Fig. 2

KM and univariate Cox analyses of the impacts of clinicopathological variables on prognosis. A The heatmap showed the logrank test p-value of 7 important clinicopathological variable groups for DFS and OS. The color changes from blue to red when p-value decreases. The pheatmap R package was used to draw the heatmap. The names of variables indicate the method of grouping as follows: “LNM_-_Pelvic_PelvicAortic” represents classifying as LNM negative, only pelvic LNM and pelvic and para-aortic LNM; “Histology_-_with_dominant” represents classifying as pure NECC, with NEC differentiation, and NECC is dominant in histology. B Univariate Cox regression of all important clinicopathological variables for DFS and OS, hazard ratio, and 95% confidence interval was shown in forest plot using the “forestplot” R package. #Import into univariate Cox model as consecutive variable originally. ##Import into univariate Cox model as consecutive variables by numbering, stage from early to late, LNM from negative to high, or increase from only pelvic to pelvic and para-aortic LNM, DIM from superficial to deep