How to explain in an article that some of the first-order variables in a second-order variable are not significant？

[mukji]

As shown in the figure, there are two first-order variables that are not significant in the second-order formative variable. The first-order variables are reflective variables. The LV scores are used to replace the original first-order variables here. The path is the p-value, and two of the first-order variables are not significant. I would like to ask if this situation can be explained in the article? The variable source article uses Amos to develop the five dimensions of this variable, but does not use pls to verify the second-order construct. If I want to delete these two first-order variables, is there a reason to explain in the text? Because I have not found any literature that mentions the same problem.Any paper, literature, or tutorial? Thanks a lot！！！！