Element-wise mean of several big.matrix objects in R, with na.rm = TRUE

Question

This question is quite similar to the one asked here: Element-wise mean of several big.matrix objects in R.

However, now I aim to calculate the element-wise mean of 18 filebacked big.matrix objects (each > 4GB) taking into account that na.rm = TRUE.

For normal sized matrices I would do the following:

#Generate a list of 3 example matrices
my.list <- list(matrix(c(1:9), nrow = 3), 
                matrix(c(10:18), nrow = 3), 
                matrix(c(19:27), nrow = 3))
my.list[[1]][1,1] <- NA
my.list[[3]][2,1] <- NA

#Take element-wise mean with na.rm = TRUE 
test <- apply(simplify2array(my.list), 1:2, mean, na.rm= TRUE)

#This results in:
> test
      [,1] [,2] [,3]
 [1,] 14.5   13   16
 [2,]  6.5   14   17
 [3,] 12.0   15   18

I'm using a 64x Windows10 machine with 16GB of RAM.

Any suggestions are very welcome. Thanks!