CD8 and CD56, CD8 and CD68, CD8 and PD-1, CD8 and PD-L1, CD56 and PD-1, CD56 and PD-L1 were significantly closer with pLN metastasis. These data have demonstrated that CD8+ T cells, NK cells and macrophage may be involved in pLN metastasis of CC. To sum up to above two points, these data further proved that the interaction between T cells and NK cells, as well as T cells and macrophages may promote the metastasis of pLN metastasis in CC.
Lastly, based on randomly selected sixty percent patients, a diagnostic prediction model was established and the AUC can reach 0.843. The nomogram incorporates five items of PD-1, PD-L1, the average distance of CD56 to PD-1, the average distance of CD56 to PD-L1, and the average distance of PD-1 to PD-L1. By internally validation with the remaining 40 percent of cases, a new ROC curve was emerged and the AUC reaching 0.888.
Summary with this information, T cells interact with NK cells and macrophages through PD-1 and PD-L1 to promote pLN metastasis of CC.
Our nomogram can serve as an effective preoperative predictive tool to assess pLN status in CC patients.
Cervical biopsy is an essential step before the diagnosis of CC. In the process, we can take part of cancer tissue through cervical biopsy for mIF detection. HALO system is further applied in quantitative and spatial analysis. By the diagnostic prediction model, we can predict pLN metastasis preoperatively, thus avoiding unnecessary routine pLN dissection. In this way, we can avoid additional surgical trauma and possible complications for patients.